222222 i

Information Worker Productivity Enabled by IT System Usage: A Complementary-Based Approach Natallia Pashkevich

Information Worker Productivity Enabled by IT System Usage A Complementary-Based Approach

Natallia Pashkevich

Abstract

Assessing the conditions of productivity of individual workers who process infor-mation and use IT has been a concern for many researchers. Prior studies have applied different theoretical foundations to study the relationship between IT use and produc-tivity at individual level in post adoption scenarios and have provided mixed results. In the last decades, the proposition that there is a need for a set of factors to be changed in a synchronized fashion when using an IT system has received particular attention. Very little, however, is known about the configurations of these factors at individual level. To investigate this gap, we have designed a new research model of an infor-mation worker’s individual productivity when a more aligned IT system is used in a synchronized manner with both individual and organizational factors. The formulated research model is grounded on the complementarity theory, functioning here as a meta-theory guiding the linking of productivity theory, Kirton’s adaption-innovation theory, and several theoretical bodies on the structure of production processes and human resource management. The formulated model was tested in two empirical stud-ies – a longitudinal quasi-randomized field experiment and an online experiment – conducted to investigate configurations of complementary factors that increase productivity when a new, more aligned IT system is used.

Overall, the two studies shed important light on configurations of complementary factors and the improvement of the research design to study their impact on IT-ena-bled productivity. The obtained results contribute to the research that focuses on in-dividual information worker IT-enabled productivity as well as research that rests on the complementarity theory with new configurations of complementary factors that, when matched correctly, can increase individual productivity of information workers. Eventually, the studies presented here advocate that further research is needed to in-crease our in-depth understanding of complementary factors and their impact on in-dividual IT-enabled productivity of information workers. Keywords: Complementarity; Individual level; Information worker; IT use; IT-ena-bled productivity.

©Natallia Pashkevich, Stockholm University 2016 ISBN 978-91-7649-498-1 Printed in by Holmbergs, Malmö 2016 Distributor: Stockholm Business School, Stockholm University

To Volha and Renato

Acknowledgements

The results from five years of dedicated, intensive and exciting work are pre-sented in this dissertation. It would never have been finalized without support and help from a number of people and research groups.

First and foremost, I wish to express my sincere gratitude to my supervisors, Professor Darek Haftor and Professor Thomas Hartman. Darek has been supportive since the days I took my first research steps. He supported me both academically and emotionally through the whole research process. Thanks to him I had an opportunity to run two unique empirical studies and test the research model formulated as a part of this dissertation. I am very thankful to Darek that during the whole five years he gave me reasonable freedom and moral support. Thomas was also very supportive during the whole period of my study. Our joint meetings stimulated my thinking and gave me a feeling of a constant progress and encouragement.

Besides my supervisors, I would also like to thank my pre-opponents, Pro-fessor Jan Löwstedt, Professor James Sallis, Professor Anita Mirijamdotter, Associate Professor Fabian von Scheele, and Assistant Professor Olov Isaks-son for their valuable and instrumental comments during my research mile-stones. I am thankful for the insightful comments and challenging questions which broadened my outlook and stimulated me to consider my research from various angles and perspectives.

My sincere thanks also goes to faculties from the Sauder School of Busi-ness at the University of British Columbia (UBC) in Vancouver, Canada, where I was a visiting scholar in the winter of 2016. I am very grateful to Professor Izak Benbasat for sharing his valuable knowledge and providing me with critical comments regarding the research model as well as empirical in-vestigations. I am also grateful to Assistant Professor Adam Saunders for in-sightful comments about IT-enabled productivity research in general and at the individual level in particular. I wish also to thank Assistant Professor Ning Nan for an exciting course which gave me an understanding of classical and contemporary Information Systems research topics.

My sincere appreciation goes to professors, colleagues and friends from the Swedish Research School of Management and IT and the Gunilla Bradley Centre for Digital Businesses. I thank all of you for the stimulating discussions during seminars and meetings. Your advices and comments on my research are inestimable. I am also grateful to the Stockholm Business School for the enjoyable time and comfortable research conditions. I would also like to thank

the pharmaceutical company that provided me with the unique opportunity to conduct my research in real life settings. I am also thankful to people that took part in the online experiment and made it possible to get valuable results.

I thank some particular key people in my life that believed in my research abilities and supported my research journey in one or another way. I am very thankful to Lene Fenger Clausen who not only introduced Scandinavia to me but also made a great contribution at the beginning of my studies. I am also thankful to Miranda Kajtazi who demonstrated me what the research journey looks like and instilled confidence in my own strength.

Last but not least, I would like to thank my family for all their patience and encouragement. For my wise parents who gave me freedom in choosing my own life direction. For the presents of my great sister Volha who is an engine in all my life aspects since I was born. She always encouraged and still en-courages me in all my undertakings. I am thankful to Volha for her availability and active participation in discussing my research any day and time. And most of all I am very thankful to my husband Renato for being here and supporting me during all stages of my studies. He also made it possible for this disserta-tion to be completed by sharing his professional knowledge and helping me with the practicalities of the online experiment. Thank you all for allowing me to grow as a research scientist!

Contents

Abstract ......................................................................................................... iv

Acknowledgements .................................................................................... vii

List of Figures ............................................................................................ xiii

List of Tables ............................................................................................... xv

1. Introduction ...................................................................................... 17 1.1 Research motivation and significance ................................................ 19 1.2 Research question and aim ................................................................. 21 1.3 Key concepts ....................................................................................... 22

1.3.1 Information work and information worker ................................... 23 1.3.2 IT use for work ............................................................................. 24 1.3.3 The concept of productivity .......................................................... 25 1.3.4 Non-complementary approach ...................................................... 26 1.3.5 Complementary approach ............................................................. 27

1.4 Scope and constraints ......................................................................... 28 1.5 Dissertation outline ............................................................................. 29

2. Literature review .............................................................................. 33 2.1 Individual productivity factors in the context of information work ... 34

2.1.1 Individual differences ................................................................... 40 2.1.2 Job/task design .............................................................................. 42 2.1.3 Human resource management practices ....................................... 44 2.1.4 IT use ............................................................................................ 49

2.2 Individual IT-enabled productivity: non-complementary approach ... 50 2.2.1 Conceptualization of IT use and individual productivity ............. 51 2.2.2 IT use, multitasking, and individual productivity ......................... 54 2.2.3 IT-enabled communication patterns and individual productivity . 56 2.2.4 The impact of individual and organizational factors on IT-enabled

productivity .................................................................................. 60 2.3 A shift towards a complementary approach in IT-enabled productivity

studies ................................................................................................. 65 2.4 Complementary factors of productive IT use at firm, establishment

(plant), and individual level ................................................................ 68 2.5 Summary of current knowledge .......................................................... 73

3. Formulation of research model ....................................................... 77 3.1 The notion of complementarity .......................................................... 77 3.2 General research model and hypotheses ............................................. 80 3.3 Complementarities of individual information worker productivity .... 85

3.3.1 Complementarities between cognitive style and operational production mode ........................................................................... 86

3.3.2 Complementarities between cognitive style and training mode ... 89 3.3.3 Complementarities between cognitive style, operational

production and incentive modes ................................................... 92 3.3.4 Complementarities between cognitive style and decision-making

mode ............................................................................................. 94

4. Research methodology ..................................................................... 99 4.1 Overview and rationale for research approach ................................... 99 4.2 Study 1: Sales representative productivity ....................................... 103

4.2.1 Research settings ........................................................................ 104 4.2.2 Conceptual set-up and complementarity configurations ............ 112 4.2.3 Measurement ............................................................................... 116 4.2.4 Participants ................................................................................. 122 4.2.5 Data access and data collection procedure ................................. 124 4.2.6 General analytical strategy ......................................................... 126

4.3 Study 2: Software programmer productivity .................................... 130 4.3.1 Experimental design, procedure and pilot testing ....................... 130 4.3.2 Participants and experimental material ....................................... 134 4.3.3 Operationalization of the main constructs .................................. 137 4.3.4 Dependent measures ................................................................... 140 4.3.5 General analytical strategy ......................................................... 141

4.4 Validity and reliability ...................................................................... 143 4.4.1 Conclusion validity ..................................................................... 143 4.4.2 Construct validity ........................................................................ 145 4.4.3 Internal validity ........................................................................... 147 4.4.4 External validity .......................................................................... 148

4.5 Methodological limitations ............................................................... 150 4.6 Research ethics ................................................................................. 152

5. Data analysis and research findings .............................................. 155 5.1 Study 1: Sales representative productivity ....................................... 155

5.1.1 Assessment of main assumptions ............................................... 155 5.1.2 The intervention (operational changes) effect ............................ 160 5.1.3 Robustness check and persistence of the intervention effect ...... 168

5.2 Study 2: Software programmer productivity .................................... 171 5.2.1 Assessment of main assumptions ............................................... 172 5.2.2 Output interpretation ................................................................... 174 5.2.3 The source of interaction ............................................................ 181 5.2.4 Effect size, power and final results ............................................. 187

5.3 Summary of the research findings .................................................... 188

6. Discussion ........................................................................................ 193 6.1 Discussion of empirical results ......................................................... 193

6.1.1 Discussion of empirical results: Study 1 .................................... 193 6.1.2 Discussion of empirical results: Study 2 .................................... 196 6.1.3 Discussion of empirical results: Summary ................................. 200

6.2 Value of the research model ............................................................. 201 6.3 Theoretical contributions .................................................................. 203 6.4 Managerial implications ................................................................... 210 6.5 Limitations and avenues for future research ..................................... 212

7. Summary and conclusions ............................................................. 219 7.1 Summary of this research ................................................................. 219 7.2 Contributions in conclusion .............................................................. 221

Appendix A ................................................................................................ 225

Appendix B ................................................................................................ 228

Appendix C ................................................................................................ 230

Appendix D ................................................................................................ 234

Appendix E ................................................................................................ 236

Appendix F ................................................................................................ 239

Appendix G ................................................................................................ 244

Appendix H ................................................................................................ 247

Appendix I ................................................................................................. 251

References .................................................................................................. 253

Sammanfattning ........................................................................................ 277

List of Figures

2.1 Scope and boundaries of the literature review ………………... 34 3.1 Cube-view of complementary interaction……………………... 78 3.2 General research model of complementary factors and their

value range…………………………………………………….. 81

4.1 The evolutionary process of the methodology of inquiry……... 102 4.2 The overall process with less aligned IT system support……… 106 4.3 The overall process with more aligned IT systems support…… 108 4.4 Study design conceptual set-up………………………………... 113 4.5 Types of training and education in a pharmaceutical company. 119 4.6 Timeline of major events………………………………………. 125 4.7 Percentage of participants from different geographic areas……. 135 5.1 Normal distribution histograms……………………………….. 159 5.2 Scatterplots of standardized residuals…………………………. 159 5.3 Parallel trend assumption for a number of calls and products

sold before intervention……………………………………….. 160

5.4 Intervention effect on a number of sales calls………………….. 163 5.5 Intervention effect on a number of products sold……………… 167 5.6 Estimated marginal means of time and quality scores…………. 180

List of Tables

2.1 Summary of current research topics and factors affecting indi-vidual productivity in the context of information work……….

37

2.2 Position map…………………………………………………. 71 3.1 Operational process structure characteristics………………… 86 4.1 Pre-existing sales practices and its IT system support………... 107 4.2 New sales practices and functions supported by the more

aligned IT system…………………………………………….. 110

4.3 Proposed designs/complementarity set-ups………………….. 114 4.4 Descriptive statistics…………………………………………. 123 4.5 Description of data collection and data sources………………. 124 4.6 Key components of the DID model…………………………... 129 4.7 Research design for the experiment and expected results…….. 131 4.8 Experimental sessions………………………………………... 132 4.9 Participation outcomes for four months of online availability... 134 4.10 Descriptive statistics…………………………………………. 136 4.11 Operationalization of the main constructs……………………. 137 4.12 Key components of the factor effect model for repeated

measures ANOVA…………………………………………… 142

5.1 Criteria (assumptions and restrictions) for performing DID

analysis……………………………………………………… 156

5.2 Correlation matrix and VIF diagnostics……………………… 157 5.3 Skewness and kurtosis of residuals…………………………... 158 5.4 DID results demonstrating the impact of the intervention

(treatment) on sales calls, controlling selected covariates (De-sign 1, 2, 3 and 4)……………………………………………...

161

5.5 DID results demonstrating the impact of the intervention (treatment) on sales calls between pairs of designs……………

164

5.6 DID results, demonstrating the relationship between the num-ber of products sold and treatments, controlling selected co-variates (Design 1, 2, 3 and 4)………………………………...

165

5.7 Robustness check…………………………………………….. 168

5.8 Persistence of the intervention impact (time placebo set in the quarter 4, year 2014 and quarter 2, year 2015)………………...

170

5.9 Criteria (assumptions and restrictions) for the repeated measures ANOVA……………………………………………

172

5.10 Skewness and kurtosis of the dependent variables…………… 173 5.11 Test of homogeneity of variances…………………………….. 174 5.12 Main effect for complementarity set-up……………………… 175 5.13 ANOVA test for between-group differences…………………. 175 5.14 Assignment means…………………………………………… 176 5.15 ANOVA test for within-group differences and interaction…... 177 5.16 Post hoc test for within-group differences……………………. 179 5.17 Interaction means…………………………………………….. 180 5.18 Independent samples test for time and quality scores………… 182 5.19 Repeated measures one-way ANOVA for assignment (re-

ported by complementarity set-up)…………………………... 185

5.20 Effect size and power for time and quality scores…………….. 187 5.21 Summary of expected versus obtained results: Study 1………. 189 5.22 Summary of expected versus obtained results: Study 2………. 190 6.1 Major theoretical contributions………………………………. 205 6.2 Limitations and avenues for future research………………….. 212

17

1. Introduction

This introductory chapter comprises five sections. First, we present the rationale and background as to why this research is important and, consequently, should be undertaken in addition to the current body of knowledge. Second, we formulate the research question and aim addressed in this research. Third, we clarify key concepts that are frequently used throughout the present research. Finally, we present an overview of the constraints and monograph outline.

The activity and success of economic organizations such as banks, insurance companies, recruiting firms, financial, businesses, consulting and accounting services firms, software and data processing companies increasingly depend upon the IT use and productivity of information workers whose work primarily consists of the production, analysis, collection, processing, manipulation and distribution of information (Davenport, 2011). These organizations heavily invest in IT to equip their employees with appropriate technology hoping that doing so will increase their productivity (Sundaram et al., 2007; Xue et al., 2012). However, despite substantial investments in IT, many organizations do not experience the anticipated productivity growth (Dwivedi et al., 2015; Taherdoost & Keshavarzsaleh, 2015), suggesting a need to better understand how the use of IT can lead to increased productivity of an individual information worker.

Recent research investigating IT-enabled productivity demonstrates that the full potential of existing IT is not exploited and a broad effect of IT use on individual productivity of information workers is not well defined (Aral & Van Alstyne, 2011; Aral et al., 2012a; Webster, 2012). Moreover, frequent decisions about acquiring or using new IT are based on executives’ speculation and negotiation rather than on consistent knowledge about IT impacts (Tallon, 2014). Indeed, despite the wide penetration of IT in workplaces, little is still known about how IT use affects individual productivity and managers are facing the challenge of creating the conditions necessary for IT-enabled productivity growth of the individual information worker. Yet, without an understanding of productive IT use by individuals, organizations cannot readily identify productive information system implementation (Marchand et al., 2002) and, therefore, guarantee IT project success.

Current literature shows that a number of factors, including IT, affect individual information worker productivity (e.g. Zmud, 1979; Drucker, 1999;

18

Kessels, 2001; Parker et al., 2001; Frey & Osterloh, 2002; Kozlowski & Bell, 2003; Jain & Kanungo, 2005; Appelbaum et al., 2008; Aral et al., 2012a). However, as will be evident from the literature review in the subsequent chapter, these studies provide only a partial understanding of what conditions individual information worker productivity and it is difficult to generalize the majority of these studies. Moreover, in current studies, a non-contingency notion is assumed that one model fits all situations, mostly neglecting relationships between contextual variables and the associated performance. Given these limitations, research is needed to explore specific conditions under which IT use can have a positive impact on individual information worker productivity. The results of such research can help us better understand how individual information worker IT-enabled productivity can be increased while providing managers with a clear strategy for IT use with respect to individual productivity and contextual richness of information-intensive environment.

Along with contingency thinking, a variety of recent theoretical and empirical evidence suggests that a positive impact of IT use on productivity rarely occurs without potentially complementary1 changes at different economic levels2. For example, complementarities are shown to be a strong predictor of IT-enabled productivity at macro-, meso-, and micro-levels (Bugamelli & Pagano, 2004; Lim et al., 2004; Lee et al., 2005; Jorgenson, 2007; Aral et al., 2012b; Tambe et al., 2012). Although there is a substantial evidence of positive complementarity effects on IT-enabled productivity at more aggregate levels of the economy, we still lack knowledge on complementarities of productive IT use at the individual level (Ennen & Richter, 2010; Brynjolfsson & Milgrom, 2013). To the best of our knowledge only two studies have made attempts to identify complementarities of productive IT-use at this level. For example, Athey and Stern (2002) experienced challenges to demonstrate a synergistic effect between the adoption of a specific health related application and working practices on emergency health care outcome. Autor et al. (2003) demonstrated that computer capital is a complement in performing complex communication and decision-making tasks, yet without showing the exact impact of the identified complementarities on productivity indicators. Therefore, little work has been done to investigate complementary factors and how to synchronize them in order to increase IT-enabled productivity at the level of the individual information worker from theoretical and empirical perspectives. In this research, we have developed a framework for addressing this gap by assuming

1 Complementarity is the notion meaning that several factors need to coexist in a synchronized manner to produce a desired outcome. 2 See, for example, Brynjolfsson et al. (2002); Bugamelli & Pagano (2004); Lim et al. (2004); Hu & Quan (2005); Kohli & Grover (2008); Bloom & Van Reenen (2011); Aral et al. (2012b); Tambe et al. (2012); Brynjolfsson & Milgrom (2013); Schryen (2013); Chae et al. (2014).

19

the so-called complementarity view, which holds that productivity benefits from IT use are to be regarded as produced by a number of individual and organizational factors that are synchronized in such a manner as to increase productivity.

1.1 Research motivation and significance There are several arguments why it is important to conduct this research. First, the introduction of various sophisticated ITs into human, social, business and industrial affairs has created several kinds of positive effects at different economic levels. Yet, productivity is a key and central indicator in all economic activity (Porter, 1990). Productivity increase at national level is associated with improved living standards (Brynjolfsson, 2003). At industry level, productivity increase allows industries to compete with other sectors of the economy (Dedrick et al., 2003). At company level, productivity increase automatically leads to increased capabilities in cost reduction, innovativeness and customer satisfaction (Clemons et al., 1993). Productivity increase at the workplace is associated with better motivation and working environment improvement (McNeese-Smith, 1996). Therefore, economic growth and development cannot be sustained without improvements in productivity.

Second, in recent decades, the structure of the workforce has been changing significantly due to the emergence of the information age. Whereas the category of material workers was predominant just a few decades ago (D’Agostino et al., 2006), now the information sector occupies a dominant position (Wolff, 2005; Lal, 2005; Karmarkar & Apte, 2007). According to the latest labor statistics data, information workers account for as much as half of the labor force in developed countries: for example, 53% in Sweden, 54% in Denmark, 58% in Switzerland and 56% in Canada and Australia (ILO, 2008)3. This category of workers, unlike others, presents sustainable growth and the cost of their provision, which is the most significant input of production costs, is relatively high (Castells, 2011). While information workers use IT as the main production tool, we have little knowledge about how it affects individual productivity.

Third, the emergence and stable growth of the information workers’ rate in the total workforce and the continuing development of IT imply that the competitive advantage of businesses will increasingly depend on their ability to make information workers more productive (Hopp et al., 2009). A clear understanding of how information workers create value and what factors 3 Information workforce data was derived from the International Labor Office and calculated based on Dordick’s theory (Lal, 2005).

20

affect their productivity will allow practitioners to reduce costs, to control the production process, to determine high and low productivity staff, to improve service quality, to make better decisions and, at the same time, to determine directions for further development. Knowing the best way to use a particular combination of available resources allows companies to do more with the same level of costs, something that is central in a situation of restricted resources.

Fourth, in today’s highly competitive environment, there is an increased emphasis on individual achievement and productivity from IT use in post-adoption contexts (Jain & Kanungo, 2005; Aral et al., 2012a). For example, both the technology acceptance model (Davis, 1989) and the task-technology fit model (Goodhue & Thompson, 1995) have long been applied to understanding the success of IT use. Yet, a critical evaluation of both models lays a foundation for this research. Although acceptance of IT by end users is a good indicator of the initial success of the system (Davis, 1989), such acceptance does not necessarily imply productivity gains (Jain & Kanungo, 2005). In other words, an individual can adopt a new IT system and perceive it as useful and easy to use, yet may not be productive per se just by using this IT system. By studying the role of the task-technology fit on individual performance, previous studies have also found it difficult to demonstrate positive relationships between IT use and individual performance (Goodhue & Thompson, 1995; Lucas & Spitler, 1999). Moreover, neither the technology acceptance model nor the task-technology fit model reflect such a point that people, due to differences in cognitive styles, can differ in their acceptance and use of IT. Yet, the aforementioned studies point to a possibility of additional factors, including the need to consider an individual and task being performed together with IT use in order to increase individual performance.

Fifth, the focus of the current research into the impact of IT use on productivity arrived at the so-called “nano” level (Wu et al., 2009; Aral et al., 2012a). In this research, the concept of nano4- (individual/process/task) level is used interchangeably and refers to the level of analysis which indicates the relationship of individuals and integration of these individuals into activities executed in a professional context. The focus on individual level can be explained by the fact that “…Only by understanding the individual level of productivity, can researchers and practitioners begin to build theories and models that deal with the diffusions and synergies that occur when individuals are grouped into work teams, departments, organizational systems and economies.” (Ruch, 1994, pp. 105-106). While the individual level of investigation is important, studies at this level5 applied different theoretical

4 Hereafter referred in this text to individual level. 5 See, for example, Czerwinski et al. (2004); Kvassov (2004); Jain & Kanungo (2005); Aral et al. (2006); Sundaram et al. (2007); Chung & Hossain (2009); Wu et al. (2009); Aral & Van Alstyne (2011); Webster (2012).

21

foundations to investigate the impact of IT use on productivity and, therefore, the results are mixed and difficult to compare. This implies that narrow perspectives are not good enough to account for how productivity gains are generated by individuals.

Sixth, in the last decade, the focus of the existing research arrived at the point that, if we are to be able to understand the actual impact of IT on productivity, we need to look beyond the technology and concentrate on complementary factors (Bloom & Van Reenen, 2011; Brynjolfsson & Milgrom, 2013). These studies have addressed the need for the joint and well-synchronized adaptation of IT use and innovative human resource management practices to increase productivity at firm, establishment, and individual levels. All this suggests that there may be several ways to organize tasks and operations of an organization to benefit from IT use in terms of productivity and that there may be some way that currently is not known or employed in practice.

Finally, complementarity theory, with its recently acquired popularity in organizational economics, states that changing only one or a few factors at a time in organizational settings may not come close to all the benefits that are available from a coordinated effort (Brynjolfsson & Milgrom, 2013). While the complementarity theory offers a broad perspective to show the importance of introducing a “system of complements” to increase productivity, we still lack knowledge about complementary factors that affect IT-enabled productivity, especially at the individual level (Ennen & Richter, 2010). Particularly, we have little knowledge about which individual and organizational factors can be complementary in a situation when a new, more aligned IT system is used by information workers. In this research, we present a new research model for addressing this gap aimed at research productivity growth at the individual level when IT is used, jointly and in a synchronized manner, with both individual and organizational factors. Therefore, an understanding of productive IT use by an information worker is of critical importance and relevance.

1.2 Research question and aim It has already been mentioned that in general there is a dearth of knowledge about complementarities of productive IT use at the individual level. In particular, we lack knowledge of what individual and organizational factors have to be synchronized when a more aligned IT system is used in an information-intensive environment to increase individual productivity. Therefore, the research question is formulated as follows:

22

What are the configurations6 of complementary factors that influence productivity, considered at nano-level (i.e. individual/process/task) in the context of an information-intensive environment, when a more aligned IT system7 is used?

Therefore, the aim of this research is to develop and test a new research model of complementary factors that affect individual information worker IT-enabled productivity.

In order to answer the research question, we reviewed relevant literature, formulated the research model and tested it in a longitudinal quasi-randomized field experiment of sales operations and an online experiment of software programming. By studying two information-intensive occupations, including sales representatives and software programmers, we were able to test the formulated research model of complementarities and to shed important light on their effects on IT-enabled productivity at the individual level of an information worker. Both sales representatives and software programmers are good examples of an information worker since both use non-trivial IT systems as their main production tool and require cognitive skills to process information which is an input and output of the production process (North & Gueldenberg, 2011). Below, we examine central concepts to establish basic conceptual foundations for this research, in which the analysis of the link between IT use, complementary factors and productivity of information workers at the individual level is carried out.

1.3 Key concepts In this section, we outline several central concepts used in this research. First, the concepts of information work and information worker are elaborated. Second, the concept of IT use for work is defined. Third, the concept of productivity at the individual level within the scope of this study is elaborated and defined. This section ends with the notions of non-complementary and complementary approaches.

6 The way factors are arranged into the system. 7 IT system that offers information and information processing functionality that is more ad-justed and tuned to support a specific kind of work process.

23

1.3.1 Information work and information worker Typically, economic organizations and their work are categorized based on the ‘form’ of a products’ delivery to and consumption by the receiver, offering the goods-service continuum (Vargo & Lusch, 2004). However, products can also be conceived in terms of their ‘content’, with regard to being material or information (Porat, 1977; Dority, 2006), offering two types of work – material work and information work.

The main distinctive features of material and information work have been summarized earlier (North & Gueldenberg, 2011), and are as follows. First, in contrast to material work, information work is characterized by specific work object which is intangible since information is an input and output of the production process. Second, information work requires cognitive skills of employees involved in information processing. Third, the production process in information work is characterized by a close interaction between individuals and IT use that is essential, supportive if not one of the most important production tools.

An accountant creating a report, a journalist preparing an article, an architect working on a project, a programmer writing software for a particular purpose, a physician summarizing the symptoms and likely diagnoses of a patient, a financial adviser analyzing a client’s investment profile, a manager of the company trying to come up with a long-term strategy to make his or her business more profitable – these are all examples of information workers.

Given the centrality of the distinctive features between the material and the information work, in this research, we adopt the definition of an information worker developed by Szabo and Dienes (1998) who consider an information worker as an information-dependent, technology reliant, educated employee who uses data and information as the main inputs of the job, whose work time is spent engaged in professional tasks, and whose major product of work is the distillation of information. We consider this definition as appropriate for this research, since it covers the essential distinctive features of information workers and demonstrates the complexity of information work.

However, we recognize that the distinction between material and information workers is approximate. For example, material work and workers may be compared to information work and workers with the help of practical examples of two jobs, i.e. a wall painter and an architect. For example, the painter receives a specification for which walls to paint, with which color and when, all of which is information communicated to the cognitive faculties of the wall painter, who then opens a can of paint and uses various brushes to paint the specified walls. The planning and coordination of the actual painting are a kind of a given function that is informational and that employs the painter’s cognition. Yet, the painting work itself is done using the worker’s physical capabilities. The architect, on the other hand, who receives a

24

specification for a desired house, uses his or her cognition and also an IT system to design a new house for the client; here both the governance function (coordination and communication) and the core function are information processing. The mere physical activity of the architect is to translate the conception present in the architect’s cognition into information symbols in the IT systems.

This illustration of two jobs and workers, one material and one information, shows the key distinction between the two kinds of jobs and workers: material workers conduct work governance with their cognition, hence informational, while the core functions are conducted physically, with the worker’s body. On the other hand, information jobs and workers conduct both the governance function and the core function with the cognitive processing, possibly using the human body to operationalize or transfer information from cognition to another human cognition or an IT system. From that we may derive that our material work and workers have a core function that receives materials as input and generates materials as output, while the opposite is true for information work and workers who receive information input and generate information output. It also follows clearly that IT systems are particularly useful for supporting the execution of the core function of the information workers as these systems process information symbols (generate, store, transform, and transfer) (Simon & Newell, 1964), hence can automate at least some parts of human cognition. A third and blended category of work and workers may be those whose core functions do both, for example a worker who receives a shipment of goods into a warehouse may produce a receipt with the information that the goods have been received while having physically carried in the shipment.

1.3.2 IT use for work

To achieve a productivity increase, IT tools have to be used by individuals (Jain & Kanungo, 2005). Previous studies have made different attempts to define the concept of IT use. For example, Jain and Kanungo (2005) perceive the concept of IT use as the amount of time an individual spends using an IT system. Burton-Jones and Straub (2006) include structural and functional dimensions in the concept of IT use. Sundaram et al. (2007) developed the concept of IT use, which is based on frequency, routinization, and infusion dimensions. Venkatesh et al. (2008) conceptualize IT use concept in terms of duration, frequency, and intensity of its use. While the aforementioned studies applied different theoretical foundations to conceptualize IT use and mostly collected self-reported data, the results of the impact of IT use on individual performance are difficult to compare. Despite a variety of approaches in

25

defining IT use concept, we explain how we understand this phenomenon in terms of this research.

To conduct their work, information workers use a variety of IT tools, including word processing, spreadsheets, emails, Internet access, social media tools, etc. Besides these numerous IT tools, information workers also use IT systems that support a specific kind of work process and are essential for information processing. This research is limited to a particular situation when a more aligned IT system is deployed and used to support operational processes. Previously, it was proposed that in order to investigate the impact of IT use on performance, there is a need to focus on the functions that this IT use offers (Schienstock et al., 1999). In essence, IT can be used for acquiring, storing, processing, and distributing information (Simon & Newell, 1964). Therefore, with regard to our particular situation, IT use is understood here as the use of skills-extension, technological tools (in our case, IT systems) that assist individuals in the completion of their professional tasks through information acquisition, storing, processing, and distribution. Given these points, such a definition of the IT use concept can be generalizable for different types of information work when a more aligned IT system is used.

We also take into account two ways of thinking towards IT use that have been formulated in information system discipline (Alter, 2013). One conception regards the use of IT merely as a work tool that is an aggregate of software and hardware that are used by individuals. Another approach conceives the use of IT in terms of systems, or more specifically recognizes the broader context in which IT is used, including who uses the system, where, how and when. Tool thinking is criticized by its restricted view on IT use and misleading towards understanding how IT is used in organizations, including its contribution to productivity and performance (Brynjolfsson, 2003). Indeed, there is evidence that two similar firms (in the same market and industry) may acquire similar IT systems, yet receive different performance outcomes (Brynjolfsson & Milgrom, 2013). In this research, we support system thinking and recognize that the context interacts with the actual use of IT and therefore both influences it and is influenced by. Therefore, in our definition of IT use concept we assume both a functional perspective and the context of its use.

1.3.3 The concept of productivity In a broad sense, productivity is defined as the ratio of output to input (Syverson, 2011). Yet, this simple relation was challenged by Ramirez and Nembhard (2004) who summarized methodologies of information worker productivity measurement. Particularly, they demonstrated that there is no one acceptable approach to measure productivity of this category of workers. Moreover, the intangible nature of input and output of the operational process

26

in an information-intensive environment makes an assessment of productivity complicated and therefore requires further research. Due to the complexity of information work, the first step on a way to study productivity in an information-intensive environment has to be based on a clear understanding of what dimension of the productivity concept is under investigation (ibid.).

Two main dimensions of productivity are often discussed in economic and business literature, including efficiency (the ability to produce using resources without waste) and effectiveness (the ability to produce a desired result) (Tangen, 2005). In this research, we specifically focus on the concept of productivity understood in terms of internal individual work efficiency, namely how much output can be generated from a given amount of input. The focus on this dimension is explained, on the one hand, by the main concern of this research to demonstrate how available factors can be used together with a new, more aligned IT system to increase individual information worker productivity. On the other hand, the complementarity theory by itself is largely based on a premise of the efficient use of available resources. Therefore, we have limited ourselves to a certain productivity dimension that was mostly suitable for the present research.

1.3.4 Non-complementary approach

Most of the studies addressing the relationships between IT use and individual productivity of information workers can be classified into two categories depending on how they define the relationships between IT use and surrounding factors as being either non-complementary or complementary. Individual information worker productivity studies have mostly been focused on the application of the non-complementary approach and identified the impact of a single element, mostly IT use on individual performance. The gist of these studies is that using IT may or may not guarantee increased individual productivity. For example, Kvassov (2004) argued that IT use affects individual productivity through temporal dimensions and time personality. Jain and Kanungo (2005) showed that the “nature of IT use” positively affects individual productivity. Yet, Sundaram et al. (2007) found that IT effect on individual productivity depends on the stage at which an IT tool is utilized (e.g. frequency, intensity or infusion). Although Aral et al. (2006) demonstrated that a more central position in an email network is associated with greater individual productivity, Aral and Van Alstyne (2007) showed that the size of the email network is not associated with individual productivity.

By analyzing previous studies, some interesting inferences can be made. For example, studies that applied a non-complementary approach came to conclusions that technologies can have an ambivalent impact on individual productivity. This was explained by the environments in which individuals

27

operate (Haner et al., 2009). Some studies also identified that other factors such as personal characteristics (Kvassov, 2004; de Koning & Gelderblom, 2006; Deng et al., 2008), information work infrastructure, and managerial practices that are not mutually exclusive may affect the relationships between IT use and individual productivity (Sonnentag, 2003; Hoop et al., 2009). This implies that a complementary approach may help us identify how the use of IT together with surrounding factors can increase individual productivity of information workers.

1.3.5 Complementary approach

In general, the resources of the firm may affect each other in three possible ways: independent, substitutive and complementary (Parmigiani & Mitchell, 2009). An independent relationship exists when changes in the level of one element do not affect the value of another element. A substitution relationship exists when the increasing value of one element diminishes the value of another element. A complementary relationship exists when a change in the level of one element magnifies the impact of another element.

Complementarity theory is not a regular specific empirical theory; it is rather more of a generic theory, as it says that two or more factors tend to complement each other in a certain manner to produce a certain outcome at a certain time (Milgrom & Roberts, 1990), yet it does not specify which factors complement each other. There are two investigative approaches in the complementarity theory: the interaction approach and the systems approach (Ennen & Richter, 2010). The interaction approach is based on the analysis of the interaction between a limited number of clearly identified factors. While this approach provides a high degree of granularity, it does not take into account contextual settings that may affect the relationships between the variables investigated (Porter & Siggelkow, 2008). In contrast, the systems approach investigates the impact of a system of multiple factors on performance outcomes. The main distinctive feature between the aforementioned approaches is that while the interaction approach investigates the nature of the actual interactions between the studied factors, typically only a few, the systems approach studies the performance effects of the entire system of multiple factors. According to the nature of the research problem, we follow the systems approach since we have very little knowledge about which systems of complementary factors are needed to increase information worker productivity when a more aligned IT system is used. When the system of complementary factors that generate productivity gains has been established, it can then be possible to study the nature of the interactions of the identified factors based on the interaction approach.

28

In essence, a complementary approach offers a broad perspective to show the importance of introducing a “system of complements” to obtain greater IT use returns in terms of individual and task productivity. For example, Kraemer and Danziger (1990) emphasize that the interaction between people, computer technology and tasks performed is essential to explain an impact of IT on individual performance. Yet, the absence of an empirically grounded system of complementary factors at individual level points to the need for conducting further investigations.

1.4 Scope and constraints Although this research is theoretically and empirically grounded, several central research constraints have to be mentioned. Among them are the following. First, the scope of this research is limited to individual level. Yet, in an information-intensive environment, some tasks have to be conducted in groups. We acknowledge that the complementary approach can be applied at the group level. This will require theorizing of specific group level complementarities and how they interrelate with each other. However, this research does not account for either group level or jobs that require high or intensive interactions with coworkers.

Second, this research relies on the systems approach of the complementarity theory, which is based on the analysis of a system of multiple factors. The results of studies using this approach are characterized by a higher degree of finding complementary relationships between elements of the system (Ennen & Richter, 2010). Yet, the application of the interaction approach, which is based on the analysis of the interaction between a limited number of clearly identified factors, can provide a greater degree of granularity in terms of two factors’ interaction. However, the systems approach satisfies the purpose of this study to uncover configurations of complementary factors that positively affect individual IT-enabled productivity. After configurations being uncovered with the systems approach, the interaction approach can further be used to study the interaction between specific complementary factors.

Third, it is well-known that productivity requires both efficiency and effectiveness. Due to the complexity of output quantifying and qualifying, the productivity measurement in this study is based on resource utilization, mainly on the input of the productivity ratio. In other words, the efficiency criterion for productivity of the information worker and IT use measurement assessment is predominant. Although the qualitative dimension of the productivity ratio is very relevant, this research focuses primarily on the production process with a close interaction to IT use. Besides, empirical

29

measures are limited to some degree by the different kinds of information work productivity metrics available and perceived as relevant. Further studies can capture more potential IT productivity benefits.

Fourth, the approach assumed in this research is to black box IT use as an aggregate (unit), independent of its content, and to look for contextual configurations of IT use. We focus on productivity from the use of a mandatory IT system and not on ITs that are optional to use. Therefore, we limit ourselves not to a specific type of IT system, but to a particular situation when an IT system, with the functionality closely aligned with the work process, is used in an information-intensive environment to increase individual productivity.

Fifth, another potential limitation of this research is that the data was gathered from two information-intensive occupations – sales representatives and software programmers. These particular occupations do not present all information-intensive occupations. Yet, this constraint helped us limit the scope of our analysis. While there are good reasons for the suitability of such a focus, extending this research by focusing on other information-intensive occupations can further validate the generalizability of the research model.

Despite a set of constraints, this research is one of the few attempts to study complementary factors in relation to individual information worker IT-enabled productivity in a post-adoption context by introducing specific complementary factors of productive IT use. This research advances the focus beyond a single relationship between IT use and productivity of employees. It also highlights the importance of complementary factors in a situation where a more aligned IT system is used. The studies presented here and the obtained results have implications for theory as well as practice. The primary theoretical contribution lies in discovering a set of complementary factors in relation to individual productivity when a more aligned IT system is used. These particular factors can be used and manipulated by managers in the process of productive IT use.

1.5 Dissertation outline This dissertation is a monograph and consists of seven chapters. A more detailed outline is provided below.

• Chapter one begins with arguments about the need and relevance for this particular research from theoretical and practical perspectives. This chapter is organized around a discussion of research within IT-enabled productivity at individual level in an information-intensive environment. This discussion provides arguments to support the objective, aim and research question of this

30

research and, therefore, the need to design the research model and test it empirically. A description of key concepts and research constraints is also presented in this chapter.

• In chapter two, we review theories and studies addressing individual

productivity factors. More detailed attention is given to studies investigating IT-enabled productivity at individual level from a non-complementary perspective. We also review the research findings from IT-enabled productivity studies that studied complementary factors at firm, establishment (plant), and individual level. This chapter ends with a summary of the literature review that lays a foundation for the development of the research model.

• Chapter three presents the development of the research model. This

chapter begins with introducing the complementarity theory that is being applied within studies on IT-enabled productivity at firm, establishment and, to a lesser extent, individual level. Additional theoretical sources such as Kirton’s adaption-innovation theory, literature on the structure of production processes and human resource management that form the basis for the model development are described. The research model is further developed and the hypotheses are constructed together composing the theoretical framework of this research.

• Chapter four is devoted to the research methodology. This chapter

provides a detailed rationale of the research methods chosen and describes why the proposed research design is appropriate to answer the above-stated research questions and accomplish the aim of the research. The methods applied in the study are focused on a longitudinal quasi-randomized field experiment and online experiment that guided the empirical study on data collection. Assessment of the validity and the reliability is further interpreted and explained. Strategies applied to ensure ethical principles and standards for conducting research conclude this chapter.

• Chapter five describes how the obtained data were analyzed and presents

the results obtained in a longitudinal quasi-randomized field experiment and experimental study that were performed in order to test the research model and investigate configurations of complementary factors that, when matched correctly, can positively affect individual information worker IT-enabled productivity.

• Chapter six presents an analysis and interpretation of the obtained

empirical results. The results are examined with regard to the theoretical research model formulated earlier. This chapter summarizes the discussion, implication and concluding remarks of the empirical results. The implication

31

of the results is considered individually for theory and practice. This chapter ends with limitations and avenues for future research.

• Chapter seven presents a summary of the research undertaken indicating

how the aim of the research was achieved. It also includes discussion of the major findings and contributions to knowledge.

In summary, this monograph begins with a literature review. Next, the theoretical framework is developed based on the current body of knowledge. After this, an account of the research design is presented, so as to explain how the obtained data were collected, structured and analyzed. The explanation of methodological choices and a discussion on the subject of validity and reliability are also given. Further, the actual empirical work and obtained results are depicted. Finally, an extensive analysis of the results together with conclusions and suggestions for possible future research are explicated and discussed.

32

33

2. Literature review

In this chapter, we review the literature sources which guide this research. The purpose of this chapter is to identify and show the current state of knowledge with reference to an assumed research question. In particular, we demonstrate what is known from previous studies about the factors that condition the productivity of an individual information worker using an IT system. By evaluating the strengths and limitations of previous studies, we show what constitutes the research gap with regard to the empirical focus assumed here.

In order to investigate the current state of knowledge in the sphere of individual information worker productivity, we conducted a comprehensive literature review based on recent publications from leading journals within information systems, psychology, organization studies, operations management, management sciences, and economics. Articles were selected according to keywords identified in the title and abstract and then classified and analyzed based on the following criteria: the assumed research questions, the employed research methods, the obtained results, limitations and finally future research questions to be addressed.

The scope and boundaries of the literature review are presented below (Figure 2.1). In the literature review, we demonstrate that the present research focus of IT-enabled productivity has moved from material work to information work. Although a number of factors have been identified that affect individual information worker productivity, such limitations as partiality and generalizability make these results difficult to compare. Further, we demonstrate that IT-enabled productivity studies at different economic levels widely apply a complementary approach and demonstrate that the role of complementarities is significant. Eventually, we demonstrate that there is a dearth of knowledge about complementary factors at the individual level of information worker IT-enabled productivity that constitutes the research gap addressed in this research.

34

Figure 2.1: Scope and boundaries of the literature review

We begin this chapter with the literature review of theoretical bodies selected that investigate individual productivity factors with a particular focus on information work. We then review studies on IT-enabled productivity of information workers at nano-level that assumed a non-complementary approach. Further, we review studies of IT-enabled productivity at different economic levels. We summarize the research findings from those studies at macro-, meso-, and micro-levels and demonstrate a noticeable shift towards complementary thinking. We also sum up complementary factors of productive IT use at firm, establishment, and individual level. This chapter ends with a summary of what is known from previous studies to demonstrate the research gap addressed in this research. Particularly, the key finding from this review shows that there are only two empirical studies that address complementarities of individual information worker’s productivity when IT system is used; these two studies show a very rudimentary understanding of such productivity and hence call for further research.

2.1 Individual productivity factors in the context of information work

Productivity improvement as a research topic has been widely studied since the beginning of the 20th century when Frederick W. Taylor began to investigate labor efficiency in manufacturing. The so-called classical approach was based on the principle that productivity increase can be achieved

35

through standardization of the working process, a clear structure and division of work among employees (Taylor, 1947). The problem of motivation was also addressed. For example, it was established that unlike internal motivation, employees need to be rewarded externally based on their performance. Managers became responsible for setting up reward systems, selecting appropriate employees, task analysis, and process optimization. In general, productivity improvement was expected to be achieved through the rational structuring of work and the provision of compensation incentives for higher levels of output.

In contrast to the classical approach based on a mechanistic view of an organization, the human relations approach brought the idea of human variability into organizational life (Tausky, 1978). A number of physical and social factors that affect individual productivity have been researched during numerous experiments at the Hawthorne Power Plants. Together, the results of the experiments showed that a concept of motivation is more complex than payment only and that productivity improvement can be achieved through a better understanding of the individual’s psychological and social needs at work. According to the human relations approach, personnel management, leadership, organizational structure, job design and personal adjustment of employees with the organization can improve productivity to a greater degree than external rewards.

The human relations approach made a significant step forward in providing insights into human behavior in the working process and humanization of the work organization. Yet, despite technological development, growing problems in organizations, including turnover, productivity decrease, absenteeism led to the emergence of the socio-technical systems approach that made an attempt to combine principles of classical and human relations approaches (Daft, 2000). According to this approach, an organization is viewed as the interaction between the technical (tasks, tools, location) and social (individual and interpersonal needs of employees) factors within the whole. In order to design work and increase productivity, both technical and social systems need to be addressed equally. A distinctive feature of this approach is that individuals received autonomy for organizing and controlling their work with the aim to improve performance.

The contingency approach emerged as an extension of the socio-technical systems approach. This approach highlights the importance of a possible contingent fit between the nature of tasks, organizational structures and environmental influences for organizational performance (Luthans, 1973). Therefore, managers should not seek for one best way to structure and manage the organization, but instead, they should focus on the situational and contextual factors that influence management decisions. According to this approach, managers cannot rely on established rules and policies as the only guides for their choices. Instead, they have to evaluate each individual situation and make specific decisions related to those situations. For example,

36

managers cannot rely on motivational factors alone in a situation of decreasing productivity. They have to conduct more detailed analysis of the true cause of the drop in production and respond to the problem at hand.

The aforementioned approaches were formulated as a response to changes in the nature of work, work organizations, technology, economy, and society. Those approaches formulated main trends in management thinking and are still influential in the research and managerial communities. Because earlier studies were mostly conducted to optimize the working process of material workers, the operations management field followed this tradition as well. However, the structure of the workforce has changed significantly since that time. Whereas the category of material workers was predominant in the past, information workers occupy a dominant position in developed countries at the present time (Wolff, 2005; Apte et al., 2012). This implies that future economic growth will depend more on improving the productivity of information workers than on further improvement of material workers’ productivity (Hopp et al., 2009).

Different fields beyond operations management, including organizational psychology, ergonomics, economics, sociology, marketing, and information system discipline have produced a number of studies focused on factors affecting the individual employee’s productivity. In the past, it was identified that productivity of a single worker is affected by a number of factors, including well-being, ability to perform, skills and attitudes, motivation, job design, satisfaction and technical competence (Clements-Croome & Kaluarachchi, 2000). Those factors, in turn, are influenced by such factors as indoor and outdoor environment, occupation, organization and personal circumstances. Indeed, individual productivity is affected by a number of factors that range from the physical environment to psychological well-being (Davies, 2005). For example, Clements-Croome and Baizhan (2000) proposed that individual productivity depends on a cluster of variables, including personal, social, organizational and environmental variables. Yet, from a managerial perspective in contrast to the material worker, information worker productivity improvement has presented an essential challenge since the emergence of this category of worker.

Previously, Drucker (1999) proposed six general factors that determine information worker productivity, including (1) clear task definition, (2) self-management and autonomy, (3) continuous innovation as a part of the work, (4) continuous skill development and learning, (5) focus on quality rather than quantity, (6) the information worker being considered as an ‘asset’ rather than a ‘cost’. This implies that information workers have to clearly understand the task at hand and be responsible for their own productivity. They have to be responsible for continuous innovation and learning at work. The primary measure of information worker productivity should be quality rather than quantity. Last but not least, information worker productivity requires that the information worker is treated as an asset rather than a cost. Further research

37

on information worker productivity has focused on strengthening some of the aforementioned factors8 or introducing new ones such as motivation (Lord, 2002; Morgeson & Humphrey, 2006) and personal skill training (Kraiger, 2003; Hopp et al., 2009; Singh & Mohanty, 2012). Other studies proposed to focus on clusters of personal, process and organizational individual productivity drivers (Antikainen & Lönnqvist, 2006; Cequea et al., 2011). Nevertheless, new challenges in individual productivity improvement have flourished since the widespread investments in and use of numerous IT tools.

Therefore, information worker productivity has been addressed from different perspectives and a number of models and frameworks to improve individual productivity have been developed. Over the last decades, the following research topics were formulated in managerial disciplines to address individual information worker productivity factors (Table 2.1), including the impact of (i) individual differences, (ii) job/task design, (iii) human resource management practices (e.g. motivation, including performance-based rewards, employee skills, empowerment and participation mechanism in decision-making), and (iv) IT use.

Table 2.1: Summary of current research topics and factors affecting individual productivity in the context of information work

Topic Factor Characteristics Individual differences

Demographic factors

Age; Gender; Marital status; Experience; Education; Professional orientation, etc.

Personality-related factors

Five Factor Model; Creative Personality Scale; Innovation Styles Profile; Myers-Briggs Type Indicator, etc.

Cognitive style factors

Adaptive versus innovative; Intuitive, analytical, and integrated; Analytic, conceptual, directive, and behavioral, etc.

8 See, for example, Sokoya (2000); Mani (2002); Lord (2002); Sonnentag (2003); Horwitz et al. (2003); Love & Edwards (2005); Thompson & Heron (2005); Garg & Rastogi (2006); Sahinidis & Bouris (2008); Rose et al. (2011); Frick & Drucker (2011); Lund et al. (2012); Dewhurst et al. (2013).

38

Table 2.1: Continued from previous page

Topic Factor Characteristics Job/task de-sign

Motivational factors

Autonomy; Task identity; Skill variety; Feedback from the job; Significance; Task variety; Job complexity; Specialization; Problem solving; Information processing

Social factors

Dealing with others; Friendship opportunities; Social support

Work context factors

Physical demands; Work conditions; Ergonomics

Human resource manage-ment practices

Motivation, including perfor-mance-based re-wards

Practices that elicit high motivation, including: • secured, enjoyable and challenging job; • adequate pay; • ability to perform the job; • feedback on performance; • recognition from colleagues; • goal orientation; • freedom; • peer and achievement recognition; • exposure to smart colleagues; • an opportunity for self-advancement,

etc. Employee skills Practices aimed to offer employees opportu-

nities to obtain task-related skills and im-prove personal development

Empowerment and participation in de-cision-making

Practices aimed at enabling employee influence on the decision-making process by empowering employees via greater re-sponsibility and access to resources

39

Table 2.1: Continued from previous page

Topic Factor Characteristics IT use The concept of IT

use

The nature of IT use (the level of IT use so-phistication, different approaches to IT use, and an intention to explore new ways of tech-nology use); Duration of IT use; Adaptive system use (trying new features, feature substitution, feature combination, and feature repurposing); The extent of IT use (frequency, routiniza-tion, infusion); Technology orientation (individual’s pro-pensity and analytical skills for using IT), etc.

IT use and multi-tasking

IT-based multitasking is a concurrent/simul-taneous performing multiple tasks over a cer-tain period of time

IT-enabled commu-nication

Email network structure; Betweenness centrality; Structural diversity of email network; Email network size; Communication channel bandwidth; Access to diverse and non-redundant infor-mation through email networks; Status of individual’s social; Information rich networks

Individual and or-ganizational factors of productive IT use

Individual differences (demographic charac-teristics and personality type); Absorptive capacity; User support and training; Customer pressure; Peer use of technology; Learning and performance orientation; Previous performance; Work infrastructures; Self-efficacy, trust, system quality and infor-mation quality

The above research topics will be discussed in more detail below, to

summarize the factors of information worker productivity which received particular attention, and to demonstrate our present focus on IT use and its effect on individual productivity in an information-intensive environment.

40

2.1.1 Individual differences It has previously been established that individual difference variables, including (i) demographic variables, (ii) personality-related variables, and (iii) cognitive variables can affect individual performance (Zmud, 1979). Agarwal and Prasad (1999) further define these variables in detail. For example, demographic variables are personal characteristics, including age, gender, marital status, experience, education and professional orientation. Personality-related variables refer to cognitive structures maintained by an individual to facilitate adjustments to situations (e.g. locus of control, extroversion/introversion, need for achievement, risk-taking propensity, evaluative defensiveness, and anxiety). Cognitive variables (cognitive styles) represent an individual’s inherent mode of organizing and processing information that is independent of cognitive ability.

Traditionally, research in information systems has focused on whether demographic variables influence the use of IT and if so, how. Individual studies have focused on previous computer experience (Thompson et al., 1994), age and the use of the Internet (White et al., 1999), gender and learning (Arbaugh, 2000). Harrison and Rainer (1992), for example, demonstrated that such individual difference variables as male gender, more experience of computers, younger age, lower math anxiety, and a creative cognitive style were associated with more effective computer use. Sharit et al. (2004) identified that younger participants performed better in computer-based tasks than the older participants. In contrast to traditional studies, Knight and Pearson (2005) came to the conclusion that the role of such demographic variables as age and gender is diminishing due to significant changes in work structure (a higher average workforce age, a greater number of women, and a higher level of education) and the use of IT. Indeed, IT use has reached a high level of maturity and it is expected that any employee knows how to use it.

Different personality types have also been studied in relation to individual information worker performance, including the Five Factor Model (Barrick & Mount, 1991), Creative Personality Scale (Gough, 1979), Innovation Styles Profile (Miller, 1986), Myers-Briggs Type Indicator (Myers et al., 1998). In general, current evidence demonstrates that while personality type instruments are significant attributes of individual behavior and performance outcomes, not all personality dimensions were associated with individual performance (Barrick et al., 2001). For example, Tett et al. (1991) found that among the “Big Five” personality dimensions, including extraversion, emotional stability, agreeableness, openness to experience and conscientiousness, only agreeableness and openness have non-zero correlations with individual performance. Barrick & Mount (1991) demonstrated that only conscientiousness was related to individual performance. Barrick et al. (2001) came to the conclusion that only emotional stability and conscientiousness are

41

associated with individual performance. The same inconclusive results were demonstrated together with other personality types (Smither, 1998; Coffield et al., 2004). The overall conclusion from those studies was that in order to identify meaningful relationships between personality and individual performance, a more theoretically valid measurement of constructs is required.

A number of research attempts have also relied on cognitive styles as core predictors of individual IT-enabled performance (e.g. Benbasat & Taylor, 1978; Couger & Zawacki, 1980; Blaylock & Rees, 1984). This particular focus on cognitive styles can be explained by the fact that in contrast to a material worker, cognitive and creative abilities are required from an employee performing information work. Different typologies of cognitive styles have been developed in psychological literature. Kirton (1976), for example, proposed to examine an adaptors-innovators dimension. Agor (1984) recommended to distinguish between three broad types of cognitive styles: the intuitive, the analytical, and the integrated. Rowe and Mason (1987) proposed four different cognitive styles: analytic (task-oriented), conceptual (creative, intuitive), directive (practical, power-oriented), and behavioral (people-oriented, supportive). Successive studies have indeed demonstrated that cognitive style can be a key determinant of individual behavior and predetermine individual workplace actions (Sadler-Smith & Badger, 1998; Hayes & Allinson, 1998).

Since previous research has shown that an information worker may be understood in terms of his or her cognitive style, where different topologies are proposed, the most dominant and empirically tested is the continuum between the innovative or creative style versus the adaptive or systematic style (Kirton, 1976). Previous studies demonstrate that these styles show some covariance with task types (Baer et al., 2003; Sagiv et al., 2010), incentives (Amabile et al., 1994; Baer et al., 2003), and training activities (Ee, 1998; Ee et al., 2007) in productivity increase. However, it is not clear what the nature of this relation is and whether any other characteristics show covariance and thereby condition workers’ productivity.

To summarize, the aforementioned studies show that individual differences, including demographic variables, personality-related variables, and cognitive variables can affect individual productivity. Yet, cognitive style, particularly continuum from adaptive to innovative, is shown to be a potentially potent variable capable of influencing productivity of an information worker at the individual level. Reviewed studies also show that individual performance of a particular cognitive style can be affected by the tasks to be conducted, incentive and training modes. Hence, the contingency approach is required in order to increase information worker productivity. Despite ongoing debates in personal psychology about individual differences of employees involved in information-intensive work, knowing specific individual characteristics of employees, particularly cognitive style, can help

42

managers successfully match the nature of jobs with personal characteristics to achieve higher productivity results.

2.1.2 Job/task design Job/task design as an approach to structure a person’s job is also shown to be an important driver of individual performance (e.g. Loher et al., 1985; Fried & Ferris, 1987; Parker & Wall, 1998; Sokoya, 2000; Parker et al., 2001; Humphrey et al., 2007). A well-designed job is a job that can bring job satisfaction, involvement in the working process and, therefore, affect individual productivity positively (Bates, 2004). Much of the information worker productivity research is based on job characteristics model (Hackman & Oldham, 1976), which states that core motivational job characteristics, including skill variety (the extent to which an individual has to use various skills to perform a job), task identity (the extent to which an individual can complete a whole and identifiable piece of work), task significance (the extent to which a job impacts colleagues and organization), autonomy (the amount of freedom an individual has in carrying out work), and feedback (the extent to which an individual can see the impact of the work) together can drive individual productivity of information workers.

In addition to the five job/task characteristics identified by Hackman and Oldham (1976), five other job characteristics were discussed in the literature (Humphrey et al., 2007), including task variety (the extent to which an individual conducts various tasks at work), job complexity (the extent to which a job is difficult to perform), specialization (the extent to which a job requires the use of specific knowledge and skills), problem solving (the extent to which a job requires solving non-routine problems), and information processing (the extent to which a job requires monitoring and processing of information). While these factors are shown to harm employee well-being to a certain degree (for example, high job complexity can increase stress and workload) they all promote positive performance outcomes.

Other characteristics of work design are considered equally important in relation to individual productivity. For example, due to increased use of teams in organizations, in addition to motivational characteristics of job design and their impact on individual performance, it was proposed to focus also on social characteristics, including dealing with others, friendship opportunities, and social support (Hackman & Lawler, 1971; Karasek, 1979). Although we recognize the importance of these factors for performance outcomes, we disregard them in this research since team work lies outside the scope of this study. Physical and environmental work context characteristics may also affect job satisfaction and performance outcomes (Campion & Thayer, 1985). Yet, these characteristics are mostly investigated in such fields as ergonomics

43

and biomechanics rather than management (Humphrey et al., 2007) and thus are also disregarded in this research.

There are also other theoretical perspectives that examine the design of work in relation to individual productivity. For example, sociotechnical systems theory (Emery & Trist, 1969) seeks the interdependencies between people, technology, and work environment in order to enhance job satisfaction and improve individual productivity. Techno-structural change models (Galbraith, 1970; Lawrence & Lorsch, 1967) addressed the need for different structural models in job design due to new IT systems’ emergence. Activity theory (Leontiev, 1978; Vygotsky, 1978) states that at the center of work design there exist three levels of work activities, including the motivation of activity, actions, and operations that can affect the behavior of employees and, therefore, their performance. Quality improvement theory (Juran, 1974; Deming, 1986) puts forward that the kernel of individual productivity is a continuous work process improvement. This theory postulates that before redesigning a work process, there is a need to understand what constitutes a process and which activities are included in the process.

The aforementioned theoretical streams have been widely criticized for their inappropriate applicability in new work environments, general nature of principles and discontinuous character in restructuring job design (Kelly, 1992; Torraco, 2005). However, their emergence demonstrates that both information technologies and nature of information work affect job and task design. Moreover, these streams recognize that work process (the manner in which the job and work tasks are accomplished) is an essential factor that may impact individual productivity of information workers.

Information work process is often characterized by a diversity of tasks (Czerwinski et al., 2004; Webster, 2012) and, as Drucker (1999) highlights, the most important challenge for information worker productivity is to clearly understand and define the task at hand. Information workers are involved in value-adding tasks that are related to their information worker role. Besides those tasks, information workers also perform administrative tasks (creating and managing reports, emails, and documentation) and interaction tasks (presentations and providing feedback) that consume an essential amount of working time (Lund et al., 2012). Since information worker tasks are too heterogeneous and complex to motivate creating a single model, it is recommended to concentrate on value-adding tasks to study individual productivity of such category of workers (Hammer et al., 2004). In this research, we follow this proposition, yet acknowledge that information work is characterized by a diversity of working tasks.

In general, an information work process is based on problem solving and information processing (Reinhardt et al., 2011). It is recognized that there are different kinds of human information processing, starting from fully structured processing, where the problem is well-defined with a limited number of alternative solutions and ending with fully unstructured processing,

44

where the problem is ill-defined with numerous alternative solutions (Simon & Newell, 1964). This distinction shows a clear impact on individual performance (MacCormack et al., 2001). In contrast to the unstructured/flexible process, inputs and outputs in a more structured process are typically carefully specified and its activities have a pre-defined order (Abdolmohammadi & Wright, 1987). When we refer to the cognitive style continuum from adaptive to innovative (Kirton, 1976) as described above, it is shown that adaptors demonstrate better performance when they are involved in structured, pre-defined tasks, while innovators perform better in complex tasks that are loose in structure (Baer et al., 2003). As we have previously proposed that information work may require different types of personalities, we may also expect that a particular match between structural task diversity and cognitive style may significantly affect individual productivity.

2.1.3 Human resource management practices Much of the research on individual productivity has focused on the importance of human resource management practices that affect individual behavior and drive individual productivity. Currently, a consensus is emerging that high-performance human resource management practices are focused around three areas: (i) motivation, including performance-based rewards, (ii) employee skills, (iii) empowerment and participation in decision-making (Snape & Redman, 2010). These practices are also shown to act synergistically in relation to performance (Alfes et al., 2013) and constitute a high-performance human resource management program (Snape & Redman, 2010). These practices are reviewed in more detail below.

• Motivation, including performance-based rewards

Despite its intangible nature, the concept of motivation is a fundamental

factor for both material and information workers’ productivity (Frey & Osterloh, 2002). Motivation can be characterized by different dimensions, including a secure, enjoyable and challenging job, ability to perform the job, feedback on performance, recognition from colleagues, and adequate pay (Pinder, 2011). A number of content theories, including McClelland’s need theory (McClelland, 1967), Herzberg’s two-factor theory (Herzberg et al., 1959), and self-determination theory (Deci, 1975) have been developed to establish underlying factors of motivation that drive individual performance. In contrast, process theories such as expectancy theory (Vroom, 1964), equity theory (Adams, 1963), goal-setting theory (Latham & Locke, 1979) investigated the processes which underlay work motivation and its impact on individual performance. Both types of theories provide the foundation for a

45

significant amount of current research and managerial practice (Lord, 2002; Mani, 2002; Kubo & Saka, 2002; Morgeson & Humphrey, 2006).

Incentives and rewards are considered as the prime factors that can affect motivation and workplace performance (Danish & Usman, 2010). In general, it is shown that employees prefer to have high pay levels, individually oriented pay systems, job-based pay, fixed rather than variable pay and flexible benefits (Barber & Bretz, 2000). Petroni and Colacino (2008) demonstrate that engineers consider the compensation level important for establishing status and experiencing recognition. Yet, financial rewards are not always a panacea for better performance. Particularly, non-financial rewards, including goal orientation, task complexity, freedom, supervisory styles and perceived evaluation are shown as important for enhancing individual performance (Chesbrough, 2003; Horwitz et al., 2003; Thompson & Heron, 2005). In particular, it is established that information workers prefer peer recognition, exposure to smart colleagues, an opportunity for self-advancement and task complexity (Hopp et al., 2009). Moreover, it is also shown that non-monetary rewards such as achievement recognition, small non-cash awards, and photos in public location are strong predictors of intrinsic motivation and individual performance among information workers (Markova & Ford, 2011). All this implicitly implies that individuals have different preferences in incentives, and thus, these preferences can affect their motivation and performance.

While motivation is a hot topic in relation to performance outcomes, the effect of individual differences in incentive preferences (particularly cognitive abilities) and incentives on performance outcomes has often been ignored in existing research (Kossowska et al., 2010). However, previous evidence demonstrates that personality can be a primary predictor of motivation and work performance (Day et al., 2002; Tett & Burnett, 2003). For example, Tett & Burnett (2003) demonstrated that individuals feel more motivated in work settings that allow them to express their traits that are positively valued by others. Earlier research has also shown that the cognitive abilities of individuals together with motivation represent two basic determinants of work performance (Dunnette, 1976; Hunter, 1986). Moreover, earlier research stressed the importance of interaction between cognitive ability and motivation in relation to work performance (Pinder, 1984). Yet, at the present time no unified approach exists for studying simultaneous effects of individual differences in cognitive abilities and motivation on work performance. When we refer to the previously proposed distinction between adaptive/innovative cognitive style, we can presume that individuals may differ in their preferences in incentives (Amabile, 1996), and thus, when matched correctly this can affect their motivation and performance outcomes in a positive manner.

46

• Employee skills development Since information-intensive work is characterized by rapid changes,

besides individual characteristics, job design, and incentives, skill development has an important role in achieving productivity gains (Drucker, 1999; Kessels, 2001; Aguinis & Kraiger, 2009). Empirical studies have demonstrated that training practices can have a positive impact on individuals’ motivation, organizational commitment, job satisfaction, work performance and individual productivity (Sahinidis & Bouris, 2008; Rose et al., 2011). Yet, researchers and practitioners are concerned with identifying optimal training strategies, including tools, methods, and content (Kraiger, 2003; Singh & Mohanty, 2012).

Generally, research in training practices investigates either learning principles or specific teaching approaches and methods (Kraiger, 2003). For example, theories arising from cognitive and industrial psychology have become a basis for research into human learning (Womer, 1984; Sutton & Barto, 1998; Howell & Cooke, 1989; Lord & Maher, 1991). Those studies mostly addressed the need for primary understanding of how individuals acquire, organize and master skills before focusing on training program development. Existing literature also investigates different forms of learning in information-intensive work. For example, Eraut (2000) claims that learning types such as formal, informal and incidental learning are equally important for efficient performance of tasks. Ryu et al. (2005) studied three types of learning processes, including learning-by-investment, learning-from-others, and learning-by-doing. Those three types of learning processes are found to be necessary for achieving an optimal level of knowledge to perform a task. Both learning-before-doing (push) and learning-by-doing (pull) are also found to be appropriate in an information-intensive environment (Pisano, 1994, 1996). Although there is a long tradition in identifying the impact of learning principles on individual performance, the implication of learning for developing training programs is still not fully explored by researchers and practitioners (Aguinis & Kraiger, 2009).

Therefore, skills training is demonstrated as an important factor that can positively influence individual performance (Satterfield & Hughes, 2007). Previous studies, however, demonstrate that despite the quality of training, motivation to learn and mastery orientation are shown to be important in relation to performance outcome (Gagne & Medsker, 1996; Schank & Joseph, 1998). For example, the difference in learning outcomes between individuals who were exposed to the same training program were explained by differences in learning style orientation as a sub-characteristic of cognitive style (Hayes & Allinson, 1997; Sadler-Smith & Smith, 2004). These findings lead to a potentially existent interrelation between particular learning style and training strategies that may affect individual performance outcomes. When we refer to adaptive/cognitive style we can expect that when a particular cognitive style

47

is aligned with particular work related training practices, this can lead to more effective learning and thus positively affect individual productivity. Yet, little knowledge exists on how to organize effective training programs with a focus on cognitive style to drive individual performance (Armstrong et al., 2012). • Empowerment and participation in decision-making

Previously, it has been established that to manage high performance environment, besides motivation-enhancing and skills development; empowerment-enhancing (using practices enabling employees influence on decision-making) is also required (Delery et al., 1997). Organizational behavior and management literature demonstrate that high levels of worker control over decision-making together with employee participation and involvement in decision-making are all associated with a higher level of job satisfaction, well-being, intrinsic motivation, and performance outcomes (Conger & Kanungo, 1988; Laschinger & Wong, 1999; Ozaralli, 2003; Chen et al., 2007). However, in a competitive and turbulent environment where information workers are prevalent (Wolff, 2005), delegation of decision rights brought about more complex uncertainties for both researchers and practitioners (Bloom & Van Reenen, 2011).

On the one hand, broadly applied empowerment initiatives among employees may limit managers’ ability to predict and standardize work processes, and therefore negatively affect employees’ performance in particular and organizational performance in general (Zabojnik, 2002). On the other hand, while some individuals prefer more empowerment in decision-making with more autonomy, others may experience such environments as stressful and cognitively demanding (Wilkinson, 1998). In response to these challenges, it was recommended to focus on more selective empowerment interventions, by delivering decision-making to specific groups of employees, while placing fewer demands on others (Sisselman & Whitt, 2007). For example, it has previously been identified that empowerment requires a particular cognitive style that is characterized by self-awareness, self-control, self-confidence and self-efficacy (Velthouse, 1990). The behavior of such individuals is usually characterized by innovation, independence, responsibility, involvement, persistence and commitment (ibid.). However, while cognitive style is often mentioned as a potential determinant of decision-making preferences, little research has been done into this fit (Biron & Bamberger, 2010). Current companies also often ignore the various cognitive preferences of their employees and assume that all employees perceive decision-making in a similar way (Auh & Menguc, 2007).

To manage the delicate balance in giving employees decision-making authority without at the same time losing this authority over employees, it was proposed that empowerment has to be considered not as an absolute term but rather as a degree of delegation of decision rights (Laschinger & Wong, 1999).

48

It was also highlighted that different decision areas require different degrees of empowerment, depending on job content and context. Psychology literature also demonstrated that some individual characteristics are more appropriate for empowerment than others (Velthouse, 1990; Sisselman & Whitt, 2007). Moreover, these individuals experience greater intrinsic motivation when they are exposed to greater autonomy in decision-making (Amabile, 1996). While different individuals prefer different empowerment in decision-making, managers can match these preferences with a certain degree of decision-making in order to enhance their performance. All this implies that there is a contingency fit between empowerment in decision-making, specific individual characteristics, job content and context that can positively affect individual performance, yet this fit is not sufficiently researched.

In establishing decision-making authority, the role of leadership and organizational culture is highly important (Malone, 1997). Good supervisors are usually aware of areas that are potentially fruitful for empowerment and can help those employees who are exposed to decision-making by instructing them and offering advice. These supervisors can also be sensitive to employees’ preferences in decision-making authority. Organizational culture also plays an important role in delivering decision-making authority to meet environmental challenges (Appelbaum et al., 1999). For example, one of the dimensions of organizational culture is to be ensured that power distance is established in such a way that employees feel comfortable in interactions across hierarchical levels (Hofstede, 2001). Therefore, both leadership and organizational culture can affect individual performance outcomes in the event of careful delegation of decision-making authority to employees.

Given the above, human resource management practices are shown to be important factors that affect individual performance. Previous studies also argue that complementarity sets of human resource management practices, including motivation, skills development and empowerment rather than individual practices can lead to higher levels of performance outcomes at both individual and organizational levels (Takeuchi, 2009; Alfes et al., 2013). However, most of the reviewed studies assume that all individuals perceive human resource management practices in a similar way (Biron & Bamberger, 2010). Yet, there is evidence that people who performed satisfactorily in one job setting can perform differently when they transfer to a different job setting (Sadler-Smith & Badger, 1998). This implies that individual productivity may also depend on a particular match between the individual’s characteristics and specific clusters of human resource management practices. This premise is further investigated in relation to the research question of this research in the next chapter. In the next sub-section, we review another research inquiry that investigates the impact of IT use on individual productivity and is addressed in this research.

49

2.1.4 IT use Technology is an important resource in both material and information work. In contrast to material work where technology is embedded in the equipment or used to support established tasks, information technology is used in information work in a profound manner and can help workers in decision-making and the generation of creative solutions (Hopp et al., 2009). This presents an essential challenge for both researchers and practitioners.

Earlier information systems research made attempts to identify the relationships between IT use and IT-enabled performance by focusing on such factors as IT system design, individual cognitive style and task characteristics (Dickson et al., 1977; Benbasat & Taylor, 1978; Benbasat & Dexter, 1982; Blaylock & Rees, 1984). For example, it was established that IT system characteristics (data representation and output formats) and/or individual attributes of decision makers (cognitive style, aptitude, and past experience) can affect decision maker performance. (Dickson et al., 1977). It was also proposed that specific cognitive styles should be included in research studying the effect of IT use on individual performance (Benbasat & Taylor, 1978; Blaylock & Rees, 1984). Moreover, it was concluded that the match between task environment (multiperiod inventory/production decision-making game) and cognitive style can also affect individual performance (Benbasat & Dexter, 1982). While these studies shed light on the relationships between factors affecting IT-enabled productivity, the focus on a cognitive style was criticised as an unsatisfactory basis for information systems research (Zmud, 1978; Huber, 1983). For example, Huber (1983) concluded that since there is little empirical evidence to demonstrate that the inclusion of the cognitive style improves information system design, the use of this factor is misleading. At the same time, cognitive style was supported by other researchers as a promising factor of effective IT use (Robey, 1983; Ramaprasad, 1987). Although the aforementioned studies showed that there is a potential for the joint impact of individual attributes, information system characteristics, task environment on individual performance, the results were mixed and inconclusive due to the lack of underlying theories and explorative empirical studies.

Over the last decades, a number of studies have made attempts to demonstrate how IT use can improve individual productivity of employees in information-intensive occupations by applying different theoretical approaches. Yet, since the concept of IT use as such is complex and multi-dimensional (Grgecic & Rozenkranz, 2010) a number of studies have addressed the need to define and measure the concept of IT use to demonstrate its effect on individual productivity (Jain & Kanungo, 2005, 2013; Sundaram et al., 2007; Sun & Fricke, 2009). Another stream of studies investigated the human actor position in an email network and demonstrated that individual

50

productivity can be improved when an individual is located in the center of a network with diverse contacts (Aral et al., 2006; Chung & Hossain, 2009; Wu et al., 2009; Aral & Van Alstyne, 2011). A number of studies demonstrated that system quality, information quality, service quality, intention to use a system, and user satisfaction can positively affect individual and organizational productivity (D’Ambra & Rice, 2001; McGill et al., 2003; Gyampah & Salam, 2004; Iivari, 2005; Halawi et al., 2007). In the next section, a more detailed analysis of studies on IT-enabled productivity at the individual level will provide the foundation for this research.

To summarize, the review of studies on information workers’ productivity demonstrates that individual productivity is conditioned by a number of different factors, including specific personal characteristics of employees, job/task design, human resource management activities (e.g. incentives, skills development, and empowerment of decision-making authority) as well as IT use. Literature review also demonstrates that most studies addressing individual productivity focus on one or a few factors only. However, these factors on their own seldom show a definitive tendency of influencing productivity directly but rather show a contingent nature or variability depending on other factors or the situation at hand. Bearing this in mind, below we review studies that investigate the relationships between IT use and information worker productivity at individual level in more detail.

2.2 Individual IT-enabled productivity: non-complementary approach

Although productivity of an information worker at individual level is a challenging and underexplored area of research (Aral et al., 2012a), a number of studies have made attempts to uncover the relationship between IT use and individual productivity from different perspectives (e.g. Jain & Kanungo, 2005; Sundaram et al., 2007; Deng et al., 2008; Aral et al., 2012a). In general, studies on individual IT-enabled productivity of information workers differ in their respective approaches to the nature of the problem. Yet, some core topics have emerged in research publications over the last decades. Most of the studies reviewed below address (i) conceptualization of IT use and its impact on individual productivity, (ii) IT use, multitasking, and individual productivity, (iii) IT-enabled communication patterns and individual productivity, and (iv) the impact of individual and organizational factors on IT-enabled productivity. This section is based on the systematic review of papers from several leading management sciences and economics journals,

51

which analyze how IT can affect individual productivity of information workers.

2.2.1 Conceptualization of IT use and individual productivity At the present time, one of the prevalent views to demonstrate and explain the relationship between IT use and individual productivity is based on the re-conceptualization of IT use construct as such. For example, Boudreau (2003) noticed that previously accepted conceptualizations of IT use in terms of time, reliance and diversity are misleading in understanding successful IT use, since they limit the variety of ways of technology use. Boudreau (2003) suggests that in order to demonstrate productivity gains, a new concept such as quality of IT use (the ability to correctly exploit software in particular circumstances) is required. Moreover, Burton-Jones and Straub (2006) claim that inappropriate measuring of IT use may lead to opposite conclusions in empirical research. To support this argument, Sun and Fricke (2009), by developing a construct of “adaptive system use”, came to the conclusion that only a richer conceptualization of how IT is used can explain its significant impact on individual performance, including task productivity, management control, and task innovation. Recently, a number of studies made some attempts to investigate the relationship between IT use, individual productivity, and performance outcomes by developing different concepts of IT use quality in contrast to a narrow conceptualization of IT use as such. These studies are reviewed below in greater detail.

In order to investigate individual IT-enabled productivity of information workers, Jain and Kanungo (2005) developed a multi-construct concept of the nature of IT use. These researchers explain the need to develop this concept by the variety of ways individuals use IT. For example, they claim that while some individuals prefer to exploit the functionality of the system, other individuals tend to limit their interaction with the system. This concept of the nature of IT use is defined as “…the degree to which a person differs from others in the way he or she uses a particular information system” (p. 115). The concept consisted of three dimensions: (i) the level of IT use sophistication, (ii) different approaches to IT use, and (iii) an intention to explore new ways of technology use. The relationship between the use of IT (the number of hours respondents used an IT system), the nature of IT use and IT-enabled productivity has been tested based on survey data collection. In general, the results indicated that both IT use (the use of emails and spreadsheets) and the nature of IT use were positively related to IT-enabled productivity among a wide range of information workers, including middle management, supervisory and technical staff. The results imply that the variety of IT systems as well as duration are both important for individual productivity. Yet, the study is not without its inherent limitations, such as the

52

use of only two applications (emails and spreadsheets) and self-reported measures.

The study by Jain and Kanungo (2005) has been further developed by introducing a task-technology fit model and tested among 483 individuals from six organizations belonging to the telecommunications, defense, government and public sectors, financial and banking, aerospace and the hospitality sectors. (Jain & Kanungo, 2013). The results demonstrate that empirically the developed model explains nearly 60% of the variance in performance impacts of IT use, meaning that the developed model provides a fuller understanding of the impact of IT use on individual performance. While both studies are premises towards developing a process theory-oriented approach for IT-enabled productivity and provide a fuller understanding of the relationships between IT use and individual performance, some issues require further discussion. For example, in both studies, only a limited number of technological applications such as emails and spreadsheets were used to study the impact of IT use on individual productivity. Yet, an information-intensive environment is characterized by a broader variety of IT tools. For example, besides emails and spreadsheets, information-intensive workers currently use different work-related IT applications, smart technologies such as tablets and iPhones, Internet-based applications, etc. Moreover, self-reported measures of both IT use construct and productivity are subject to a potential bias (Szjana, 1996).

Another conceptualization of IT use such as adaptive system use has been used to study the relationships between IT use and task productivity of administrative assistants (Sun & Fricke, 2009). The concept of adaptive system use is multidimensional and includes such dimensions as (i) trying new features of an IT system (using new features), (ii) substitution feature (replacing a current with a new feature), (iii) combination feature (using two or more features), and (iv) repurposing feature (using a feature in a way that was not planned by the developer). The results demonstrate that adaptive system use accounts for a significant part of the impact of IT use on the job performance of administrative assistants, meaning that using an IT system adaptively can have a strong impact on job performance. Hou (2012) adds that end-user computing satisfaction also leads to increased use of an IT system and individual performance. This implies that the role of the individual is not less important than the use of an IT system in relation to performance outcomes.

An essentially different approach studying IT-enabled productivity in a post-adoption context was suggested by Sundaram et al. (2007) who applied the theory of planned behavior to study salesperson performance. The researchers proposed that in order to understand how IT use affects individual performance, it is necessary to study the relationships between the extent of IT use (the level at which an individual has incorporated technology within the work structure) and individual performance output. Three dimensions of

53

the extent of IT use have been proposed such as (i) frequency (the extent to which a person uses IT), (ii) routinization (incorporation of IT use in routine work pattern), and (iii) infusion of IT use (a stage when an individual maximizes the potential of IT). In general, the study presents empirical evidence that in contrast to frequency and routinization, only infusion makes a positive impact on both IT-enabled salesperson performance and IT-enabled administrative performance.

Another study in the same empirical settings reveals that technology orientation (a salesperson’s propensity and analytical skills for using IT) is one of the most important factors affecting individual performance (Hunter & Perreault, 2006). In more detail, it was shown that technology orientation has a direct impact on salesperson efficiency and, through the effective use of information and smart selling behavior, affects performance. Mathieu et al. (2007) identified that the use of the new sales technology has a positive impact on performance changes in the same empirical setting. Yet, by describing the main shortcomings of the study, the researchers claim that the obtained results are tentative and field experiments are required to identify the relationships in more detail. Moreover, they recognize that not all technological interventions can be effective and an account of salesperson training, the availability of assistance and adequate transition time is required to study the impact of new technology on individual performance. It is also noticed that the incorporation of individual competencies and personalities offers potentially fruitful directions for further research. Therefore, while only the relationship between IT use and individual performance was studied, it is recognized that some complementary factors can potentially be important for IT-enabled performance.

In general, different approaches to the conceptualization of actual IT use grounded on attitudes and behavior theories enrich our understanding and provide empirical evidence of productive IT use in post-adoption contexts. Still, these studies are inherent to potential limitations that could affect the obtained results. First, the studies described above are mostly based on studying one particular type of IT in specific working contexts at one point in time, that limits generalizability of the obtained results. Second, quite a different operationalization of IT use construct was proposed, rendering the results difficult to compare. Third, mostly self-reported measures were used to operationalize both independent and dependent variables. It is well-known that self-reported measures may be biased (Podsakoff et al., 2003). Indeed, we lack studies that could provide objective, or IT user independent, data of IT use and individual productivity. Eventually, while the reviewed studies mostly focused on IT use measuring and its relation to individual performance, some of them do recognize that other individual and organizational factors can potentially affect the established relationships. Yet, while recognizing the importance of additional factors for IT-enabled productivity, little is proposed about what particular factors have to be taken into account and how to align

54

them to increase individual IT-enabled productivity. Hence, more research is required to enhance our understanding of why individuals perform differently when the same IT is used.

2.2.2 IT use, multitasking, and individual productivity While one stream of studies of IT-enabled productivity encourages further research towards a better conceptualization of IT use; over the last decade, a concept of IT-based multitasking (computer-based multitasking, or technological multitasking) and its impact on individual performance has received noticeable attention from the research community (Czerwinski et al., 2004; Bell et al., 2005; Aral et al., 2006, 2012a; Appelbaum et al., 2008; Buser & Peter, 2012). Increased use of IT tools has indeed facilitated multitasking – concurrent performance of multiple tasks over a certain period of time (Benbunan-Fich et al., 2011). The view that IT-based multitasking can increase productivity of information workers when managed correctly (Wasson, 2007) is very much in line with common sense. However, this increased opportunity for multitasking has brought with it some challenges to understanding its actual impact on individual productivity. This sub-section presents elaboration on potential reasons for the contradictory results on the impact of IT-enabled multitasking on individual performance.

By conducting a diary study, Czerwinski et al. (2004) showed that a number of switches among tasks during the working week led to a significant time leakage in task performance for information workers and negatively affected individual performance. Aral et al. (2006) demonstrated an inverted-U relation between multitasking and individual productivity, meaning that multitasking improves productivity up to a certain point, yet after this point productivity decreases. In a controlled experiment, Adler and Benbunan-Fich (2012) identified that IT-enabled multitasking was associated with a diminishing marginal return to efficiency. However, decreasing line was associated with performance effectiveness (accuracy). As a conclusion, the researchers state that the nature of the relationship between multitasking and performance depends on the performance metrics used. Alternatively, Madjar and Shalley (2008) argue that multitasking provides better outcomes by prompting a more efficient use of time and allowing ideas to mature.

One more negative effect of IT-based multitasking as a communication interruption has been studied widely over the last years. In the past, it was established that information workers switch among tasks quite a significant number of times during the working process, which increases the time to return to the previously executed task (Czerwinski et al., 2004; Iqbal & Horvitz, 2007). Indeed, interruption frequency and order of task complexity negatively affect cognitive load and task performance (Basoglu et al., 2009). For example, Sykes (2011) argues that with the implementation and

55

dissemination of computer tools, the number of IT-based communication interruptions is growing. In fact, Marulanda-Carter and Jackson (2012) claim that any task interrupted by email takes one-third longer to complete than task without interruption. This occurs due to the disturbance of employee’s concentration and email addiction problem.

By studying the available literature on behavior, managerial, and technological aspects of multitasking, Appelbaum et al. (2008) pointed out the existence of the IT-based multitasking paradox, meaning that although overall multitasking has a negative effect on individual performance, it did not restrict organizational productivity growth. The effect of IT-based multitasking, especially at individual level, becomes even more intriguing since, despite the growing number of research in this area, it is still unclear what the effect of IT-based multitasking on individual performance is. In relation to this, Jez (2011, p. 161) points out that: “…while multitasking might include switching between several tasks, it also means being interrupted”. Probably, one of the reasons for the contradictory results in current studies can be explained by a dual effect (switching to another task after having completed the foregoing versus interruption while conducting a task) of multitasking practices.

Another challenge can be in the definition of multitasking as such. While a number of studies analyzed the impact of multitasking on individual productivity, there is still no consensus about the exact meaning of multitasking. For example, Cherwinski et al. (2004) and Bell et al. (2005) define technological multitasking as rapid task switching activities while performing a job in close interaction with IT. As an example of multitasking behavior, they suggest checking emails or using instant messaging during a meeting. Contrary to this, Kenyon and Lyons (2007) refer to multitasking as a simultaneous performance of two or more activities during a particular time period. According to Aral et al. (2006), multitasking is the act of taking on parallel projects simultaneously. Bannister and Remenyi (2009) divide multitasking on conscious (performing one task at a time) and subconscious (performing several tasks simultaneously). Benbunan-Fich et al. (2011) define computer-based multitasking as a performance of more than one computer-based task concurrently. Moreover, the researchers recommend using tasks, technology and timing issues in order to measure the construct of multitasking. Thus, the various studies of multitasking assume varying definitions of what the concept of multitasking refers to. This implies that these studies measure different kinds of multitasking and, consequently, the results of these studies are not fully comparable. Hence, the concept of multitasking creates obstacles to identifying its potential benefits and by itself requires extensive research.

Therefore, current studies demonstrate that the existence of the multitasking paradox can be partly explained by the inconsistent conceptualization of multitasking practice as such. Moreover, the research findings by Bannister and Remenyi (2009) lend support to the claim that common assumptions about multitasking for information workers are not yet

56

well-developed due to a dimensional matter of the research problem. On the one hand, common sense suggests that if an information worker is given IT tools to quickly and effectively switch his or her work between different kinds of task, jobs and projects, and in some cases, conduct simultaneously several tasks – then it would be a given contributor to productivity increase of the individual. Also, some empirical studies claim that productivity can be increased due to a worker’s multitasking when using an IT system (Aral et al., 2006; Adler & Benbunan-Fich, 2012). On the other hand, the inclusiveness of these empirical studies and also their incoherence with regard to the notion of what is really multitasking, shows that there is much research to be conducted in order to measure and identify the actual nature of the relationship between multitasking and performance of the individual. Therefore, one way forward for this research is to assume that multitasking is inherent in the operational set up of a worker, whether it is a structured or flexible way of working, probably in different ways. If that is the case of being inherent then there is no real need to decompose such operational processes or ways of working to see exactly how much multitasking is conducted. Rather, multitasking may be black-boxed with the kind of work processes utilized.

2.2.3 IT-enabled communication patterns and individual productivity

The emergence of powerful communication devices and continuous connection to the Internet has enabled communication networks to grow greater and faster (Castells, 2011). This fact can explain an increasing focus among the research community on the use of electronic communication devices and their effect on individual performance. Social network theoretical premises (Freeman, 1979; Burt, 1992) have become a framework for a new scientific direction on individual productivity in an information-intensive environment A number of pioneering studies, reviewed below, demonstrate the applicability of social network analysis to uncover the mechanism of individual productivity in terms of the communication aspect of information work.

The first study which presented how the communication aspect of IT use affects the productivity of employees at the individual level was conducted by Aral et al. (2006). The productivity of information workers, namely recruiters, was measured at individual level mainly through the analysis of electronic communication networks and IT use. The results of this study demonstrated that IT use (the use of an Executive Search System and external proprietary databases) positively affected recruiters’ productivity. The results also demonstrated that the use of databases was positively associated with multitasking and generated revenue for the firm per unit time. However, more

57

multitasking was associated with increased project output, but with diminishing marginal returns which led to a trade-off between workload and efficiency. The main conclusion of this study suggests that workers who were at the center of their organization’s email flow tended to be more productive on average than their less well-connected colleagues. Undoubtedly, this kind of research is highly innovative, though provoking, and sheds light on the process of IT value creation and information worker productivity. Nevertheless, some essential limitations are presented below, thus further advancing our knowledge of information worker productivity.

In contrast to the existing theoretical arguments on the relationship between social network parameters and economic advantages, Aral and Van Alstyne (2007) argued on the scarcity of theoretical explanations of the link between network structures, the nature of distributed information and individual performance. This theoretical aspect, starting from Burt’s, at some point, contradictory assumption about the positive relationship between structurally diverse networks and individual performance which is achieved through better access to new and non-redundant information and, at the same time, on the inefficiency of structural redundancy in terms of inability to receive new information from additional contacts (Burt, 1992, 2004). Indeed, the results present some evidence of non-linearity in the relationship between network structure, information content, and performance advantages. Network size and diversity were positively correlated with the total amount of novel and diverse information. When network sizes increased, the marginal increase in information diversity became less pronounced. As in previous studies, network diversity was positively correlated with individual performance, yet diminishing marginal productivity returns to novel information were noticed. In addition, there was little effect of human capital attribute variables on access to novel information.

Another study by Aral and Van Alstyne (2011) also used classical social network arguments regarding the ability to receive more novel and diverse information from structurally diverse networks (rich in weak ties and structural holes). Contrary to these settled statements, the hypothesis of the study stated that stronger ties can provide greater novelty over time in particular circumstances, i.e. even if strong ties in cohesive networks provide less novel information, the interaction between actors will be more frequent and the bandwidth higher. According to the study, two factors affected the rate at which novel information is received. These factors are structural diversity and channel bandwidth. Moreover, both network diversity and channel bandwidth predicted access to diverse and non-redundant information. Yet, in the environment of redundant and more rapidly changed information with a large number of topics, bandwidth was a better predictor of access to novel information while the opposite was true for network diversity. In general, the obtained findings demonstrate that access to diverse and non-redundant information was positively associated with individual productivity.

58

Through monitoring of the use of emails and instant messaging, Wu et al. (2009) reached a conclusion that human capital and status of an individual’s social contacts is no less important a predictor of productivity than social network topology. For example, it was hypothesized that strong ties to managers highly affect employees’ productivity. The cumulative human capital inside one’s social contacts was suggested to be one of the core parameters determining individual productivity. The results demonstrated that social network parameters of diversity and centrality were positively associated with performance at individual and project levels. However, it was suggested to explore the content of the network in terms of ties to managers and powerful individuals in addition to network topologies. Yet, in contrast to the study by Aral and Van Alstyne (2007), network size was not associated with individual performance. Finally, individuals’ performance was more determined by social network characteristics at project level than at individual level.

While the characteristics of electronic communication network topologies and their impact on individual performance have been studied widely, little attention has been paid to the characteristics of communication channels. This aspect underpinned the study by Wu et al. (2008) on the link between the structure of communication networks and individual productivity in information-intensive work settings. Both social network theories and the information richness theory, which exhibits differences in communication channel capacity for immediate feedback based on the ability of information to change understanding within a time interval were employed in the study. The bridge between both theoretical aspects was raised to improve understanding of which types of social network structure are more productive in terms of information transfer complexity across different communication media and, thus, justify individuals’ media choices for different types of professional tasks. In general, the hypothesis of the study that various communication modes lead to differences in the type of information being transferred has been supported by empirical investigation. Moreover, communication network with high cohesion was associated with higher performance due to the advantages of using face-to-face communication to transfer complex information, and is especially effective in solving complex tasks.

Theoretical arguments on the relationship between information-rich networks (rich in structural holes and low in cohesion) and improved performance also set up the study by Wu (2012). In comparison to previous research in which the consequences of different social relationships on individual performance had been explored, the question of how information-rich networks can be created formed the basis of the study. In general, the study investigated social network theories with a focus on ties contents and showed that the benefits of information from information-rich networks can have both instrumental and expressive properties, and in particular, an effect

59

on worker productivity and job security. According to the obtained results, information diversity was positively correlated to billable revenue and social communication was associated with job security.

Due to measurement difficulties of information work, most of the presented studies relied on the triangulation methodology to collect the necessary data. As shown in the reviewed studies, a multi-method approach was necessary to fill in gaps involving aspects of communication and performance metrics in information work that are difficult to measure directly. Methodologies such as questionnaire surveys, interviews, observations, and self-reports have been predominant. For example, Aral et al. (2006), Aral and Van Alstyne (2007, 2011) conducted interviews to understand the production processes of executive recruiters and qualitative survey to identify factors of production function, as well as to take into account such human capital factors as age, experience, education, IT use skills and information-seeking behavior. Hence, these studies collected unique data and carried out careful analysis to demonstrate the impact of IT-enabled communication on individual productivity of information workers.

In general, the reviewed studies complement each other and mostly address social network properties such as network positions, ties strength, the presence of structural holes (Aral et al., 2006), information content and bandwidth of structural holes (Aral & Van Alstyne, 2007, 2011), capabilities of communication media to process rich information (Wu et al., 2008), human capital attributes and status in social contacts (Wu et al., 2009) and ties characteristics (Wu, 2012) in relation to individual information worker productivity. In particular, they demonstrate that employees who are at the center of their organization’s email network tend to be more productive on average than their less well-connected colleagues. Moreover, network diversity and centrality are also positively associated with individual performance. Status in social contacts is also shown to positively affect individual performance. Therefore, the reviewed studies present unique evidence on (i) what constitutes an effective communication network both from email and verbal communication perspectives, (ii) how the relationships between actors create performance benefits, and (iii) how position and status in the network can affect individual performance.

Several managerial implications of the obtained results can be derived from the reviewed studies. For example, there is a good comparison of social network analysis as “X-rays” for the assessing of interrelationships between actors (Cross et al., 2003) through which managers can identify specific patterns of collaboration between employees to improve their performance. At the same time, knowing which actors are important in facilitating information flows and which ones are peripheral actors, allows managers to optimize communication structure, to organize effective collaboration, to enhance economic outcomes and react in time when a so-called broker leaves the company. Actor position and status in the organization make an impact on

60

the accessibility of this actor which can slow down the work of the whole organization when the important decisions must be taken immediately. Moreover, the results on the physical proximity and verbal communication network (Wu et al., 2008) may have some implications for co-allocation of individuals in the company in order to increase their productivity.

There is no doubt that the transfer of information is one of the most important activities of information workers. However, the whole process of information processing and treatment can be conducted in different ways. Consequently, new questions arise. Is it enough to consider only one information processing activity in order to reduce uncertainty, determine and justify productivity of an individual in information-intensive environments? In which way are other phases of information processing related to the productivity of this type of workers? For example, according to Simon’s information-processing theory (Simon & Newell, 1964), digital computers execute the same sequence of information processes as human minds, including storage (memorizing), transfer (retrieving or communication), transformation (codification or computation), and generation (creation) of information. Moreover, most of the reviewed studies are based on a specific production process. Thus, how may the obtained results be generalizable for all information-intensive environments? All this suggests that focus should not only be on information transfer/logistics when measuring IT-enabled productivity, but also on other functions (transformation, generation, storage) which reflect the whole complexity of information work. Moreover, although many information-intensive occupations require frequent interaction with colleagues and work in teams, there are also jobs, such as software construction, analytics, web design and graphics that do not require extensive collaboration or communication. Thus, application of social network analysis becomes inappropriate when studying IT-enabled productivity and, hence calls for the need to use other theoretical approaches.

2.2.4 The impact of individual and organizational factors on IT-enabled productivity

The fourth stream of studies, unlike the previous ones, addresses the impact of IT use and individual performance from a broader perspective by analyzing other, no less important factors through which IT can affect individual productivity. For example, Kvassov (2004) investigates the impact of IT use on information worker productivity through temporal dimensions and time personality in terms of the degree of poly-/monochronicity (preferences in performance one/many tasks at a time). It has conclusively been shown that personality type is one of the major mediators between IT use and individual task performance. Moreover, it has been indicated that IT use makes a

61

significant impact on all temporal dimensions such as schedules and deadlines, coordination, autonomy, a pace of work and allocation of time. In general, the study confirmed that IT use affects productivity by reducing the time spent on activities and increasing the number of tasks performed per unit of time.

In another study, Deng et al. (2008) explored the concept of absorptive capacity (the ability to acquire, develop and apply new knowledge to work) as a preliminary factor of effective IT use. The researchers hypothesized that absorptive capacity, through enhanced IT utilization for problem-solving/decision support, enhances task productivity and innovation in the IT-enabled engineering work context. Indeed, the results indicate that the absorptive capacity in terms of task knowledge, computer knowledge, and problem-solving modalities is a predominant factor of effective IT use. This implies that absorptive capacity can be an essential individual characteristic which determines different effects of IT use. Yet, the concept of absorptive capacity as an asset used in this study has been criticized by inconsistent conceptualization as an asset versus capability (Roberts et al., 2012).

To study the impact of personal characteristics, de Koning and Gelderblom (2006) explored the effect of age and skills of employees in the link between IT use and individual performance. Based on a survey in the printing industry and wholesale trade sectors, the researchers achieved the following results. First, unlike younger workers, older workers use IT less actively in their work. Moreover, this category of employees is shown as one that uses less complicated applications and has difficulties in using IT. Second, it is demonstrated that skills have a strong moderating effect on the relationships between IT use and individual performance. Yet, it is also shown that training in IT use does not have any significant effect on individual performance. One explanation for this result as elaborated by the researchers is that in the field of IT learning by doing is more effective than formal training.

While personal characteristics were recognized to a certain degree as relevant for investigating the impact of IT use on individual performance, other studies recognize the importance of organizational settings such as training and education, information work infrastructure, managerial practices. In more details, Ahearne et al. (2005) showed that only under high levels of user support and training, the use of an IT system enhanced efficiency and effectiveness. The researchers also demonstrated that low levels of user training and support decreased both efficiency and effectiveness of sales representatives. Moreover, it was established that individuals with low IT use and high levels of training (support) were not productive. These results prove that only adequate user support and training can enhance individual performance. The study was one of the first attempts to study the impact of organizational settings on individual performance particular categories of information workers such as sales representatives. Yet, some aspects of the study design do require further development. For example, self-reported

62

measures were used to operationalize the main constructs of the study which are subject to response bias. Second, evidence of causality can be questioned, since the study is not longitudinal in nature. Third, it is well-known that other factors of organizations can affect the link between IT use and individual performance such as incentives, empowering behavior, and organizational culture. Finally, the study has been conducted within one company that is a potential threat for results’ generalizability.

In relation to this, Jelinek et al. (2006) support the aforementioned study adding that training, customer pressure, and peer use of technology affected IT use and enhanced individual performance. The technology-acceptance model, the theory of reasoned action and the guiding principles of interactional psychology and individual goal orientation all formed a theoretical framework of the study. The study demonstrates that both learning and performance orientations have a strong positive impact on the intention to use a new IT system. Customer pressure, initial and continuous user training are also shown to positively affect the intention to use a new IT system. The aforementioned factors all serve to demonstrate a positive effect on individual productivity. Moreover, it is demonstrated that previous performance has a direct impact on performance of individuals after the introduction of a new IT system. Yet, as in the previous study, this study was conducted within one company based on self-reported measures that are potential threats for generalizability and response biases.

Other organizational settings have been investigated in the study by Haner et al. (2009). By supporting the fact that the working infrastructure (the quality of the personal technology set) as one of the factors to increase individual productivity the researchers demonstrated that flexible and supportive work infrastructures had a particular impact on information worker productivity. This concept has been operationalized by the ICT quality index which includes items that are related to effectiveness (using the right means at work), efficiency (the effort made to achieve these goals), task-related communication (effort and intensity) collaboration quality and organizational process efficiency. The results demonstrate that as more information work requires autonomy and flexibility, the more mobile and higher quality technological equipment is required. A high-quality infrastructure is identified as a strong premise of individual information worker performance.

A set of studies have been conducted to test the DeLone-McLean model of information system success in relation to individual productivity (DeLone & McLean, 1992, 2003). The model explains the relationships between six of the most critical dimensions of success such as system quality, information quality, service quality, intention to use a system, user satisfaction, individual and organizational productivity. The distinctive feature of the model is that it can be applied at both individual and organizational levels. The model has been tested in different settings and the relationships between constructs such as system quality, information quality, use and user satisfaction, and

63

individual performance are shown to have moderate support. For example, D’Ambra and Rice (2001) found that information quality is associated with quality of work and time savings. Similarly, Kositanurit et al. (2006) identified that system quality, information quality, and service quality are the most important factors that influence individual performance. Burton-Jones and Straub (2006) claim that a system’s use if conceptualized correctly demonstrates a positive impact on individual performance. Halawi et al. (2007) identified that intention to use, service quality, system quality and user satisfaction positively affect job performance. Contrary to the obtained results, McGill et al. (2003) found that the intended use is not significantly related to individual impact. McGill and Klobas (2005) have not found any significant relationship between system quality and individual productivity. Yet, the researchers demonstrate that user satisfaction is a significant factor that affects individual productivity. Similarly, Iivari (2005) found that unlike user satisfaction, the influence of system use was not related to individual productivity.

A set of information system success studies in terms of individual performance demonstrates that essential work has been done in this area. Yet, while the assessment of information system success was reported consistently, there is little consensus among researchers and practitioners on the exact impact of the model constructs on individual performance. As demonstrated above, current studies provide many inconclusive and contradictory results. Researchers explain such contradictory results by relying on determinants of computer acceptance that may be inadequate in determining the impact of various IT systems on individual performance (Gyampah & Salam, 2004). As mentioned earlier, there is little recognition that acceptance of technology may lead to increased individual performance. Moreover, the information system success model is IT system centric, i.e. based on the assumption that individual worker performance is determined by characteristics of the IT system as such in an isolated manner from other surrounding factors. This is rather different from the position assumed in this research which states that the performance of the information workers is conditioned by the use of IT, together with other non-IT factors that have to be synchronized as a whole.

By studying the relationships between IT use and individual performance, Sandhi (2013) came to the conclusion that IT use and the nature of IT use cannot alone impact IT-enabled productivity performance. Self-efficacy, trust, system quality and information quality are no less important factors of effective IT use. Therefore, it is remarkable that a set of current studies acknowledges the importance of other surrounding factors of effective IT use while not providing any evidence of any synergistic effect between those factors (or more precisely, how these factors can complement each other) and their impact on individual productivity (Milgrom & Roberts, 1990). There is still a lack of knowledge about optimal clusters of practices in relation to IT use that fit together and increase individual performance.

64

Overall, while most of the studies in the past on the impact of IT use on individual performance were based on understanding how people accept and adopt IT (Davis, 1989; Goodhue & Thompson, 1995), current literature demonstrates that there is a growing and continuous activity in the area of IT-enabled productivity with a focus on post-adoption scenario. This can be explained by the fact that acceptance does not necessary imply productivity gains (Jain & Kanuno, 2005) and the task-technology fit does not always lead to performance improvement (Lucas & Spitler, 1999). Further, the literature review demonstrates that there is a growing interest in individual IT-enabled productivity in information-intensive occupations from clerical to upper management levels, including recruiters, consultants, sales representatives and engineers. The contribution from current studies of individual IT-enabled productivity to information systems and economic research is presented below in more detail.

First, a number of studies began to pay more attention to the concept of IT use and enrich our understanding of IT system use in post-adoption contexts. Yet, more attention is required to investigate the link between IT use and individual performance outcomes, especially productivity. Moreover, different conceptualizations of IT system use exist, which make operationalization of this concept blurred and applicable only in specific contexts. Second, current studies demonstrate the relevance of the concept of multitasking from the perspective of IT use and its impact on individual productivity. Yet, the concept of multitasking alone requires additional research. Third, a number of studies investigate the communication aspect of information work in relation to individual productivity. However, we have to acknowledge that information work consists of more activities than the transfer of information alone. Finally, a number of studies have explored the impact of individual and organizational factors in relation to individual IT-enabled productivity by applying different theoretical backgrounds and frameworks which are difficult to align.

To summarize, most of the reviewed studies on IT-enabled productivity at individual and task level are built on the assumption that there is one set up of factors for productivity gains. Yet, taken together, these studies show that such a unidimensional view is inadequate for our research purpose. First, it is not enough to conceive individual worker productivity in terms of one or two factors only. The varieties of factors investigated by the many studies, especially, show clearly that individuals’ productivity is conditioned by a set of factors that may interact with each other. Second, the current studies show a kind of variance that suggests that different set-ups of factors in a different context may produce productivity gains. Thus, it becomes obvious that post-adoption contexts require a new theoretical ground that not only assumes an unidimensional and technology-centric view in relation to individual performance. This new ground has to be based on a premise that individual performance basically depends on how technology is used together with other

65

non-IT factors that have to be synchronized as a whole. We demonstrate the shift towards this perspective in IT-enabled productivity studies at different levels of the economy in the next section.

2.3 A shift towards a complementary approach in IT-enabled productivity studies

It has already been mentioned that our research focus is on individual IT-enabled productivity. This level of analysis is important since it is assumed here that only an understanding of individual level productivity can help build theories and models to address the synergies that emerge when individuals are grouped in teams, organizations, and economies (Ruch, 1994). However, in the last decades, the issue of how IT influences productivity and economic growth has been hotly discussed among researchers at macro- (Bugamelli & Pagano, 2004; Melville et al., 2004; Lee et al., 2005; Jorgenson, 2007; Arvanitis & Loukis, 2009; Dedrick et al., 2013), meso- (Lim et al., 2004; Hu & Quan, 2005; Han et al., 2011), and micro-levels (Brynjolfsson, 1993; Brynjolfsson & Hitt, 1996; Dehning & Richardson, 2002; Bloom et al., 2007; Kohli & Grover, 2008; Aral et al., 2012b; Tambe et al., 2012). While a review of these studies demonstrates that there is a shift towards complementarity perspective at all three levels of the economy, we lack an individual-level understanding of what complementarities affect individual IT-enabled productivity. Below, we summarize the obtained findings of recent relevant research on IT-enabled productivity at all three levels of the economy and position the present research in relation to the theoretical and empirical gap addressed here.

Generally speaking, macro-level research enables us to identify common trends in the economy as a whole. While initial research had failed to find a relationship between IT investment and productivity growth (Strassman, 1990; Loveman, 1994), current studies have demonstrated that IT is a major contributor and driving force of labor productivity and economic growth at the macro-level of developed countries (Lee et al., 2005; Jorgenson, 2007; Melville et al., 2007; Han et al., 2011). Yet, not all developed countries have received benefits from investment in IT in the same way. For example, it has been established that the diffusion and adoption of IT was delayed more than 7 years in Italy compared to the USA. The reason for this delay, as explained by the researchers, was in the missing complementary investments such as investments in a skilled workforce and re-organization (Bugamelli & Pagano, 2004). The same idea was supported by Arvanitis and Loukis (2009) to explain significant differences in IT-enabled productivity between Greece and

66

Switzerland. Supporting a complementary approach, Melville et al. (2004) emphasize that IT investments provide value, yet their impact on productivity growth depends on complementary resources such as a competitive and general macroeconomic climate. Following the same approach, Lee et al. (2005) summarized complementary factors of productive IT use at country level such as specialized information infrastructure, human resources, research and development, adaptive business models, telecommunication, liberalization, and reorganization. To summarize, the aforementioned studies indicate a growing interest among the research community in the various complementary factors that can drive IT-enabled productivity at the macro-level.

Although a positive and significant link between IT and productivity at the meso-level was established, the main concern in recent studies is whether particular industries benefit more from spending on IT, do they have some specific characteristics and distinctive features? Over the last decade, some research attempts have been made to shed light on the above-stated inquiry. For example, it was found that a relationship between IT investments and productivity is moderated by the interaction between value-chain information-intensity and product information intensity, meaning that investments in IT in industries with products and value chains that have high information intensity are associated with a significant productivity increase (Lim et al., 2004; Hu & Quan, 2005). By studying inter-industry IT investments, Han et al. (2011) noticed that IT spillovers (IT investments made by supplier industries) increase productivity of downstream industries. It was also established that in contrast to regular capital, IT investments increase productivity of firms in more competitive industries, yet not in dynamic industries where changes are difficult to predict (Melville et al., 2007). While the meso-level of analysis is important for IT-enabled productivity research, it was highlighted that IT-complementary investments at micro-level may help explain the cross-industry and aggregate productivity growth in relation to IT investments (Basu & Fernald, 2007). This suggests that there must be other, complementary factors, that influence productivity from IT use, which requires studies that focus on the micro-level.

The micro-level of the economy, together with the recognition of the organizational context in studies of IT productivity is of special research interest due to the ability to understand and measure IT investment in terms of productivity in more specific circumstances with higher accuracy and lower probability of measurement errors (Gurbaxani et al., 2000). Since the emergence of the IT productivity paradox (Solow, 1978), a number of early attempts have failed to find evidence that IT investments increase firm performance (e.g. Strassman, 1990; Brynjolfsson & Hitt, 1993; Loveman, 1994). The existence of the productivity paradox was summarized by Brynjolfsson (1993) in four hypotheses such as (i) mismanagement practices (difficulties in managing information and IT), (ii) mismeasurement of outputs

67

and inputs (productivity gains are real, but current accounting methods miss them), (iii) time lags due to learning (productivity growth from IT investments takes time to show up), and (iv) redistribution of profits (IT may be beneficial to individual firms, but add nothing to industry or country output). It was suggested that in order to understand IT-enabled productivity, each hypothesis has to be tested individually and, after that, together.

While previous studies have made attempts to test the aforementioned hypotheses and reveal the relationships between IT investments and firm productivity (Brynjolfsson & Hitt, 1996; Mukhopadhyay et al., 1997; Melville et al., 2004) in recent years, the main challenge for several studies became an explanation of the so-called “firm effect” (the effect of business capabilities as the primary driver of value) in IT-enabled productivity (Kohli & Grover, 2008). For example, it was established that the impact of IT-enabled information management capability on firm performance is mediated by customer management-, process management-, and performance management capability (Mithas et al., 2011). Another study proved that there are three-way complementarities among the adoption of IT (human capital management software), performance pay and human resource analytic practices (Aral et al., 2012b). On average, firms that adopted all factors together experienced more improved performance than firms that focused only on pairwise relationships. By applying the same theoretical framework, Tambe et al. (2012) identified that a combination of IT investments, decentralization of decision-making and external focus (the ability to respond to changes in the external operating environment) is associated with higher firm productivity. While it is difficult to provide a clear classification of the complementary factors that affect IT-enabled productivity at the firm level, findings from the above-mentioned studies are consistent with complementary relationships between managerial capabilities and investments in IT.

To summarize, while the complementary approach is shown as a promising research direction for investigating IT-enabled productivity at different levels of the economy, it was remarked that the level of investigation can play an important role (Dehning & Richardson, 2002; Schryen, 2013). For example, we have evidence that when similar firms adopt the same micro-level complementarities of productive IT use, they manifest different performance (Pinsonneault & Rivard, 1998; Hughes & Scott Morton, 2005). Moreover, Devaraj and Kohli (2003) highlighted that more research is required to understand how actual IT use affects productivity, meaning that research attention has to be directed towards the lowest levels of investigation. All this implies that there is a need to open up the “black box” of micro-processes in order to gain insights on the impact of complementary factors on IT-enabled productivity. Therefore, the shift to the nano-level is expected to remediate the absence of a well-grounded mechanism on the interaction between actual IT use and productivity. Below, we present a more detailed description of complementary factors and their impact on IT-enabled productivity at firm,

68

establishment (plant), and individual level to demonstrate the research gap addressed in this research, namely the lack of knowledge of complementary factors of productive IT use at individual level.

2.4 Complementary factors of productive IT use at firm, establishment (plant), and individual level

By investigating why some firms benefit more from IT use than others, Pollalis (2003, p. 470) raised an important question: “…how other major elements of organizations such as strategy, people, organization structure interweave with IT impact overall business performance?”. While recent studies succeeded in demonstrating complementary relations between several factors of productive IT use at firm and establishment (plant)9 levels, very little empirical evidence exists on the nature and relationships between complementary factors of productive IT use at individual level (Ennen & Richter, 2010; Brynjolfsson & Milgrom, 2013). Below, we review these studies in more detail to demonstrate the research gap addressed here.

Complementarity theory has been widely applied to study the effect of various work practices together with the use of IT on performance especially at firm and establishment levels. For example, Milgrom & Roberts (1990, 1995) suggested that the successful use of IT in the “modern manufacturing firm” requires joint adaptation of a set of factors such as (a) frequent product and process improvements, (b) highly skilled cross-trained workers, (c) decentralized decision-making processes, (d) focus on cost and quality rather than on volume, (e) high reactivity to the customers’ demands, (f) targeted markets rather than mass marketing, (g) low inventories, (h) reliance on outside suppliers, (i) production runs oriented towards scope economies rather than scale economies. All these factors together with the use of computers have to provide complementary interactions and, therefore, increase firm performance. This proposal, on the one hand, is theoretically feasible and important as it offers resolution to various theoretical challenges that could not be handled positively without complementary notion. On the other hand, there are still very few empirical studies on complementarities of productive IT use at individual level. Hence, this lack of studies also motivates the present

9Establishment is a place of residence or business with its staff and possession. An establish-ment is typically part of a firm, which may have several establishments and which may manifest different levels of performance even though belonging to the same firm, which makes produc-tivity assessments at establishment level relevant. Plant is a building for the manufacture of a product.

69

research as it attempts to both provide empirical support for the complementarity theory and also extend this theory as such into the individual worker domain. Below, recent empirical evidence on complementary factors of IT-enabled productivity at firm, establishment (plant), and individual level is reviewed to give us an opportunity to identify where research attention is required.

Previous research explored the complementary relationships between IT use, decentralization of decision rights, team-oriented production and demand for skilled labor by using panel data across 272 American companies from different industries (Brynjolfsson et al., 2002). The findings demonstrate that investments in IT strongly correlate with investments in organizational practices. Moreover, firms with high levels of computerization and the aforementioned practices were shown to experience the growth of output. In relation to this study, it was identified that the impact of computerization on firm output was up to 5 times higher in a long-term perspective (Brynjolfsson & Hitt, 2003). All this suggests that the contribution of computerization may be strongly associated with time-consuming investments in complementary inputs which can explain slow productivity growth.

There is also a significant volume of literature on the differences in productivity growth among firms in various countries that was explained by complementary relations between various firm factors. For example, Caroli and Van Reenen (2001) by using data from British and French establishments, found complementarities between skilled labor and workplace reorganization. Bresnahan et al. (2002), by examining 300 large US firms, reached a conclusion that increased use of IT, changes in organizational practices and changes in products and services together are mutually complementary. Bugamelli and Pagano (2004) by using a dataset from 1700 Italian firms, concluded that a 7-year technological gap between Italian and similar American firms can be explained by missing complementary business reorganization to investments in IT. In general, this type of studies has conclusively shown that despite country-specific factors, the ways in which firms configure their practices is no less important. Yet, the fact that those studies are considerably general in what they mean by complementarities (organizational capital, learning practices, etc.) and do not say which factors have to be configured and in which manner, points to the need to decompose the firm and investigate complementarities at the lower level such as individual level.

In recent years, there has been an increasing amount of literature on specific complementarities of effective IT use in US firms. For example, it was established that the performance impact of IT investments is greater in firms with a high level of diversification (Chari et al., 2008). Yet, Dewan and Ren (2011) noted that a complementarity fit between IT investments and diversification is stronger for related diversification than for unrelated. Another study pointed that the adoption of a specific human-resource

70

management process-enabling technology leads to greater company benefits together with the simultaneous introduction of human resource analytics and performance pay organization practices (Aral et al., 2012b). Tambe et al. (2012) have proposed three-way complementarities, including investments in IT, the external focus of a firm, and decentralized decision-making. In general, this type of studies, by discovering different complementary factors, provides empirical evidence that a system of technological, managerial and external factors has a strong joint impact on IT value creation at firm level. However, the key limitation of these studies is that they boldly assume that the introduction of local IT systems to support operations (manufacturing, supply chain, or human resource management processes) will lead to firm-level performance increase. Clearly, a firm is typically much more complex than being reduced to a single domain, be it manufacturing or supply, or human resource management. All this points to the need for studies at individual level to measure local performance variables without assuming a direct and positive effect on the whole firm.

At establishment (plant) level, based on a comprehensive list of complementarities including team-based incentive, working training, employment security, communication, job rotation and flexibility (Ichniowski et al., 1997), Black and Lynch (2004) indicated that workplace innovations such as re-engineering, teams-incentive pay, and employee voice are significant complementary predictors of productivity growth. Also, there is evidence that new IT is strongly correlated with the adoption of technical and problem-solving skills (Bartel et al., 2007). Overall, Bloom et al. (2007) concluded that beneficial use of IT depends on “internal organization” of establishments as well as from external conditions such as greater market size and well-educated employees. Yet, as previously mentioned, all these studies assume that one set of complementarities fit all establishments in order to increase productivity and that local IT systems have a global effect on performance, pointing to the need for individual level studies aimed to demonstrate which factors have to be configured and in which manner to increase IT-enabled productivity.

Although the complementarity theory has been widely applied at firm and establishment (plant) level, there has been limited use of this theory at individual level. For example, Athey and Stern (2002) explored the impact of a specific application of IT – “Enhanced 911” and operational process on emergency health care outcomes. The main findings indicate that the impact of adopted IT use increased individual performance with 1%, according to the index of the health status of cardiac patients. However, the use of new IT together with the adoption of new job design did not have any significant impact on individual performance. Another study by Autor et al. (2003) revealed that while computer capital is a complement in performing non-routine, complex problem-solving tasks, it accomplishes a substitution function in routine tasks.

71

Based on previous studies that applied the complementarity theory at individual level, we position the present study in the following way (Table 2.2).

Table 2.2: Position map

Study

Variables used

Obtained/expected results IT

use

Ope

ratio

nal

task

/pro

cess

Indi

vidu

al

fact

ors

Org

aniz

atio

nal

fact

ors

Athey & Stern (2002)

Yes Yes No No IT use is not complementary to the structured process.

Autor et al. (2003) Yes Yes No No IT use is complementary to complex

decision-making tasks.

The present research

Yes Yes Yes Yes

A full set of complementarities can increase individual IT-enabled productivity when matched correctly.

In the first study, different kinds of IT systems (telephone versus “911

system”) used together with operational processes (basic versus highly structured “Emergency Medical Dispatching”) were investigated (Athey & Stern, 2002). However, the researchers found it difficult to demonstrate productivity gains from these complementary factors. This can be explained, first, by difficulties in demonstrating complementary relationships between a limited number of factors (Ennen & Richter, 2010). In relation to this, the researchers recognize the importance of training activities and decision-making structure as IT-complementary factors, yet do not investigate them. Second, a rate of survival was used as a dependent variable. However, “911 system” and a structured process were deployed to address the receipt of an emergency call for help. Therefore, the very first activity in the emergency activity chain is to decrease the time from receiving the call to signaling an ambulance to go to the patient. After that, it would be rational to compare how the length of time for the ambulance to arrive after the emergency call was received. Thus, difficulties in finding IT complementarities can also be explained by the inappropriate dependent variable that is based on the whole chain activity.

72

The second study shows that there is a “complementarity effect” in the sense that IT use and non-routine tasks performed by information workers do complement each other, as it is assumed that some of the routine tasks are substituted by ITs (Autor et al., 2003). This implies that both IT use and task (process) factors are complementary to each other. While the study derives conclusions based on the premise that the cost of input is reduced for the same output, the study does not monitor the output performance. Yet, conclusions from this study are significantly essential, since the interaction approach of the complementarity theory that has been used provides a higher degree of granularity in the investigation of complementary relationships between factors.

In general, both studies made a significant contribution to identifying complementary factors of effective IT use at individual level. However, they both ignored relevant complementary individual and organizational factors that are identified as important factors of individual productivity in management science (Mason & Mitroff, 1973; Yaverbaum, 1988; Kraemer & Danziger, 1990; Sonnentag, 2003; Mullins, 2007; Hopp et al., 2009). In the next chapter, we address this gap by developing and testing a research model where IT use, individual and organizational factors are studied together in order to increase IT-enabled productivity at individual level.

Overall, empirical evidence at firm level demonstrates that a system of technological, managerial and external factors has a strong impact on IT value creation (Dong et al., 2009). Yet, by examining a set of IT-complementary organization practices, Poon et al. (2009) came to the conclusion that clustering among organization practices makes their modeling and empirical analysis difficult. They even showed that some organizational practices together, e.g. computer use and implementation of a budget plan may affect organization performance negatively. This implies that the relationship between complementary organization practices is quite complex and is still an area where additional research is needed. Moreover, the question is not only about which complementarities to assume, but also how these complementarities interact in order to provide productivity gains.

In summary, the review of studies at firm and establishment (plant) level demonstrates that IT itself cannot enhance performance and in order to understand the whole potential of IT use, it is necessary to explore other no less important complementary factors such as training and education, incentive pay, decentralization of decision-making and process re-engineering. As shown above, the complementarity theory offers a broad framework to understand a complex effect of different factors on the performance indicators. The review of the recent studies on the complementarity theory shows that the joint impact of IT use and complementarities can be a strong predictor of improved performance. The fewness of studies on complementary factors of individual IT-enabled productivity demonstrates that the research work in this area has not

73

progressed sufficiently. Moreover, just which complementary factors are needed to increase IT-enabled productivity of information workers at individual level remains an under-researched area. This evidence lays a foundation of the identification of the theoretical and empirical gap in this research.

2.5 Summary of current knowledge In this section, we summarize the current knowledge on the theme of IT-enabled productivity with a particular focus on the so-called nano-level. Since various studies investigating the impact of different factors affecting individual IT-enabled productivity, including IT use, applied different theoretical approaches, the obtained results display considerable diversity. Accordingly, we organized the review of the literature around the most important topics in the field of individual information worker productivity; for example, individual differences, job/task design, human resource management activities, and IT use. Further, we reviewed individual IT-enabled productivity studies that applied a non-complementary approach and addressed IT-enabled productivity partially. Based on this review, we identified four research streams: conceptualization of IT use and its impact on individual productivity; IT use, IT-based multitasking and individual productivity; IT-enabled communication patterns and individual productivity; and the impact of individual and organizational factors on productive IT use. Then, we paid particular attention to IT-enabled productivity studies at different economic levels to demonstrate both an essential shift towards the complementary perspective and the lack of such studies at the individual level. The summary below provides a comprehensive overview of the previously published research that helps us identify theoretical and empirical gap addressed in this research.

Worker productivity is one of the oldest research themes in management studies. The so-called classical approach proposed that an individual’s productivity was mainly conditioned by the standardization and optimization of the work process as well as worker motivation induced by exogenous compensation schemes (Taylor, 1947). The human relations school proposed that workers’ productivity was conditioned by organizational structure, job design, an endogenously induced motivation by leadership (Tausky, 1978). The subsequent socio-technical systems approach synthesized the two previous approaches, where an organization is regarded as a system of interactions between the social, i.e. mental and social needs of workers, and the technical, i.e. tasks, tools, location (Daft, 2000). The contingency approach emerged as a further development, highlighting the importance of contingent fit between various factors, such as task, organizational structures, and the

74

environment, which may jointly impact the worker performance (Helms-Mills et al., 2008). Somewhat in parallel, studies in psychology have shown that worker productivity is conditioned by his or her well-being, skills and ability to perform, attitudes, motivation and satisfaction, and job-design, among other factors (Clements-Croome & Kaluarachchi, 2000); these factors of an individual worker’s productivity may be categorized into personal, organizational, social and environmental (ibid.). Yet, these approaches and studies mostly emerged in a pre-IT system era and did not distinguish between material and information workers.

Peter Drucker (1999), a key advocate of the notion of the information worker and his or her peculiar conditions, proposed that the key factors that condition information worker productivity include task definition, degree of self-management or autonomy, continuous innovation as well as learning and skill development as part of work, and focus on quality rather than quantity (Drucker, 1999). Independent information worker studies show that central factors which condition productivity include type of worker personalities (Zmud, 1979; Agarwal & Prasad, 1999), motivation (Frey & Osterloh, 2002) and training for skills (Kessels, 2001; Kraiger, 2003), decision-making structures or degree of autonomy (Kozlowski & Bell, 2003; Schneider & Smith, 2004), type of work processes (Parker et al., 2001; Sokoya, 2000) including multi-tasking (Appelbaum et al., 2008; Bell et al., 2005), and technology support, particularly the use of IT systems (Jain & Kanungo, 2005; Aral et al., 2012a).

When IT’s penetration in the work process became widespread, both the technology acceptance model (Davis, 1989) and the task-technology fit model (Goodhue & Thompson, 1995) have been developed to understand the success of IT use. Yet, both models have been criticized for their limitations in understanding how IT use can affect individual productivity in post-adoption contexts (Jain & Kanungo, 2005). This gave rise to studies that addressed the need to investigate how IT use can affect individual productivity of information workers (e.g. Czerwinski et al., 2004; Kvassov, 2004; Jain & Kanungo, 2005; Aral et al., 2006; Sundaram et al., 2007; Chung & Hossain, 2009; Wu et al., 2009; Aral & Van Alstyne, 2011). Those studies applied different theoretical approaches and investigated the impact of various IT tools on productivity in different empirical settings. Although these studies provide a unique empirical evidence, the obtained results were mixed and difficult to draw conclusions from. One key limitation of this large body of research is that each study offers only a partial understanding of what conditions the productivity of an information worker. Another limitation is that those studies tried to investigate IT-enabled productivity based on a particular information-intensive profession challenging the generalizability of results obtained. These key limitations in the literature on individual IT-enabled productivity are further targeted by the research model formulated in this research.

75

Earlier research in the information systems field has made attempts to address partiality and investigated the impact of a number of factors that together can affect their IT-enabled performance (Dickson et al., 1977; Benbasat & Taylor, 1978; Benbasat & Dexter, 1982; Blaylock & Rees, 1984). These studies suggested that individual differences, particularly cognitive style, task design and IS characteristics might be included in research that addresses the link between IT use and individual performance. Yet, this research initiative came to a halt for a number of reasons: (i) researchers found it difficult to find underlying theories to address the complexity of interaction between individual, task and technology, (ii) particular focus on cognitive style as a basis of information systems research has been criticized by Huber (1983) as an unsatisfactory basis for information systems research, yet it was later reinstated (Robey, 1983; Ramaprasad, 1987), (iii) since technology use was in its infancy, and the first results were difficult to interpret.

Since that time, technologies became an inseparable part of our life. In information work, the production process nowadays is characterized by a close interaction between an individual and IT use (specifically, core production tools designed to assist in information processing). To overcome partiality, a specific theory of complementarities has been advanced, where its most elaborated notion is provided by the recent developments in organizational economics (Roberts, 2007). This offers a specific conception to account for, and explain how, various organizational practices may fit each other and which patterns of fit may exist (Brynjolfsson & Milgrom, 2013). This theory holds that positive complementarities emerge when the marginal return to one activity or resource increases in the presence of another activity or resource (Milgrom & Roberts, 1995). For example, the value of a newly installed ERP-system will probably increase in the presence of workflows that are adapted to its functionality as well as provided training of system users.

Complementarity theory has been widely used at different levels of the economy. Over the last two decades, however, more and more studies claim the need to move from country, via industry and firm level to the individual level of the investigation, where each aggregated level asks for de-aggregation (Melville et al., 2004; Kohli & Grover, 2008; Schryen, 2013). The individual level constitutes the last instance of investigation and has the potential to reveal the dynamics of the above levels. Moreover, studies specifically at firm and establishment level demonstrate which kinds of other factors are to be summarized, not exactly which factors and in which manner they should be synchronized, pointing to the need for more detailed studies at individual level. Yet, empirical and statistical support remains very weak with regard to which specific factors complement an information worker’s use of IT systems to increase productivity at the individual level.

For example, only two studies could be identified that used complementarity theory at individual level and both had adopted an interaction approach (Ennen & Richter, 2010). In the first study, Athey and

76

Stern (2002) explored the impact of a new IT system use in the context of 911 emergency call centers and its impact on emergency health care performance. The study shows that the use of the new IT system together with a new work process, that was supposed to offer a higher degree of support or fit than the old practices, did not produce any significant productivity increase. The second study (Autor et al., 2003) differentiates the types of work processes and shows that the use of an IT system complements the conduct of non-routine, complex problem-solving tasks. Regarded jointly, the two studies show that the interaction between the work process and IT use matters for information worker productivity yet is of a complex nature where homogenization of conditioning factors will ignore the specific dynamics between two factors. A key limitation of these two studies is the assumed partiality, as only the impact of two factors is being investigated – the work process and the IT system use. Although the two studies suggest that various kinds of factors may determine the productivity of an information worker, yet do not investigate them. The key question here is then how to fit these various factors into a joint system of complementarities that condition programmer’s productivity.

Given the above, the literature review demonstrates that productivity of information workers may be conditioned by an array of factors. Yet, a key limitation of the reviewed studies is that each study offers only partial understanding of what conditions the productivity of an information worker and there is a lack of comprehensive understanding of the productivity of the individual who uses an IT system. Another limitation of the reviewed studies is that their design is typically context-dependent (focus on one profession), challenging the generalizability of the results obtained. Further literature review of IT-enabled productivity studies at different economic levels demonstrates that there is a noticeable shift from a non-complementary to a complementary approach where a system of complementary factors that interact with each other is expected to increase IT-enabled productivity. There is also a shift from country, via industry and firm levels to the individual level of the investigation, where the last level has the potential to give new insights into the above levels’ dynamics, even though the links between the various levels are complex. While studies at the micro-level provide evidence about the kinds of factors which are complementary to productive IT use, they are silent about which factors and how exactly they have to be synchronized in order to increase IT-enabled productivity. Therefore, the present research attempts to discover what factors are needed and how to synchronize them, in order to increase productivity of an individual information worker who uses IT system. In the next chapter, a theoretical framework is designed and a conceptual model is developed to present hypothetical key complementary factors that can increase individual IT-enabled productivity of information workers at the individual level.

77

3. Formulation of research model

In this chapter, we introduce a research model for the conception of individual information worker productivity. The model adopts a systems approach of the complementarity theory (Ennen & Richter, 2010), meaning that a number of key factors are assumed to interact with each other in order to produce superior information worker productivity. By integrating empirical evidence from prior studies, we developed two theoretical hypotheses regarding complementari-ties of productive IT use. The design of the research model influenced the choice of the research methodology to be used in further empirical investiga-tion.

3.1 The notion of complementarity

Previously, we have already partially described the notion of complementarity; this notion will be reviewed here in more detail. Over the last decade, the idea that complementarities are important for reaching productivity gain from IT use has received special attention (Ennen & Richter, 2010). The notion of complementarity offers an approach explaining how organization characteristics fit each other and what patterns can exist between various factors (Brynjolfsson & Milgrom, 2013). In simple terms, a complementarity emerges when the marginal return to one activity increases when a firm does more of another activity (Milgrom & Roberts, 1995). The implementation of the complementarity theory is well defined and motivated in the study by Lee (2001, p. 193) who noted that: “…The notion that “x (e.g. IT) works better (e.g. generates more value) with a certain condition of y (e.g. an efficient process)” has been referred to as “fit”, “synergy”, or “complementarity””. It is a powerful concept because it shows that the effect of x should not be considered alone; it is always affected by another variable y”. Since the complementarity theory is based on the assumptions from contingency and configuration theories, it views an organization as a system of interdependent strategic, organizational and individual elements (Roberts, 2007). This implies that the overall performance in organizations can be explained by revealing ties of complementary interactions between elements.

78

Porter and Siggelkow (2008) by simplifying a mathematical expression of the complementarity theory, define complementarity as a function with the following property: f (x’’, y’’, z) – f (x’, y’’, z) ≥ f (x’’, y’, z) – f(x’, y’, z) (1)

for all x’’ > x’, y’’ > y’ (2) and all values of z, (3),

where x and y are complements and z is a vector of variables.

According to the complementarity theory, three conditions have to hold for two factors to be complements. Complementarity occurs and forms a system when: 1. increasing the factor x from its lower level x’ to the higher level x’’ is more beneficial when the factor y is at the higher level y’’ than at the lower level y’; 2. the relationship between x and y holds at all levels of x and y; 3. the relationship between x and y holds for all values of all the other factors z.

For the visualization of the complementarity presence, a cube view of three-way complementarities (Figure 3.1) has been used in recent studies (Tambe et al., 2012; Aral et al., 2012b).

Figure 3.1: Cube-view of complementary interaction

79

In a cube, each activity such as A, B and C can be configured in two ways, 0 and 1. For example, as mentioned before, managers may consider when in-troducing a new IT system (i) whether to invest in training and education, (ii) whether to reconsider a rewards system, and (iii) whether to give workers more freedom to organize their activities. Eight different possible combina-tions are represented on each vertex of the cube.

According to the complementarity theory, four tests have to be significant for activities A, B and C to form a system of complements (Brynjolfsson & Milgrom, 2013). These four tests are presented below. 1. A: f (1,1,1) – f (0,1,1) > f (1,0,0) – f (0,0,0) 2. B: f (1,1,1) – f (1,0,1) > f (0,1,0) – f (0,0,0) 3. C: f (1,1,1) – f (1,1,0) > f (0,0,1) – f (0,0,0) 4. The system of complements: [f (1,1,1) – f (0,1,1)] + [f (1,1,1) – f (1,0,1)] + [f (1,1,1) – f (1,1,0)] – [ f (1,0,0) – f (0,0,0)] + [f (0,1,0) – f (0,0,0)] + [f (0,0,1) – f (0,0,0)] > 0

Previously, we have argued for the need to assume a complementary based approach in order to study the conditions of productivity increase from an individual worker’s IT system use. The research on complementarity factors can be conceived in terms of two key phases. While the first phase is based on examining what factors are relevant for a complementarity system, the second phase is based on testing of a particular complementarity set-up. In this research, the identification of factors for a potential complementarity fit is grounded only on previous studies and no new empirical studies were conducted to find novel, previously unknown factors. The reason we chose the second phase is because, over several decades, a large set of individual productivity factors have been identified, however without investigating much of their interactions. Therefore, rather than conducting a search for new factors, there is a need to test for configurations of the existing factors that offer empirical support for individual factors.

After some complementarity configurations are identified and when some existing factors can be dismissed as irrelevant for those configurations, then there may be a need for explorative studies to find additional factors. However, in this situation, the identified preliminary configurations of complementary factors will offer guidance for the search of new factors that manifest complementary potential. At the present time, such guidance does not exist as no previous research has identified any stable complementarity set-up.

80

3.2 General research model and hypotheses In this section, we present the theoretical background relevant to the study of productivity of information workers using an IT system and formulate a research model. We draw from a systems approach (Ennen & Richter, 2010) of the complementarity theory to explore the integrative effect of a system of complements when an alternative, more aligned IT system is deployed to materialize complementarities of productive IT use. This theoretical base is crucial for understanding key constructs of IT-enabled productivity at the individual level and complementary relationships among them. By studying empirical evidence from prior studies, we investigated complementary factors which may be powerful drivers of individual IT-enabled productivity. Based on the theoretical background and prior research, we further developed a research model for IT-enabled productivity at individual level and illustrated its applicability by developing theoretical hypotheses regarding complementarities of effective IT use.

Our comprehensive literature review demonstrates that the use of an IT system requires reconsideration of pre-existing resources, as well as the introduction of new ones in order to increase individual productivity. Assuming the complementarity theory as a foundation implies that a set of theoretical constructs has to be allocated into one configuration, i.e. a system where each factor is specific in itself and also related in a particular manner to the other factors so that systemic properties can emerge when a new IT system is used. The key question here is then how to fit these various factors into a joint system of complementarities that condition information worker’s productivity.

In general, individual productivity of an information worker is conceived as a function of a single worker, IT tool used by the worker, a task conducted and contextual settings of the worker, i.e. processes that govern the interactions among the individual, task and IT tool in order to complete tasks (Hopp et al., 2009). These factors are grounded on a large set of independently conducted empirical studies which demonstrate that there exist a number of factors that may affect individual productivity, including worker personality (Zmud, 1979; Agarwal & Prasad, 1999), type of work processes (Sokoya, 2000; Parker et al., 2001), motivation (Frey & Osterloh, 2002), training for skills (Kessels, 2001; Kraiger, 2003), decision-making structures or degree of autonomy (Kozlowski & Bell, 2003; Schneider & Smith, 2004), and technology support, particularly the use of IT systems (Jain & Kanungo, 2005). Below, we present how we matched these factors together to formulate a research model of complementary factors.

Our conceptual model is grounded on the notion of complementarity. Complementarity theory (Milgrom & Roberts, 1990, 1995) functioning here as a meta-theory, and guiding the linking between the conception of

81

productivity (Jefferys et al., 1954), Kirton’s adaption-innovation theory (Kirton, 1976, 1994, 2003), conception of operational processes structure (MacCormack et al., 2001; Weber & Wild, 2005; Schonenberg et al., 2008) and human resource management sources (Amabile, 1996; Hayes & Allinson, 1997; Ryan & Deci, 2000; Baer et al., 2003; Ahearne et al., 2005; Sense, 2007; Bloom & Van Reenen, 2011). The aforementioned studies have helped us specify and motivate the expected patterns of effective IT use to increase individual productivity of an information worker in a situation when a more aligned IT system is used.

From the review of empirical studies, we have formulated a theoretical model where the following factors are proposed to fit each other in a particular manner: cognitive style, the structural complexity of the operational process, training activities, incentives, and decision-making structure (Pashkevich & Haftor, 2014, 2016). These factors are now configured into a proposed system of complementarities and depicted in the research model (Figure 3.2).

Figure 3.2: General research model of complementary factors and their value range

In this research, we distinguish between less aligned IT system (pre-

existing IT system) and more aligned IT system (new IT system). A less aligned IT system typically offers information and information processing functionality that is basic, generic and can easily support various variants of a certain kind of process. This IT system support is less sophisticated in terms of its alignment and integration with the work process that it is aimed to support. On the one hand, this IT system is more cost efficient since it may be used in many work contexts. On the other hand, it provides less support for automation and information processing to a work process. A more aligned IT system offers information and information processing functionality that is more adjusted and tuned to support the specific work activities within a specific kind of work process. Such IT system support is more sophisticated in terms of its alignment and integration with the work process. This IT system provides more support for automation and information processing in a work

82

process. This simple distinction is assumed to account for a firm’s reality, where executives make decisions to acquire more aligned IT systems either to replace work practices that are conducted without any IT system support and, more frequently, to replace work processes that are supported by an existing IT system with the aim to add productivity and quality into the work process.

In order to perform work and use IT systems in an information-intensive environment, there is an information worker that is an obvious, yet quite often ignored, factor in individual IT-enabled productivity studies. In information work, a key act of information processing is conducted by human cognitive faculties (Pyöria, 2005), rather than physical capabilities. This means that human cognition manifests a central factor for the comprehension of information work processes.

Studies in psychology show that humans may be conceived fruitfully in terms of various kinds of cognitive characteristics, which strongly influence the cognitive acts and thus information processing ability (Agor, 1984; Rowe & Mason, 1987). While there exist a number of partly competing and partly overlapping categorizations of human cognition, we adopt here Kirton’s scheme of adoption–innovation (Kirton, 1976, 1994). This scheme holds that human cognition may be characterized along a spectrum, from mainly adaptive behavior to mainly innovative behavior. This manifests differences in human information processing, problem solving procedures, decision-making modes, the need for autonomy as well as responses to changes. The adaptive cognitive character tends to operate in a systematic and well-structured fashion, step-by-step, and prefers explicit and stable situations that are predictable. In contrast, the innovative cognitive character tends to operate in a creative and ad hoc kind manner and prefers a lack of imposed and pre-defined structures. While no human individual is only adaptive or only innovative, repetitive empirical studies show that humans tend to be dominated by one of these cognitive types (Kirton, 1994).

We have adopted this scheme of adaptive-innovative cognitive style for several reasons. One reason is that studies show that both cognitive styles may perform well, but their behavior and productivity may depend on the context and its conditions (Kirton, 1994). Second, this scheme produces a good fit with the remaining factors included in the research model proposed here. Third is that the development within the domain of the theory of the firm (Roberts, 2007) shows that a firm’s success is conditioned by a deliberate coordination of two kinds of personalities: the systematic and the creative (Jablokow & Booth, 2006). Finally, the Kirton scheme is the most explored, elaborated and well-tested categorization of human cognitive styles, offering well-developed measurement instruments (Brown, 2001; Duff, 2004). It is thus proposed here that cognitive style constitutes a central factor that conditions IT-enabled productivity of an information worker and is relevant for the research question at hand.

83

An information worker does not operate in a vacuum, rather he or she executes a certain kind of work tasks that constitute a work process (Kock & McQueen, 1996). Previous studies show that the character of a work process influences the performance of an information worker (Keen & Scott-Morton, 1978; Amabile, 1996). While several conceptions of work process categorization are advanced (Weber & Wild, 2005), we assume here the distinction between a more structured work process versus a more flexible work process, as this distinction shows a clear impact on the performance outcome (MacCormack et al., 2001; Baer et al., 2003). A more structured work process is explicitly pre-defined in terms of its constituting activities and their order; also, the inputs and outputs are typically carefully specified (Abdolmohammadi & Wright, 1987). A structured process also manifests typically more stability in terms of recurring patterns of activities (MacCormack et al., 2001). Given this, we propose that a structured work process should be conducted by an adoptive cognitive style of an information worker while a flexible work process should be conducted by an innovative cognitive style of an information worker, as these should complement each other in relation to productivity achieved, in a manner that any alternative pairing of the two factors may not produce.

Information workers are dependent on the use of often advanced work technologies, such as work processes and IT systems (Czerwinski et al., 2004; Webster, 2012). The mastery of such work technologies influences work productivity (Hitt & Brynjolfsson, 1997) and may be supported and enacted with various training and education forms, which induces learning and thus develops work capabilities (Hayes & Allinson, 1997). Many studies have explored and established the significance of training and education of an information worker, which advances her cognitive ability to perform both new and current tasks in a better manner (Carland et al., 1994; Hayes & Allinson, 1997; Sadler-Smith & Smith, 2004). Further, the cognitive character of the information worker shows correspondence to the training form when learning outcome is monitored (Amabile, 1996). In our attempt to account for this complementarity, we propose that training can assume two modes. The first mode stipulates training that is received mandatorily in a pre-defined and well-structured fashion, prior to the initiation of the use of work technologies, such as the work process and the IT system – we can call this form for ‘push training’. The second form of training is received optionally, on-demand, whether prior to the initiation of the work technologies or after it and often for various lengths of time and with a different scope; i.e. when the information worker asks for it – we call this mode for ‘pull training’. Given this, we propose that high productivity performance of an information worker, with an adaptive cognitive style, is complemented with push-mode training while high productivity of an information worker with an innovative cognitive style is complemented by pull-mode training.

84

One of the theoretically and empirically best researched factors that conditions a worker’s productivity is motivation (Frey & Osterloh, 2002). This cognitive factor explains, as such, a significant amount of worker productivity (Kasof et al., 2007) and is also central for a firm’s performance (Roberts, 2007). Studies in psychology have established a key difference between exogenously and endogenously induced worker motivation (Amabile, 1996). These studies also show that the high performance of an adaptive cognitive style is matched to a greater degree with exogenous motivation, such as various compensation schemes, while high performance of an innovative cognitive style shows a greater match with endogenously induced motivation, such as positive appreciation from a manager (Amabile et al., 1994; Baer et al., 2003). Therefore, we propose that the model for the productivity of information worker as elaborated here includes exogenous motivation as a complement to the adaptive cognitive style while endogenous motivation complements the innovative cognitive style.

Yet another central area of information worker productivity is associated with worker freedom, autonomy, or self-management (Zabojnik, 2002). More operationally, this may be translated into decision-making freedom, authority, or power (Kozlowski & Bell, 2003). Decision-making, a clearly cognitive act, is central for many information jobs both in terms of decision frequency and scope (Hunt et al., 1989; Rowe & Boulgarides, 1992), and has shown significant impact on information worker productivity (Baer et al., 2003; Ahearne et al., 2005). Various studies in psychology show the presence of complementarity between decision-making authority, in terms of centralization versus decentralization, and an individual’s cognitive style (Amabile, 1996; Oldham & Cummings, 1996; Axtell et al., 2000). A higher degree of centralization provides an information worker with fewer decisions and more given guidelines for what to do (Baer et al., 2003). On the other hand, a higher degree of decentralization (lower degree of centralization) offers an information worker more freedom to decide on how to act and change actions, without the need to consult a manger or other decision maker, which should thus constitute a complement to an innovative cognitive style (Axtell et al., 2000).

The complementarities listed here, accounted mostly in pairs so far, are now linked together into one greater system of factors that complement each other, and where each factor is assumed to hold binary value set. This gives a foundation for the formulation of two key hypotheses subjected to empirical tests. Therefore, the following two hypotheses were formulated:

Hypothesis 1: Individuals with adaptive cognitive style will generate higher productivity when matched with a “stable” complementarity set-up that includes structured operating process, push mode training in work technology, exogenous incentives, and centralized decision-making, compared to other configurations of these factors.

85

Hypothesis 2: Individuals with innovative cognitive style will generate

higher productivity when matched with a “dynamic” complementarity set-up that includes flexible operating process, a combination of minor upfront mandatory training with optional on-demand training in work technology, endogenous incentives, and decentralized decision-making, compared to other configurations of these factors.

Therefore, we expect that adaptors rather than innovators experience

productivity advantages when a more aligned IT system is used together with a “stable” complementarity set-up. In contrast to adaptors, innovators win productivity advantages when a more aligned IT system is used together with a “dynamic” complementarity set-up. Yet, we have to mention that the key principle of the conceptual model and hypothesis development is that we do not invent totally new factors but rather start with pre-existing factors that have been established and explored independently by other studies. Our contribution is that by applying the systems approach of the complementarity theory, we propose a new and relevant system of complementarities (configuration of factors) relevant for the research question targeted. More detailed support for the proposed complementarity set-ups is provided below.

3.3 Complementarities of individual information worker productivity

In this section, we provide detailed support for the formulated research model and developed hypotheses. The section is organized as follows. First, we review studies that provide evidence of the fit between cognitive style and operational production mode. Second, we present a review of studies that have addressed the need to take into account cognitive style in developing training activities. Third, we analyze studies investigating interactions between cognitive style, operational production and incentive modes. The final part of this section reviews studies that have examined the link between cognitive style and decision-making mode.

86

3.3.1 Complementarities between cognitive style and operational production mode

For information work which is characterized by a high level of uncertainty, an effective operational process increases the ability to respond to new requirements in a timely manner (Gallivan, 2003). New IT systems together with new ways of doing work are supposed to create significant competitive advantages for companies (Ahadi, 2004). Therefore, to enrich the model we refer to the literature on the structural complexity of the operational process, which explicitly addresses various types of operational processes from highly structured to unstructured/flexible (Keen & Scott-Morton, 1978; MacCormack et al., 2001; Weber & Wild, 2005; Schonenberg et al., 2008; Rigby et al., 2016) together with a focus on individual adaptive/innovative cognitive style.

A work process has long been considered as an important contributor to individual productivity (Kock & McQueen, 1996; Thompson & Heron, 2005). For example, complex work processes (i.e. that is characterized by high levels of autonomy, identity, skill variety, significance and feedback) are expected to encourage worker motivation and, therefore, increase productivity (Hackman & Oldham, 1980). Other research has argued that the main characteristic of the operational process is its structural complexity (Markus et al., 2002). While operational processes are in theory characterized by a strict structure, in practice individuals are involved in processes with different structural complexity (MacCormack et al., 2001; Weber & Wild, 2005). In the management literature, a continuum from highly structured to unstructured processes has been identified together with core characteristics (Table 3.1.). Table 3.1: Operational process structure characteristics

Process type Problem Number of solution alternatives Judgment

Structured Well-defined Limited, well-specified alternatives

Little judgment is required

Semi-structured Reasonably defined

Limited, specified alternatives

Judgment is required

Unstructured/ Flexible Undefined Infinite/undefined

alternatives Judgment is highly required

Source: Abdolmohammadi and Wright (1987) For example, a well-defined problem and a limited number of alternative

solutions that require little personal judgment are core characteristics of

87

structured and semi-structured processes. In contrast, in unstructured/flexible operational processes the problem is ill-defined with numerous alternative solutions that require considerable judgment from an employee. Therefore, based on these core characteristics, the structured process is understood here as a process which is rigorously well-defined, has limited, well-specified solution alternatives, and where little judgment is required from the worker. In contrast, an unstructured/flexible process is understood here as a process which is ill-defined with a number of solution alternatives and requires a great deal of judgment from the worker.

It has been established in the past that the problems to be solved and their complexity may affect employees’ intrinsic motivation and performance (Amabile, 1996; Bommer & Jalajas, 2002). Specifically, complex jobs with high levels of autonomy, significance, identity, skill variety and feedback were identified as determinants of increased intrinsic motivation and job performance (Hackman & Oldham, 1980). Yet, cognitive styles may affect the way individuals respond to contextual conditions, including job characteristics and the nature of the problems to be solved (Amabile, 1996). For example, it has been found that innovators were not successful in analytical, more structured tasks in contrast to their colleagues with adaptive cognitive style (Pounds & Bailey, 2001). Although adaptors performed better at analytical tasks, innovators were more successful performing intuitive tasks (Fuller & Kaplan, 2004). All this implies that there is a possibility that different cognitive styles require different types of operational processes to generate higher productivity. Thus, we can expect that cognitive style is complementary to the operational process.

When it comes to the structural complexity of the job being performed, Kirton’s adaption-innovation theory proposes (Kirton, 1976) that adaptors prefer to operate within given paradigms and procedures. This type of individual follows traditional procedures and rules in their routines. Adaptors favor structure and a predictable work environment. There is even evidence that adaptors derive less enjoyment and intrinsic motivation from complex and challenging activities than innovators (Amabile et al., 1994). Management literature has established that adaptors scored significantly higher than innovators in structured and convergent problems with tight structure, well-defined aims and more precise ways of judgment (Kirton, 1991; Pounds & Bailey, 2001; Skinner & Drake, 2003). There is also a positive fit between adaptive cognitive style, relatively structured jobs and creative performance (Baer et al., 2003; Sagiv et al., 2010). Relying on Kirtons’ adaption-innovation theory (Kirton, 1976), and previous studies on structural complexity of the work process, we may expect that an adaptive cognitive style is complementary to the structured operational process, since adaptors find such job conditions more attractive and motivating. We thus propose that:

88

Proposition 1: Individuals with adaptive cognitive style will generate higher productivity when matched with structured operating process rather than flexible operating process.

In contrast to adaptors, innovators prefer to violate the agreed ways of doing things and are impatient with traditional rules (Kirton, 1976). Individuals with innovative cognitive style prefer to avoid structure at work and redefine problems (Kirton, 1991). Their performance is driven by the work challenge itself that fosters their intrinsic motivation (Amabile et al., 1994). Motivational orientation at work may be partially shaped by the work environment, particularly job characteristics (Amabile, 1996). For example, previously, it was established that complex and demanding jobs (high in complexity and autonomy) are predictors of increased intrinsic motivation in comparison with relatively simple jobs (Hackman & Oldham, 1980). Task-based intrinsic motivation emerges when an individual performs a task for its own sake because of his or her interest and enjoyment in the task (Amabile, 1996; Parker et al., 2001). There is also evidence that innovators tend to be more productive in complex and challenging jobs that are loose in structure, with not easily defined aims and without easy methods of judgment because of the enhanced intrinsic motivation that is a driving force of their behavior (Kirton et al., 1991; Skinner & Drake, 2003; Baer et al., 2003; Sagiv et al., 2010). Such intrinsically stimulating work mostly can be found in more complex and unstructured problems being solved. Therefore, based on the adaptive/innovative cognitive style literature and previously established evidence, we propose the following.

Proposition 2: Individuals with innovative cognitive style will generate higher productivity when matched with flexible operating processes rather than structured operating processes.

Thus, we propose that individual adaptiveness/innovativeness constitutes one key determinant for increased productivity as a complementary factor to the structural complexity of the production process. Specifically, we expect that matching adaptiveness/innovativeness to the specific type of production process (structured/flexible) increases individual productivity when a more aligned IT system is deployed and used.

89

3.3.2 Complementarities between cognitive style and training mode

In this research, we are particularly interested in a situation when a complex, non-trivial IT system is used to support the conduct of core activities by an information worker. When a more aligned IT system is used, there is a need for learning and training with regard to both IT use and mastery of the operational process. The use of new IT is a substantial challenge for managers and employees (Gallivan, 2003). In these conditions, information workers must constantly learn and assimilate new technologies and new ways of doing work. For example, information workers learn first in the implementation phase, but also later when new and modified functions are released. At the same time, information-intensive companies have to update their employees with job-related skills to meet their educational needs in order to drive individual productivity and organizational performance.

A set of studies provides evidence that work-related training is a strong complementary factor to productive IT use (Hitt & Brynjolfsson, 1997; Lynch & Black, 1998; Ichniowski & Shaw, 2003; Bartel et al., 2007). The main findings demonstrate that employees who are exposed to work-related education and training in a new IT system use it more properly and adequately than those who have not been exposed. Yet, these studies are all based on the assumption that employees learn and process information in a similar manner. Individual differences in cognitive style and predispositions towards particular learning formats are often ignored (Hayes & Allinson, 1997). Indeed, conventional training design methodologies in new technology usually miss that fact that differences in cognitive styles may determine learning performance, which in turn determines the mastery of technology. While every company intends to minimize costs of the production, benefits maximization from training programs may be one of the cost reduction factors.

In educational psychology literature and workplace learning literature, it is highlighted that there is a need for understanding individual differences before the process of employees’ selection, training and management change can be organized (Hayes & Allinson, 1997; Wynekoop & Walz, 1998; Sadler-Smith & Smith, 2004; Berings et al., 2005; Armstrong et al., 2012). There is also an argument that training may not provide all benefits when it is not connected to such an important factor as cognitive style since one of its characteristics is the way individuals learn (Honey & Mumford, 1982; Hayes & Allinson, 1997; McLeod et al., 2008). A certain cognitive (learning) style is demonstrated as not only influencing learning achievements but also being more appropriate for effective performance in specific work activities (Carland et al., 1994; Hayes & Allinson, 1997; Sadler-Smith & Smith, 2004; Berings et al., 2005). These findings point us to potentially important relationships between certain cognitive (learning) styles and teaching and educational strategies for

90

effective work performance. In particular, we can expect that when work-related training is customized to fit an individual’s cognitive style it can lead to more effective learning, which produces a higher level of mastery of work technology at hand. This, in turn, leads to a higher level of work productivity, all compared to no such fit between training mode and cognitive style. Yet, Armstrong et al. (2012) argue that at the present time, we have little knowledge or rationale about organizing training and education strategy with an orientation towards a particular cognitive style. Yet, the education literature and Kirton’s adaption-innovation theory help us specify propositions of a match between particular cognitive style and training strategies that are assumed here to increase individual productivity when a more aligned IT system is deployed and used.

Earlier empirical work established that adaptors and innovators subscribe to different achievement goals in learning that motivate them to work hard (Amabile, 1996; Ee, 1998). In particular, it was identified that individuals with adaptive cognitive style subscribe to ego approaching goals that are related to achieving particular scores in task performance and demonstrating their competence to others (Biggs & Moore, 1993; Ee, 1998; Ee et al., 2007). The behavior of such types of individuals is demonstrated as being driven by evaluating pressure and social comparisons (Amabile, 1996). Adaptors might by definition require more time to place a new technology in a suitable work framework because of their resistance to change (Kirton et al., 1991). Adaptive individuals may require more time to learn new technology and use it in an efficient way. This implies that when a more aligned IT system is deployed in the company, preliminary training activities can boost performance of adaptors since they can master new technology before performing their work activities.

Adaptors not only subscribe to ego goals but also have different preferences in the learning process. For example, adaptors learn in a sequential, detailed and linear mode (Kirton et al., 1991; Kirton, 1994). Adaptors also prefer to work with explicit rather than tacit knowledge (Bloodgood & Chilton, 2012). Therefore, adaptive individuals are less likely to explore features of new technology through the process of its use. Because of their particular learning style (achievement goals and preferences in the learning process), they are more likely to benefit when initial training in a more aligned IT system is provided together with an ability to receive constant support during the process of technology use. Hence, we expect that an initial training and education strategy is congruent with adaptive cognitive style and the way information is processed and learnt. Thus, we propose the following.

Proposition 3: Individuals with adaptive cognitive style will generate higher productivity with upfront comprehensive mandatory training in work technology and process rather than with optional on-demand training in work technology.

91

In contrast to adaptors, when it comes to the achievement goals innovators

prefer to subscribe to a mastery goal, namely becoming totally absorbed in the task and understanding it in depth (Ee, 1998; Ee et al., 2007). Innovators become intrinsically motivated by mastering a certain topic (Biggs & Moore, 1993; Ee, 1998). This type of individual is engaged in the work activity in a more flexible manner (Csikszentmihalyi, 1990). Innovators also favor working with tacit rather than explicit knowledge (Bloodgood & Chilton, 2012). This implies that when a more aligned IT system is introduced in a work context, innovators may not benefit in the same manner as adaptors from initial training and education.

By mastering a certain problem in a flexible manner, innovators may require continuous knowledge about new technology features. It has previously been established that innovators are more open to novel experience and radically different frameworks (Amabile, 1996). According to Kirton’s adaption-innovation theory (Kirton et al., 1991; Kirton, 1994), innovators prefer educational programs with loose structure with aims not easily articulated. This type of individual learns discontinuously and actively searches for new challenging ideas. Therefore, innovative individuals are more likely to explore features of new technology through the process of its use. Because of their particular learning style, they are more likely to benefit when constant support is provided during the process of technology use. Thus, based on educational psychology literature, workplace learning literature and Kirton’s adaption-innovation theory we propose the following.

Proposition 4: Individuals with innovative cognitive style will generate higher productivity with a combination of minor upfront mandatory training with optional on-demand training in work technology and processes rather than with upfront comprehensive mandatory training in work technology.

The above explicitly implies that there is a difference in learning activities and strategies between adaptive/innovative cognitive style, and consequently in demands for training programs. Individuals with polar cognitive styles, therefore, may respond differently to opportunities provided by professional development and training programs. Therefore, we expect that matching adaptiveness/innovativeness to the specific type of training and education activities (initial versus follow-up) individual productivity will increase when a more aligned IT system is deployed and used.

92

3.3.3 Complementarities between cognitive style, operational production and incentive modes

While training and education for mastering a new IT system and a new work process make up a significant factor of increased productivity, the incentive system is shown to be no less important since motivation is the major leverage for applying obtained knowledge and skills (Huselid, 1995). Indeed, early theory and research have argued that individual productivity of employees might vary in relation to work motivation (Amabile, 1996; Frey & Osterloh, 2002; Baer et al., 2003). In general, the importance of incentive as a powerful instrument for motivation, behavior change, performance, and output growth has been established by many researchers10. In particular, studies at firm and establishment levels demonstrate that incentives and reward practices can be strong complements to productive IT use (Hitt & Brynjolfsson, 1997; Black & Lynch, 2004; Aral et al., 2012b).

Over four decades, research has shown that performance of individuals may vary, depending on the level of motivation as well as on the orientation of motivation (Ryan & Deci, 2000). A large set of studies focused on the joint impact of extrinsic (incentives that come outside an individual, usually in the form of rewards) and intrinsic (incentives that come from task performance itself) incentives on creative performance of employees (Baer et al., 2003; Shalley et al., 2004; Kasof et al., 2007; Gupta, 2009; Eisenberger & Byron, 2011). Another set of studies considered the impact of a certain type of incentives on individual performance. For example, it was established that both extrinsic and intrinsic incentives can boost performance depending on task type, the way the task is presented and the kind of performance that is rewarded (Ryan & Deci, 2000; Deci & Vansteenkiste, 2004; Kasof et al., 2007; Eisenberger & Byron, 2011). A certain stream of studies highlighted that personal characteristics, particularly cognitive styles of employees (Amabile, 1996; Beecham et al., 2008; Gupta, 2009) and task complexity (Amabile, 1996; Baer et al., 2003) have to be taken into account before motivation incentives can be introduced. In a situation when a new IT system is used in the company, managers may consider the aforementioned factors in order to introduce incentives in a way that can increase employee’s productivity. We will focus below on the match between particular cognitive style and specific type of incentives in a situation where a more aligned IT system is used.

Early theory and research have argued that individuals with different cognitive styles differ in the extent in which they experience excitement and enjoyment from extrinsic and intrinsic incentives (Kirton, 1976, 1994; 10 See, for example, e.g. Skinner (1976); Deci & Ryan (1985); Harter (1978); Atkinson & Joel (1978); Amabile (1996); Ryan & Deci (2000); Deci & Vansteenkiste (2004); Kasof et al. (2007); Gupta (2009); Eisenberger & Byron (2011).

93

Amabile, 1996). In particular, Amabile (1996) argues that individual cognitive styles influence the way in which individuals respond to incentives. This implies that a certain match between cognitive style and incentives can have a positive effect on individual performance. Specifically, we argued earlier that individuals with adaptive cognitive style prefer work that is structured and predictable, and, as a consequence, tend to enjoy less the intrinsic motivation arising from less structured and challenging activities. These individuals experience enjoyment from recognition of their achievements and effort (Kirton, 1976). Thus, when adaptors are involved in a less structured operational process, intrinsic motivation cannot be a strong predictor of individual performance. For this type of individual, extrinsic rewards play the role of the intrinsic motivation (Amabile et al., 1994; Baer et al., 2003). This suggests a three-factor match between cognitive style, operational production and incentives modes. Therefore, we expect that in a situation where adaptors are involved in structured operational processes, external rewards such as bonuses will motivate them and increase their productivity.

Proposition 5: Individuals with adaptive cognitive style will generate

higher productivity when matched with the structured operating process and/or exogenous incentives rather than flexible operating process and/or endogenous incentives.

In contrast, employees with an innovative cognitive style prefer work that

is challenging and less structured (Kirton, 1976). Their behavior is primarily driven by challenging work itself and they are likely to find these properties in less structured operational processes (Kirton et al., 1991; Amabile, 1996). When innovators are involved in challenging and complex work tasks, they are in a state of cognitive stability (Baer et al., 2003). Consequently, a match between innovative cognitive style and a complex, unstructured process can be achieved, resulting in a high level of intrinsic motivation.

Intrinsic incentives such as constantly promoting feedback and a sense of autonomy (self-determined behavior) facilitate a high level of intrinsic motivation and, hence, improve individual performance (Ryan & Deci, 2000). Previous research has demonstrated that when individuals experience high levels of intrinsic motivation, they are immune to the effects of extrinsic incentives (Hennessey, 1989; Hennessey & Zbikowski, 1993). Innovators, in contrast to adaptors, value more intrinsic rewards (Amabile et al., 1994). Therefore, when innovators are placed in less structured operational processes with high autonomy and receive constant feedback, i.e. they are engaged in the internal reward program, they are expected to be more productive. In real conditions, companies combine both intrinsic and extrinsic motivators in order to increase individual performance. Yet, in terms of this study when a more aligned IT system and the production process are used, we polarize

94

external and internal incentives with regard to cognitive styles. On the basis of the aforementioned arguments, we expect:

Proposition 6: Individuals with innovative cognitive style will generate higher productivity when matched with flexible operating processes and/or endogenous incentives rather than structured operating processes and/or exogenous incentives.

Hence, it is expected that individuals with different cognitive styles will benefit differently from the structural complexity of the operational processes and different types of incentives. Incentive strategies matched to particular cognitive styles can generate greater productivity when a more aligned IT system is used. Therefore, in order to meet motivational needs for both types of employees such as adaptors and innovators, special conditions have to be taken into account when the motivation strategy is formulated.

3.3.4 Complementarities between cognitive style and decision-making mode

The final factor complementing our research model of individual information worker productivity is decision-making mode. In terms of this research, decision-making is understood here as a cognitive process resulting in a course of actions based on the preferences of the decision maker (Hunt et al., 1989). Earlier studies established that decision-making as a cognitive act has a significant impact on individual behavior and information worker productivity (Hunt et al., 1989; Rowe & Boulgarides, 1992). Although the distribution of decision rights is one of the most important directions of work organization in human resource management, in current turbulent working conditions, managers are faced with the question of how much work and decision rights have to be delegated to employees (Bloom & Van Reenen, 2011). On the one hand, decentralization of decision rights decreases the costs of information communication and transfer, increases the speed of response to market changes and may enhance productivity of employees through increasing job satisfaction (Zabojnik, 2002). On the other hand, decentralization of decision rights may increase costs due to replication of information and leads to a loss of control for the upper-level managers. This ambiguity in decision-making structure forces companies to mix centralized and decentralized forms of decision-making to achieve better results (Brickley et al., 2002).

In recent years, many companies, together with the implementation of new IT systems, have adopted greater decentralization of decision-making (Brynjolfsson et al., 2002; Tambe et al., 2012). However, current managers

95

often ignore various cognitive human resource needs and preferences in the decision-making structure development, their effect on productivity, and assume that all employees perceive decision-making structure in a similar way (Auh & Menguc, 2007). Yet, there is empirical evidence about the congruence between particular cognitive styles and preferred decision-making structure (Hunt et al., 1989; Rowe & Boulgarides, 1992). Later, Betsch and Kunz (2008) also came to the conclusion that when there is a match between preferred and applied decision-making structure, positive effects will emerge.

Decision-making structure is closely related to motivational aspects and operational autonomy/freedom (Donaldson, 2000; Locke & Latham, 2004; Hirst et al., 2011). Earlier, we developed propositions regarding different preferences among adaptors and innovators in operational processes and motivational aspects have been developed. So it is more likely that individuals with adaptive and innovative cognitive style may differ in response to whether a decision-making structure is more centralized or decentralized. Following those propositions that are theoretically and empirically grounded, we can expect that the degree of decentralization in decision-making complementary to cognitive style may create advantages that explain significant productivity variance between individuals when a more aligned IT system is used. The arguments for this assumption are provided below.

Previous studies suggest that different cognitive styles may have different motivators and needs in operational autonomy to perform work (Kirton, 1976; Baer et al., 2003; Ahearne et al., 2005). For example, adaptors prefer well-structured tasks with tight structure, well-defined aims, more precise ways of judgment and external rewards (Kirton et al., 1991; Pounds & Bailey, 2001; Skinner & Drake, 2003). According to Kirtons’ adaption-innovation theory, adaptors favor conventional rules, a sense of control, precision and discipline at work. This type of individual seeks to conform to accepted norms, solve problems within established organizational guidelines and prefers ongoing functioning of the organization (Kirton, 1976). Adaptors favor working within a given paradigm by following established ways and are less dependent on autonomy (Kirton, 1994). Individuals with adaptive cognitive style seek to maximize rewards and, therefore, receive through them a feeling of self-confidence and, based on this, achieve an understanding of whether their behavior is appropriate (Baer et al., 2003).

Centralized decision-making structure, in turn, provides clarity, predictability, precision and formality in work (Corti & Storto, 2000). When the decision-making structure is centralized, managers make decisions without employee involvement, expect employees to behave in a certain way and closely monitor their behavior (Deci et al., 1989). This structure of decision-making undermines intrinsic motivation and shifts employees’ attention toward the external concern (Deci & Ryan, 1987; Deci et al., 1989). Reduction in intrinsic motivation and focus on external motivation is then expected to enhance individual productivity of individuals with adaptive

96

cognitive style. Therefore, it is more likely that there is a fit between adaptive cognitive style and more centralized decision-making structure. Thus, we propose the following.

Proposition 7: Individuals with adaptive cognitive style will generate higher productivity when their cognitive style is matched with centralized decision-making rather than with decentralized decision-making.

In contrast, innovators, by definition, favor a sense of non-controlling and undisciplined conditions in their work (Amabile, 1996). This type of individual resists bureaucratic pressure (Kirton, 1994). Innovators, due to their cognitive nature, prefer to be intrinsically motivated by challenging and complex tasks because this provides greater involvement and self-confidence in the working process (Skinner & Drake, 2003). Innovators experience job satisfaction through high operational autonomy/freedom (Kirton, 1994). For example, Oldham and Cummings (1996) found that more innovative employees scored higher creative performance when they worked with more complex tasks and were supervised in non-controlling, supportive fashion.

When a decision-making structure is decentralized, employees are encouraged to be involved in the process of decision-making (Deci et al., 1989). This type of decision-making structure fosters operational autonomy/freedom and responsibility for independent action (Corti & Storto, 2000). Operational autonomy is the extent to which employees can determine work procedures to be followed, including sequence, pace and methods for accomplishing their tasks (Hackman & Lawler, 1971; Shalley & Perry-Smith, 2001). It was shown that a decentralized decision-making structure may enhance employees to complete more complex tasks with enthusiasm, since it gives them essential freedom and, thus, fosters intrinsic motivation (Deci & Ryan, 1987; Amabile, 1996; Hennessey & Amabile, 2010). A decentralized decision-making structure is expected to promote personal initiative at work by increasing interest in work activities (Oldham & Cummings, 1996; Axtell et al., 2000). It was established that when individuals are able to choose which tasks they will complete and how much time they can allocate to each task, they are more intrinsically motivated than individuals not given this choice (Zuckerman et al., 1978). We therefore expect that decentralized decision-making structure that promote intrinsic motivation may increase productivity of employees with innovative cognitive style.

Proposition 8: Individuals with innovative cognitive style will generate higher productivity when their cognitive style is matched with decentralization of decision-making rather than centralized decision-making.

Thus, since different cognitive styles have different preferences in decision-making structure, we may expect that the degree of decentralization

97

of decision-making complementary to cognitive style will generate greater productivity.

Given the above, we identified a number of factors that were previously researched in theoretical and empirical studies. These factors are expected to complement each other when an IT system is used. Based on the systems approach of the complementarity theory (Ennen & Richter, 2010) we expect that complementarities differ with regard to cognitive style, the structure of the operational process and human resource management practices, including training and education, incentives and decision-making structure when a more aligned IT system is used. Particularly, we expect that individuals with adaptive cognitive style will generate greater productivity when matched with structured operating process, push mode training in work technology, exogenous incentives, and centralized decision-making, compared to other configurations of these factors. In contrast, individuals with innovative cognitive style will generate greater productivity when matched with flexible operating process, a combination of minor upfront mandatory training with optional on-demand training in work technology, endogenous incentives, and decentralized decision-making, compared to other configurations of these factors. We chose these specific factors and matched them based on earlier theoretical and empirical research that has already established double or triple interrelations between them. We are aware that unobserved factors could mimic the complementary relationships. Yet, the chosen set of complementarities decreases model overload and further steps in empirical research. In the next chapter, we present the methodology of this research applied to test the formulated research model.

98

99

4. Research methodology

In this chapter, we provide a description of the research methodology utilized here in order to test the research model presented in the previous chapter. The chapter begins with a background description of the rationale of the chosen research approach. After some details from related previous studies, the pro-cess of data collection and procedures that were undertaken to test comple-mentarities of productive IT use are described. Concepts of validity and relia-bility are also addressed here. Finally, the chapter ends with a discussion of methodological limitations and the steps undertaken to ensure compliance with the research ethical norms and principles.

4.1 Overview and rationale for research approach We tested the formulated research model by conducting two independent empirical studies: a longitudinal quasi-randomized field experiment and a well-controlled online experiment. The choice of the research methodology and its rationale for the present research is explained in more detail below.

Most of the studies identifying the relationship between IT use and individual productivity have been based on survey technique. This technique was also fundamental for studies of individual IT complementarities (Athey & Stern, 2002; Autor et al., 2003). One explanation for the use of surveys could be that the study of IT-enabled productivity at individual level has always been complex, as both human and technical issues are interrelated in information work. Yet, this technique is criticized for respondent bias and low confidence in the results (Podsakoff et al., 2003). As the number of information workers increases and as the field of IT-enabled productivity matures, there is an increased demand for objective data and empirically validated results (Petter et al., 2008). Moreover, the existing literature in the area of IT complementarities and the current state of knowledge in IT-enabled productivity demonstrate that both longitudinal and experimental studies are the most appropriate methods for studying the impact of complementarity factors on IT-enabled productivity (Brynjolfsson & Milgrom, 2013).

100

Our research model has also key characteristics that impact the research process. First, the research model was formulated to identify the impact of changes in operational set-ups (changed IT systems and complementary factors) on individual productivity. Particularly, we needed to discover whether this treatment (changes in operational set-ups) is causing productivity differences due to different configurations of complementarity factors. Hence, we needed to establish different operational set-ups to identify their impact on individual productivity. Therefore, experimental design in general was appropriate for testing the research model formulated in this research.

Second, the formulated model took into account the time lag effect on productivity gains from IT use. This implies that a longitudinal research approach was necessary to capture productivity data over a period of time and demonstrate the effect of changes on productivity. In general, longitudinal field experiments allow (i) the study of a targeted phenomenon in its natural settings without artificially introducing confounding variables, (ii) the study of the effect of intervention over time, (iii) the clarification of magnitude and direction of change among variables (Hassett & Paavilainen-Msntymski, 2013). Therefore, based on the characteristics listed above, longitudinal field experiments are necessary and important when studying complementarities of productive IT use to demonstrate their effect over time.

Third, and in addition to the time lag effect, we also needed to provide better control over complementarities and their impact on IT-enabled productivity. This could be achieved by conducting well-controlled experiments to support the results obtained from field experiments since in contrast to field experiments, laboratory experiments provide better control and more accurate results (Camerer & Weber, 2013). However, in comparison with traditional laboratory experiments, recent online experiments have become even more popular since they reduce demand characteristics (the influence of the experimenters’ expectancies on the participants’ behaviour) and increase generalizability due to access to wider populations (Buchanan, 2002). The Internet also allows access to the targeted population in a fast and convenient way (Birnbaum, 2004). Reips (2002) highlights that, unlike laboratory experiments, online experiments (i) reduce costs and the amount of time spent on the experiment, (ii) increase the uniformity of the experimental procedure across participants, (iii) increase participants’ comfort due to 24-h access, (iv) increase ethical standards, since they are available to public and criticism. In general, the results from online problem-solving experiments are identified as consistent with traditional laboratory experiments (Dandurand et al., 2008). Therefore, both a longitudinal field experiment and a well-controlled online experiment respond to the research model being tested.

Fourth, our model was specifically developed for information workers. We chose two information-intensive professions, sales representatives and software programmers, as appropriate examples of information workers (Lal, 2005; North & Gueldenberg, 2011). Although the nature of work for these

101

professions is different; for example, sales representatives mostly meet customers and generate sales, and software programmers write programming codes; both professions require cognitive skills to process information which is an input and output of the production process. Moreover, both sales representatives and software programmers use non-trivial IT systems as their main production tool. Therefore, we tested the research model in two independent studies with heterogeneous information-intensive professions to increase model validity and provide its applicability across different information-intensive professions.

In the light of the theoretical background and formulated research model, a longitudinal field experiment, and a well-controlled online experiment helped us clarify the potential effects of both individual cognitive differences and corresponding complementarity set-ups on the individual productivity of information workers. Following those premises, the research strategy was based on two independent studies: a longitudinal quasi-randomized field experiment where we investigated the operational productivity of sales representatives and an online experiment of software programmer productivity. Figure 4.1 illustrates how the theoretical framework influenced the development of the hypotheses and empirical testing.

102

Note: P – proposition, H – hypothesis

Figure 4.1: The evolutionary process of the methodology of inquiry

103

Based on the existing literature on IT-complementary factors, the research model and two main hypotheses were developed. The main hypotheses were developed based on a number of propositions that were derived from studies that investigated individual factors or their double or triple interactions (Chapter 3). The research model was tested in two parallel studies. First, the model was tested in a longitudinal quasi-randomized field experiment, where we investigated sales operations and sales performance. Second, an online experimental study of software programmer productivity was carried out. Two studies were conducted to provide a rigorous and comprehensive investigation of the developed hypotheses so that both internal and external validity11 could be addressed (Cook & Campbell, 1979).

Although the two studies targeted very different contents –pharmaceutical sales and software programming – they both targeted information workers where IT systems are introduced jointly with individual and organizational complementarities. Both approaches are considered appropriate as they enable in-depth exploration and learning about the inquired phenomenon. The reason for two independent studies can also be explained by the need to identify whether the emerged patterns in one study are confirmed in the other, and thereby aspire to stable results. Each empirical study is presented below in accordance with the structure and content recommendations given by the American Psychological Association (2009).

4.2 Study 1: Sales representative productivity A longitudinal quasi-randomized field experiment (Study 1), investigating sales operations of an international pharmaceutical company, was conducted in one organization with multi sub-case set-up to assure closeness to reality. The intention of the longitudinal quasi-randomized field experiment of sales representative productivity was to test the research model’s feasibility. We achieved this by means of testing several configurations of factors as prescribed in the model. Three patterns were researched in order to identify configurations of complementary factors that influence IT-enabled productivity of sales representatives at the individual level: (i) the introduction of an IT system only, (ii) the introduction of an IT system together with partially assumed complementarities, and (iii) the introduction of an IT system jointly with a “full” set of complementarities. We hypothesized that the third

11 Internal validity – confidence that the treatment caused the difference in dependent variables. External validity – the extent to which the obtained results can be generalized to other people or settings.

104

alternative will produce the highest productivity increase, considering other circumstances being equal.

The approach for conducting an empirical study of sales representatives’ operations, how they use IT and how it affects their productivity, started by addressing various dimensions of information work such as work activities, resources and actors, rules and goals, inputs and outputs, information channels, a content of the information and kinds of IT utilized. Next, the current impact of the IT use on work process, resulting indicators, and applied measures of sales representatives’ productivity was investigated. We also identified current practices for information processing patterns by sales representatives. Therefore, we gathered large volumes of data, representing a long period of time, e.g. four and a half years (nine quarters before and nine quarters after the introduction of a more aligned IT system), which characterizes the actual performance of the targeted processes. A longitudinal study was considered as the most appropriate way to study IT-enabled productivity at the individual level in order to minimize a potential impact of negative reactions towards the novelty of the IT system being implemented and time lag required to learn a more aligned IT system.

4.2.1 Research settings

The study has been conducted in a Nordic affiliate of a global pharmaceutical company that is among the 100 largest life science corporations in the world. The company has its headquarters in the USA and operates in nearly 150 countries and had about 70000 employees globally at the time of the study’s inception. The company represents the so-called fully integrated life science company with patent protected drugs, meaning that there is a full value chain, including research and development, manufacturing, logistics, marketing, sales and aftermarket. Patented drugs imply that the company offers the market new original drugs that are protected by patents. Since this study investigates operations in the pharmaceutical business, the other parts, i.e. nutritious, medical devices, veterinary products, are disregarded as they have their own operational and organizational set-up.

The Nordic affiliate operates with its own organization and resources in four markets, including Denmark, Finland, Norway, and Sweden. The Nordic affiliate’s pharmaceutical business is organized as follows. At the top, there is a Nordic managing director who reports to the European Headquarters of the company, located in London, UK. The affiliate has four market organizations, constituted by marketing and sales capabilities, for each country market and one shared services organization which includes the Human Resource Department, Finance Department, Information System Department, Sourcing

105

Department, Health Technology Department, Communication and Public Relations Department, Regulatory Department and Medical Department.

As with any major pharmaceutical company, the company under investigation has a set of key challenges:

• How to bring forward new innovative products? • How to launch new products into the market in a successful manner? • How to maximize the patent protection of each product? • How to maximize the sales performance of products introduced?

In terms of this study, the challenge of how to maximize the sales

performance of products introduced is the main concern. As mentioned above, the company has a very large salesforce. This means that the salesforce represents a very significant cost component in the cost structure of the company. The question posed then is: how can the productivity of the sales force be increased?

Several solutions can be implemented to increase the performance of the sales operations, including:

1. market access (payers, governments in Europe); 2. patient awareness and demand; 3. health care center management (a clinic’s priority); 4. marketing and sales efficiency toward prescribers (physicians).

In this context, challenge number four will be addressed. Today, the

company has the following choices with regard to the improvement of the sales operations productivity. First, the company can decrease the number of sales representatives and thereby costs, while altering operations to maintain the same level of sales, hence productivity gain from decreased resources. Second, the company can increase the performance of the sales representatives and thereby the return on investment. Third, the company can use both approaches. In the present case, the Nordic affiliate, with its four Nordic markets has been exposed for the second approach, i.e. to increase the performance of the current sales operations/sales representatives. Therefore, at the end of April 2014, in order to achieve this objective, the company introduced new sales processes: a new sales method (Customer Focused Sales), and a new sales support system (Customer Relationship Management system), which is a dedicated computerized information system that supports the execution of sales operations and sales management. These two components were introduced globally (to all markets/countries) together with training and education kits. Each market was given free hands on how to introduce both components. Below, we describe how the overall process with less aligned IT system support was conducted.

106

• The overall process with less aligned IT system support

Before the introduction of a more aligned IT system, the company used an old version of a Customer Relationship Management (CRM) system to support sales representatives in their activities. The system was mostly used as a repository of customer information where prioritization of customers could be conducted. The system also allowed the employees to schedule meetings with customers and report output from customer interactions. In the company, old sales practices (before the introduction of a more aligned IT system) were organized in the following way (Figure 4.2).

Figure 4.2: The overall process with less aligned IT system support In general, the sales process consists of three phases: planning, execution,

and evaluation. However, before the introduction of the process business unit manager provided the main direction of activities to the sales managers in terms of priority of customers (by region, relation to the company, volume of patients) to be targeted by sales operations. Product managers provided sales representatives with information about the product to be delivered to the customer such as indication, efficacy, side effects, usability, underlying studies and competition data of competing products.

Sales representatives explored the given sales material from product manager, read various additional studies on clinical research and sales pitching and arguing. Each sales representative prepared a sales plan for the time period in terms of which customer to visit and when. The plan was checked and discussed with the sales manager and, after some modification, approved. After that, a sales representative initiated booking sales call meetings with customers (sometimes booking was outsourced to specialized call booking firms). Typically, a day before the planned sales calls, sales representatives checked the schedule, address and time slots in the pre-existing IT system. After the actual meetings were conducted, the sales representatives reported the outcome of the meeting in the IT system briefly and in an unstructured manner.

107

At the end of each week sales representatives emailed the sales manager the outcome of the past week (number of calls conducted versus planned, how many calls were spontaneous, how customers reacted to the message provided). The sales manager compiled the information from sales representatives into a report and delivered it to the business unit manager. Approximately once a month, the business unit manager, sales manager, and product manager held meetings to review how the sales activities were conducted. At the end of each semester, the conducted sales and marketing activities were assessed in terms of plan versus realized and also in relation to the sales volume of the product. All this information was used as an input for the planning of the next time period.

Pre-existing sales practices and their IT system support are summarized in Table 4.1.

Table 4.1: Pre-existing sales practices and its IT system support Level of management

Practices IT support provided

Sales strategy

Formulation of general strategies for sales and prioritization

No

Follow-up of general strategies for sales and prioritization

No

Sales operations

Search for customers Storage of basic customer information such as name, specialty, education, title, institution, addresses, phone calls, emails, sales representative responsible.

Prioritization of cus-tomers

An ability to prioritize customers by as-signing priority such as 1, 2, 3 and so on.

Performance of cus-tomer calls

An ability to schedule meetings with a given customer and to access customer information prior to the meeting.

Reporting outcome from the customer call

An ability to generate and store basic notes from conduced sales call meetings in an unstructured format.

108

As depicted in Table 4.1, formulation and evaluation of general strategies for sales and prioritization were conducted without IT support. The less aligned IT system was used by sales representatives to search for information about customers, including name, specialty, education, title, institution, addresses, phone calls, emails, and sales representative responsible. Sales representatives could also prioritize customers by assigning priority. The IT system was also used to schedule meetings with customers and to access customer information prior to the meeting. At the end of meetings, sales representatives could create notes about the results of the sales calls in an unstructured manner. Below, we demonstrate how more aligned IT system support has changed the overall sales process.

• The overall process with more aligned IT system support The more aligned IT system was introduced to support sales strategy and

operations management. In Figure 4.3 we show a circular process for the evaluation and learning from business activities together with more aligned IT system support, which consists of planning, execution and evaluation phases.

Figure 4.3: The overall process with more aligned IT systems support

After the introduction of a more aligned IT system, the planning phase starts with the formulation (or update) of the business unit strategy by a business unit manager, product manager and sales managers based on sales targets and the global product strategy. When the business unit strategy has been formulated, it is entered into the IT system, where specific information (sales targets, decomposition of sales per customer type and region, activities in terms of sales calls and marketing) is detailed. The business unit strategy is decomposed into a marketing strategy with a particular focus on marketing activities to be conducted (what, when, where, to whom). All these details are

109

registered into the more aligned IT system. The business unit strategy is also decomposed into a sales strategy for the market. This strategy includes information about target customers, how to approach them and with how many calls. The sales strategy also covers various characteristics of each customer. For example, whether or not a customer is easy to access or is a key opinion leader, actual versus potential sales, professional profiles, and preferred access channels. At this stage, the IT system checks whether articulated goals, activities, and resources at the business strategy level are consistent with those of the marketing and sales strategies.

Business unit strategy, marketing strategy, and sales strategy plans provide a basis for formulation of the customer interaction plan (CIP). CIP is a long-term (1-2 years) plan for each targeted customer in terms of a so-called sales ladder (which, briefly regarded, articulates the assumed situation of a specific customer, in such terms as customer unawareness, knowing about, or interest in a product) and in relation to this ladder, the objectives set for each customer. CIP for each targeted customer is developed by sales representatives and has to be checked and approved by a sales manager. After that, information from each CIP has to be registered in the IT system.

After the stage of the CIP formulation and registration of information into the IT system, all staff is ready for the next phase –an execution of marketing and sales plans and our focus here is on the sales operations. This phase consists of (i) preparation of customer interaction, (ii) execution of customer interaction, and (iii) reporting customer interaction. Sales representatives book sales call meetings (in some instances sales calls are booked by an external booking agency) and prepare these meetings, typically, 3-5 days before the interaction stage by using information from the IT system. The IT system provides sales representatives with information about how, when, and where to meet, what the message and goals for this particular call are, and what additional pre-reading has to be done. For example, one customer may not be aware of the advantages of the product promoted, and her particular interest is in adverse event profile, hence the sales representative prepares such a message while another customer may be an advanced user of the product promoted and is interested in studies and recommendations on the drug interaction effects.

The next stage is the actual conduct of the sales call, i.e. the interaction between the sales representative and the customer. Thereafter, sales representatives report information from the meeting into the IT system. The IT system checks the consistency between planned versus realized objectives, outcomes, activities conducted. All information after the interaction phase is reported automatically and in an aggregated manner to the sales manager on a regular basis (once a week) and on-demand. Eventually, after reporting of customer interaction, the evaluation of each CIP, sales strategy, marketing strategy and business strategy plans is conducted to identify gaps between

110

planned and actual strategies. All this constitutes an input for the next cycle of planning.

The so-called pre-active customer interaction has been described above. In this situation, customer interaction is initiated by the company’s representative. However, a situation may occur where customer interaction is initiated by the customer; this is known as re-active customer interaction. When a sales representative receives a call from a specific customer, information about this particular customer can be entered into the IT system. The IT system has a specific function called CIG (Customer Interaction Guide). This guide provides general information about how any sales representative should interact with a customer when they meet spontaneously, and what kind of information can be delivered to this customer. All spontaneous interactions must be recorded into the IT system.

New sales practices together with a more aligned IT system support are summarized in Table 4.2. Table 4.2: New sales practices and functions supported by the more aligned IT system Practices IT support provided

Pre-active customer interaction

Strategic customer planning

The more aligned IT system enables: - to formulate Customer Interaction Plan for: (i) long term objectives in terms of the Customer Focused Sales stages such as: • No-Interest • Low interest • Strong interest • Convinced • Committed (ii) further activities in terms of: • Goal(s) • Action(s) • Category • Type • Status - to find and access customer information by the

first and the last name, institution, country and city, etc.

111

Table 4.2: Continued from previous page

Practices IT support provided

Pre-active customer interaction

Pre-call planning

The more aligned IT system enables: • creation and booking of a customer call. Based

on this information, a calendar provides an over-view of dates and scheduled calls;

• to carefully prepare customer call specification (interaction) with a focus on previous interac-tions, the main objective of current interaction, measures of outcome and activities;

• to include the following details: promotional items, samples, issues, document attachments, follow-up activities.

Post-call reporting

The more aligned IT system enables: - careful reporting of executed customers calls

with the following details:

• planned goal and action; • realized coal and action; • goal/action lessons learned;

- to show which customer call is not reported yet; - to include additional details such as promotional

items, samples, issues, document attachments, follow-up activities.

Re-active customer interaction

Re-active planning and reporting (spontaneous contacts)

The more aligned IT system enables planning and re-porting spontaneous contacts. Such contact has to be reported into the IT system.

Strategic follow-up and evaluation

The more aligned IT system is used for strategic fol-low-up and evaluation.

The more aligned IT system is not simply a repository of customer

information. It helps employees prepare work and analyze data, alerts sales representatives about customer status enabling them to operate in an appropriate and recommended way. The system provides detailed information about customers and activities executed with each customer. With the more aligned IT system, new customer contacts can be carefully planned and booked with a particular focus on place, time, objective, activity, outcome, and history of previous interaction. Moreover, the system enables reporting of

112

an executed customer call in relation to the planned goal, realized goal, and lessons learned. At the same time, the system shows which executed customer calls have not been reported yet. In addition, the more aligned IT system enables planning and reporting of spontaneous contacts by providing guidelines for how the interaction can take place.

Therefore, the more aligned IT system can be used to improve sales performance, enhance communication and collaboration, and most importantly, to increase the productivity of sales representatives. As the main activity of the sales company is to sell products, the motivation for implementing the system was to sell more, increase revenue and increase market share. The situation when a more aligned IT system was introduced in the company offered us a unique opportunity to test the developed model in real conditions. A detailed description of how the study was performed is presented below.

4.2.2 Conceptual set-up and complementarity configurations The more aligned IT system was specifically designed for sales representatives to facilitate their daily work and store necessary information about customers. This IT system was installed in the company with various operational-organizational configurations. This study investigated sixteen distinct business units within the Nordic affiliate. We studied four products12 (A, B, C, and D), where each product was marketed by a dedicated business unit, for each market/country (Denmark – DK, Finland – SF, Norway – NO, and Sweden – SE) that gave sixteen business units. These sixteen units were selected on the basis of being different products, having a large enough sales force and being available for study. The affiliate also had additional business units, however they were not included in this study.

The study investigated four different designs or set-ups (configuration of variables) where we assumed that everything else was the same, both within the firm and in the market (Figure 4.4). These designs were introduced in the company in such a way as to ensure that performance differences observed are due to differences in the developed designs, and not due to particular products and markets. The four different design patterns were allocated into four-by-four operational structure of the firm, i.e. with four different products and four different markets, so as to neutralize the significance of the product as such,

12 Product A was launched in 2007 and had three competitors. Product B was launched in 2009 and had eight competitors. Product C was launched in 2009 and had two competitors. Product D was launched in 2011 and had five competitors.

113

as well as the market as such, with regard to their significance on sales performance.

Figure 4.4: Study design: conceptual set-up

A description of the developed designs (complementarity set-ups) is pre-sented below (Table 4.3).

Participants in the first group (Design 1) acted as a control group and operated in an unchanged mode (i.e., in the way the whole organization operated prior to the introduction of the more aligned IT system). Twenty-one participants were involved in this design.

Participants in the second group (Design 2) received the more aligned IT system, yet maintained the same operational set-up as prior to the change (i.e., as in Design 1). Sales representatives received initial (mandatory) training and education. They also had an opportunity for follow up (on-demand) training and education. Sales representatives had the same incentives as in Design 1 such as a bonus system with a maximum up 6 months’ salary on realized sales objectives. Decision-making authority was split in such a way that key deci-sions were made by the sales manager and minor decisions were made by sales representatives. Nineteen sales representatives were involved in this design.

Participants in the third group (Design 3) received the more aligned IT system together with a new and specific type of sales process. Since cognitive style was disregarded, customers were assigned randomly. Yet, in contrast to previous designs, obligation to follow the process was not mandatory. Sales representatives received initial (mandatory) training and education and had an opportunity for follow up (on-demand) training and education. Sales representatives had a bonus system with a maximum up 6 months’ salary on realized sales objectives. Decision-making authority was split in the same way as it was in Design 2, i.e. that key decisions were made by the sales manager and minor decisions were made by sales representatives. Twenty-four sales representatives were involved in this design.

114

Table 4.3: Proposed designs/complementarity set-ups

Key characteristics Pre/existing operation

set-up

Design 1 (No change)

Design 2 (Structured

partial change)

Design 3 (Semi-

structured partial change)

Design 4 (Full change)

IT system Less aligned Less aligned More aligned More aligned More aligned More aligned

Cognitive style Disregarded Disregarded Disregarded Disregarded Adaptors Innovators

Organizational factors

Operational process

Customer allocation

Customers as-signed ran-domly to sales representa-tives

Customers as-signed ran-domly to sales representatives

Customers as-signed ran-domly to sales representatives

Customers as-signed randomly to sales repre-sentatives

Easy to access cus-tomers assigned to adaptors

Difficult to access customers as-signed to innova-tors

Obligation to follow the process

Obligation to follow each step in the process is mandatory

Obligation to follow each step in the pro-cess is manda-tory

Obligation to follow each step in the pro-cess is manda-tory

Obligation to fol-low each step in the process is non-mandatory

Every step is man-datory

Every step is non- mandatory

Training and education No No Mandatory and On-demand

Mandatory and On-demand

Mandatory and On-demand

On-demand

115

Table 4.3: Continued from previous page

Key characteristics Pre-existing

operation set-up

Design 1 (No change)

Design 2 (Structured

partial change)

Design 3 (Semi-

structured partial change)

Design 4 (Full change)

Incentives External External External External External External and Internal13

Decision-making structure

Centralized Centralized Centralized Centralized Centralized Decentralized

Productivity Primary Number of

calls Number of calls

Number of calls

Number of calls

Number of calls

Number of calls

Secondary Number of products sold

Number of products sold

Number of products sold

Number of products sold

Number of products sold

Number of products sold

Number of participants All participants 21 19 24 27

13 The local Worker Unions did not accept to remove bonuses from innovators, as it was considered as unfair.

116

In the fourth group (Design 4), the “full”, or comprehensive set of complementarities was assumed, based on the above-formulated hypotheses. Participants in this design were conceived in terms of adaptors and innovators and received different complementarity set-ups according to their cognitive style. First, sales representatives with adaptive cognitive style were assigned to customers that were easier to access, unlike the other customer group that was harder to access. Yet, the obligation to follow each step of the process was mandatory. Sales representatives with adaptive cognitive style received initial (mandatory) training and education. They also had an opportunity for follow up (on-demand) training and education. Sales representatives had the same incentives as in Design 1 such as a bonus system with a maximum up 6 months’ salary on realized sales objectives. Decision-making authority was split in such a way that key decisions were made by the sales manager and minor decisions were made by sales representatives. Twenty sales representatives had adaptive cognitive style. Second, sales representatives with innovative cognitive style were assigned to customers that were difficult to access. Yet, the obligation to follow each step of the process was non-mandatory. These sales representatives were only offered follow up (on-demand) training and education. A bonus system with a maximum up 6 months’ salary, on realized sales objectives, was used as before. However, sales managers were instructed to provide extensive feedback to the innovators with regard to sales representatives work activities and their outcomes, a feedback that was much more extensive than conventional feedback. Decision-making authority was split in such a way that key decisions were made by sales representatives and minor decisions were made by sales managers. Seven sales representatives had innovative cognitive style. Therefore, in total twenty-seven sales representatives were involved in Design 4. In the next sub-section, we shall describe how the main variables in the study were defined and operationalized to provide robustness and replicability of the research design.

4.2.3 Measurement Several steps have been undertaken to develop measures required by the research design. The use of a more aligned IT system included the introduction of new sales processes and a set of human resource management practices which were found to be strong complementarities of effective IT use. Organization factors included four parameters such as structural complexity of the operational process, presence and type of training and education, incentives, and decision-making structure. Therefore, this sub-section includes a detailed description of independent, dependent and control variables.

117

• Independent variables

The operationalization of independent measures such as IT system, KAI-adaptive/innovative cognitive style, the structural complexity of the operational process, types of training and education, incentives and decision-making structure is presented below.

1) IT system In this study, we distinguished between less aligned (pre-existing) IT

systems and more aligned (new) IT systems. As we mentioned before, pre-existing IT systems were largely used to support sales operations. Sales representatives mostly used the system as a repository of customer information and reported output from sales calls in an unstructured manner. The more aligned IT system enabled the following functions:

• strategic customer planning by formulating CIP, finding and accessing customer information; • pre-call planning by creating and booking a customer call; • post-call reporting by entering detailed information of an executed cus-tomer call; • planning and reporting spontaneous contacts; • strategic follow-up and evaluation. The observation of these IT system measures was as follows. First, we

received documentation with the description of two IT systems. Second, we received screenshots from the more aligned IT system where some of the key functions were presented. We also monitored the use of the less and more aligned IT systems and received log data from the more aligned IT system.

2) KAI – Adaptive/innovative cognitive style Kirton Adaption-Innovation Inventory (KAI) (Kirton, 1976) has been used

to assess employees’ cognitive style (Appendix A). Over the past decades, the KAI has demonstrated acceptable psychometric properties (Brown, 2001). This measure of adaptive/innovative cognitive style has been shown as highly reliable and valid (Bagozzi & Foxall, 1995). The KAI is a 32-item instrument with a 5-point scale in which respondents indicate on a scale ranging from “Strongly disagree” (1) to “Strongly agree” (5) to present themselves as a certain type of person. The inventory is based on three subscales – sufficiency of originality, efficiency, and rule/group conformity. An overall KAI score is derived from a sum of all 32 statements and a theoretical mean of 96. The scores can range from 32 to 160. Individuals with an adaptive cognitive style score in the 60-90 range. Individuals with an innovative style score between

118

110 and 140. Usually, the scale is scored so that adaptors are lower and innovators are higher than the theoretical mean.

3) Structural complexity of the operational process The sales process typically consists of different steps, including selection

(targeting), booking calls, preparation, execution, and reporting results. The pharmaceutical company developed two different sales processes by taking into account their structural complexity. Two major criteria have been taken into consideration: (i) ease of access to customers and (ii) obligation to follow the process (process flexibility).

The company polarizes the ease of access to customers from “easy to access” to “difficult to access”. The following notion of the “ease of access to customers” was supplied by the company. If a physician acts as a Key Opinion Leader (KOL) then she or he is typically harder to access, i.e. to meet face-to-face for a sales call. The reason for this is that a KOL frequently works only part time with clinical work, i.e. meeting patients, being also engaged research and/or advisory work (to government or companies), because of superior competence. This has a dual effect. One the one hand, it is important for any company to “win” a KOL as she or he influences both formally and informally the choice of products made by other physicians, meaning that all competing companies wish to access the KOL. On the other hand, the KOL spends only part of his or her time on clinical work, which is the forum for dialogue between prescribers and sales representatives, hence, less time is available for meeting different sales representatives. The KOL knows about his or her position versus the sales representatives and is more restrictive in conducting meetings with them. Therefore, sales representatives who are assigned to KOL are involved in higher process complexity. Sales representatives who are assigned to standard prescribers – so-called “easy to access customers” are involved in a less complex sales process.

The structural complexity has another criterion which is the obligation to follow the process. The company distinguishes between mandatory (i) step-by-step process performance with a more structured process versus (ii) do-as-you-wish with one mandatory step – sales call report and thus a less structured production process. For this reason, this study uses the definition of the structural complexity of the production process received from the company. The structured process is a process by which sales representatives are assigned to standard prescribers and follow step-by-step process performance. The flexible process is a process where sales representatives are assigned to KOL and follow a “do-as-you-wish” process with one mandatory step – the sales call report.

The observation of the operational process was as follows. First, we received a process specification from the company detailing the activities, their order, and conditions. Second, in the more aligned IT system, we also

119

observed the number of logs conducted in this IT system, which are related to the activities of the process.

4) Training and education

The use of a more aligned IT system and the operational process may be

complemented by training and education in order to achieve productivity gains. A new IT system may require salespeople to acquire new skills on how to use it, and to learn about the new operational process. The company provided their sales representatives with two types of training and education. Figure 4.5 illustrates how training and education activities are organized in the company.

Figure 4.5: Types of training and education in a pharmaceutical company Mandatory training and education activities consisted of initial training in

the more aligned IT system and the operational process. On-demand training and education provided employees with further online and phone support on-demand after this IT system was implemented.

We had access to several kinds of training and education observations. First, we received a detailed specification of the training and we agreed with this specification from the company (Appendix B). We also received documentation of the educational package. In addition, the company supplied data of which employee received which training.

5) Incentives

As stated in the previous chapter, motivation can have two principal sources, one that is external and one that is internal to an information worker. We operationalized incentives in the following way. Sales representatives that

120

were motivated by external incentives received a salary bonus program by which, dependent on the fulfillment of pre-defined sales objectives, each sales representative could receive six months’ salary on an annual basis on top of the basic or fixed salary. The bonus realization was conditioned by the degree of fulfillment of the sales objectives set, implying that the more the objectives were fulfilled, the more bonus was awarded to the representatives, with a maximum of the equivalent of six months’ salary per year.

A second complementarity set-up for inducing worker motivation addressed the need to provide internal, or endogenous motivation. This was operationalized by allocating more sales managers’ monitoring and coaching time to sales representatives. More specifically, each sales representative who was exposed to internal incentives had on average three times more frequent contact with his or her sales manager; this included, on average, one physical meeting per week instead of one per month (such as co-traveling, co-detailing at a customer site, or a one-to-one meeting), and on average three to four telephone calls per week, instead of one telephone call. During these meetings, sales objectives and sales activities were discussed, including problem solving. Those sales managers that were selected to provide these measures received additional coaching training from the company’s human resource department and were instructed to set objectives that constituted professional stimulation to provide professional feedback on the performance outcomes.

The observation of these motivational measures was as follows. First, we received preliminary documentation from the company about the detailed incentive program. Second, we informed the company about which sales representatives have to be exposed to internal and external incentive program.

6) Decision-making structure

An objective measure has also been applied to investigate the effect of decision-making on individual productivity. We differentiated a centralized decision-making as a delivery of decisions to the sales manager, rather than to sales representatives. This implies that sales representatives must comply with the process and ask the sales managers when they intend to deviate from their budget (expenses). In contrast, a decentralized decision-making structure was characterized by a delivery of decisions to the sales representatives, rather than to sales managers. By this, we mean that sales representatives decide on how to access a customer and may deviate from the formal process and have the freedom to manage on their own budget (expenses).

The observation of the decision-making structure was conducted using documentation supplied by the company. A centralized decision-making structure, for example, stipulated that sales representatives had a pre-defined activity plan for each sales call and a pre-defined budget for each call. Decentralized decision-making structure stipulated that sales representatives had a pre-defined annual budget per customer together with objectives per

121

customer, however, they were free to plan and execute the activities per sales call as they wanted, without the need for approval.

• Dependent variables

In order to investigate the relationship between complementarities and IT-enabled productivity, we differentiated between first- and second-order output measures. We operationalized sales representatives’ performance by using two measures to characterize a sales representative’s productivity. These measures consisted of a number of sales calls (face-to-face meetings) in relation to the duration of time worked and a number of products sold. Both metrics were obtained from company records over a period of four and a half years and are consistent with the way in which the company measures sales representative productivity. It should be noted that the actual sales of pharmaceutical products can also be influenced by other factors rather than sales calls alone, such as marketing activities aimed at prescribers, any initiatives directed at consumers, or where possible, access measures conducted at local level, i.e. a county and/or a clinic (health center), and also at national level, including pricing and reimbursement activities. Yet, in this study, these other factors may be regarded as given and their effects are assumed to be neutralized due to the four-by-four experimental design used in this study.

The choice of dependent variables is as follows. Unlike most industries, pharmaceutical sales representatives are engaged in detailing, presenting, discussing and arguing aspects related to a given product directly with prescribers. However, there is never a commercial deal made as such, as the effect of the interaction between the sales representative and the prescriber occurs when the prescriber prescribes a drug for a patient. The function of the sales representatives and the calls he or she makes to the prescriber is, therefore, to influence the prescriber to choose the product that the sales representatives represent. Selling of pharmaceutical products may be understood in terms of repetitive buying behavior by the prescriber, often meaning short buying cycles, in the context of competitive interchangeable products being available at hand. This implies that sales representatives need to maintain a frequent presence at the prescriber’s office. Therefore, the underlying assumption of the first productivity measure such as a number of sales calls is that the more calls conducted per sales representative, the greater productivity is achieved, as the key function of a sales representative is to conduct such sales calls with the prescribers. Hence, the measure of a number of sales calls is the number of sales calls (visits) a given sales representative has conducted during a given period of time. Therefore, a number of calls was operationalized as the average number of sales calls per quarter made by every sales representative to customers targeted by the company.

122

There are different ways of measuring product sales. For example, product sales can be measured in absolute numbers of money units or in absolute numbers of units of products sold; they can also be measured in relative terms of money units, market share of the total market versus the competition’s share, in relative terms of product units as market share of the total market and measured against the competition’s share. The methods described above may also be measured against sales objectives and whether the actual sales meet the objective or not, and how well. In order to maintain confidentiality, the company provided us only with the number of products sold in absolute numbers as the final measure of output. Nevertheless, as the primary goal of this research is to identify what sales effects are caused by the introduced changes, the number of units of products sold is the relevant second-order measure to be employed in this study. Furthermore, available alternatives that focus on units of money or relative measures only may potentially induce error as price per unit of a given product may change over time.

• Control variables

Based on existing literature on individual productivity of information workers (Aral et al., 2006) we expected that the following factors would influence dependent variables besides the independent variables of interest. As demographic differences between respondents may have independent effects on their productivity several traditional demographic and human capital variables have been taken into account, including gender, age, level of education, marital status, experience as sales representative, company, and industry, to control observable differences measured by questionnaire response. We also controlled the year when the product was launched and the market mean of sales calls as sales changes may also be conditioned by market dynamics.

4.2.4 Participants Our study comprised all sales representatives operating within the sixteen business units addressed here. Their duties included selection/targeting of prescribers, booking, preparing, execution of sales calls and follow- up/reporting after the sales calls have been conducted. The company under investigation provided a good sample frame for testing the developed model as it fulfilled major conditions for the research: (i) the sales representatives were in a situation when a more aligned IT system had recently been implemented, (ii) there were four markets and four products, thus sixteen cases or fields of study, to neutralize potential sales performance impacts driven by other external factors such as market, customer, competitions,

123

pricing, or internal such as marketing activities or internal culture, (iii) the sample was large enough for advanced statistical analysis. The Nordic affiliate also satisfied our research needs because it offered an environment where the developed complementarity set-ups could be introduced. Moreover, the company allowed us to collect output metrics over an extensive period of time.

Descriptive statistics of participants taking part in the study are presented in Table 4.4. Table 4.4: Descriptive statistics

Factor № Min. Max. Mean Std. Dev.

Adaptive/innovative cognitive style 91 72 119 91.76 10.266

Gender 91 0 1 0.56 0.499

Year born 91 1959 1990 1977 5.944

Marital status 91 0 1 0.67 0.473

Education 91 1 3 1.67 0.603

Years of experience in industry 91 3 30 10.13 5.369

Years of experience in sales 91 3 20 6.82 3.466

Years of experience in the company 91 1 11 4.65 1.747

Average number of sales calls per sales representative per working day in a par-ticular quarter

91 0.9 4.1 1.91 0.722

Average number of products sold by sales representative / quarter 91 149 3170 957.45 802.945

The initial sample consisted of 98 sales representatives. Questionnaires

about adaptive/innovative cognitive style and control variables were sent to all sales representatives. The participants were assured of confidentiality of their answers and advised that the information they share will not be disclosed to other individuals. This data acquisition procedure gave a 100 % response rate. Yet, during quarter 4 of 2014 two sales representatives left their jobs. During quarter 3 of 2015 the company reallocated five sales representatives to different business units. This has been done to increase the positive stimulation of the employees working with other products and therapy areas. Those participants who left their jobs and were reallocated to different business units, were removed from the analysis. Therefore, the final sample size for the analysis was represented by 91 subjects. Of these 91 subjects 44% were men and 56% women. The average age was 39 years. Most participants

124

had a Master (52%) and Bachelor degree (42%). Participants also reported their working experience. For example, on average, participants had 10 years of experience in the sales industry, 7 years of experience in sales and 5 years of experience in the company. On average, each sales representative made 1.9 sales calls per working day and sold 957 product units per quarter. The whole data collection procedure is described in more detail below.

4.2.5 Data access and data collection procedure This study was conducted over a period of four and a half years to examine the impact of complementary factors on individual IT-enabled productivity of sales representatives in a large international pharmaceutical company. Multiple data sources have been used to collect all necessary data. Table 4.5 provides a description of data collection and data sources. Table 4.5: Description of data collection and data sources Data Source Time

Pre-

impl

emen

tatio

n da

ta Organizational description Meetings, discussions

and documentations January 2013 – January 2014

A less aligned IT system de-scription

Meetings, discussions and documentations

January 2013 – January 2014

A more aligned IT system de-scription

Meetings, discussions and documentations

January 2013 – January 2014

The structure of the opera-tional process

Meetings, discussions and documentations

January 2013 – January 2014

Adaptive/innovative cognitive style and demographic characteristics

Questionnaire (human resource department)

February 2013

Productivity outcomes before the intro-duction of a more aligned IT system

Accounting function, sales support system

May 2014 – September 2014

Productivity outcomes after the intro-duction of a more aligned IT system

Accounting function, sales support system

May 2014 – July 2016

The study was initiated in January 2013 when the managing director of

the Nordic Affiliate and the director of human resource department agreed to the study design after a set of meetings and discussions. Data were collected based on the timeline presented in Figure 4.6.

125

Figure 4.6: Timeline of major events

126

The data collection began with pre-implementation details about organizational settings, the structure of the operational process and characteristics of pre-existing IT system. A precise description of a new IT system configuration and the production process was collected through discussions with business units’ managers and human resource department. At the same time, data about the launch time of the main products and a number of competitors were collected.

In order to identify adaptive/innovative cognitive style and demographic characteristics (age, gender, level of education, working experience) of the employees, the human resource department administered questionnaires to employees in the units under investigation in February 2014. Confidentiality was assured by assigning a unique code to each sales representative approached. The human resource department mailed the questionnaires back to the primary researcher with a response rate of 100%. The questionnaires were then analyzed.

In March 2014, a number of instructions with changes based on the developed designs (configuration set-ups) were delivered to business unitmanagers (Appendix C). The new sales support IT system was provided to The Nordic affiliation by the company head office and implemented with suggested complementary factors at the end of April 2014 with the aim of increasing individual productivity of employees.

Data collection of productivity outcomes (number of calls and products sold) from the previous nine quarters aimed to constitute a benchmark, or reference point, and nine quarters after the operational changes began in May 2014 and were finalized in July 2016. These data were collected from the CRM-system log data. The general analytical strategy of the collected data is presented in the next sub-section.

4.2.6 General analytical strategy In Study 1, the outcomes were observed for a number of treatment groups (designs) in the pre-implementation and post-implementation periods and we investigated the effect of a treatment (the introduction of a more aligned IT system together with complementarity changes) on the outcome by comparing the average change over time for the treatment groups and control group; therefore, a difference-in-difference (DID) approach was appropriate for analyzing available data (Angrist & Pischke, 2008). We used the number of sales calls (face-to-face meetings) and the number of products sold as dependent variables in relation to the duration of time worked by an individual. The rationale of the analytical strategy is presented below in more details.

127

Previously, Athey and Stern (1998) summarized the most common approaches to analysis of complementarities (interaction) between various variables. These include: (i) correlation approach; (ii) reduced form exclusion restriction approach, and (iii) productivity approach. The correlation approach is primarily based on the positive correlation between various variables, even after controlling for exogenous variables. Current studies widely use this approach (Tambe et al., 2012; Aral et al., 2012b). For example, based on this approach a correlation was established between such variables as decentralization, external organization focus; human resource monitoring and performance pay practices.

The reduced form exclusion restriction approach is based on the idea that a factor which makes an impact on one variable will not be correlated with another variable unless those variables are complementary to each other. Yet, Athey and Stern (1998) argue that this approach is not applicable for testing the complementarity effect when more than two choice variables exist. Moreover, application of this approach requires reconsideration of restriction procedures.

The productivity approach is based on the objective production function modeling with a set of variables, including the interaction effects between variables as complementarity parameters. For example, Ichniowski et al. (1997) applied this approach when studying the complementarity effect between various human resource management practices and output in the steel industry. This approach may be generalized and extended to three or more complements. In addition to the above-mentioned tests, Brynjolfsson and Milgrom (2013) developed a ‘cube view’ and applied the system test in order to visualize the complementarity effect between companies that adopted the system and companies that adopt only one of the complementarity factors.

All aforementioned approaches were used to test data that is mostly of a continuous nature. For example, those studies tested a three-way interaction that is presented when the interaction of two variables differ with the levels of the third variable (Wickens & Keppel, 2004). What approach and tests are appropriate for artificially created independent variables of a categorical nature? For example, the correlation test demonstrates whether factors under investigation have to be adopted together when a more aligned IT system is deployed in the working process. This test is not appropriate when independent variables are dummies as they lack scales and correlation is a measure of association between scale variables.

The productivity approach examines whether or not the hypothesized complements are more productive when adopted together or separately. Therefore, our general analytical strategy started with a theoretical model specification development. Like most of the studies that estimate the effect of complementary factors on productivity indicators (e.g. Ichniowski & Shaw, 2003; Bartel et al., 2007; Tambe et al., 2012; Aral et al., 2012b) we started with a generalized production function describing output (Q) as a function of

128

computer capital (IT), individual capital (IC) and organization capital (OC). The nature of the sales process makes it possible to develop a specific model of productivity. Potential output (number of products sold) by individual i in quarter t, or output Qit, is a function of three main inputs: IT system, individual adaptive/innovative cognitive style, and organizational factors (human resource management practices).

f(ITit, ICit, OCit) =>Qit (4)

Since sales representatives generate revenue by conducting sales calls and selling products, we defined real output (Qit) in terms of primary (number of sales calls per time unit) and secondary (number of products sold per time unit) dependent variables. Therefore, our general model specification had the following formula:

f(ITit, ICit, OCit) => No. of Callsit => No. of Products Soldit (5)

In Study 1, we intended to demonstrate that productivity of sales representatives has changed after the introduction of the more aligned IT system together with complementarities. We had four designs (treatment groups) and productivity data was measured at pre- and post-treatment time periods. Since the outcomes were observed for a number of treatment groups in the pre-implementation and post-implementation periods and we wanted to investigate the effect of a treatment on an outcome by comparing the average change over time for the treatment group and control group, a difference-in-difference approach was appropriate for analyzing available data (Ashenfelter & Card, 1985).

Formally, DID analysis is conducted by running the following regression:

Yit=αi + β1Treati + β2Postt + β3Treati * Postt + β4 CONTROLSit + ϵit,, (6) The notations and their interpretations from the above-presented formula are described in Table 4.6.

129

Table 4.6: Key components of the DID model

Notation Interpretation Yit Dependent variable (output measure) for treatment group i in

quarter t. We differentiated between first- and second-order output measures. The first-order measure consisted of the number of sales calls (face-to-face meetings) in relation to the duration of time worked and the second-order measure was the number of products sold. Yet, the first-order variable was in our main interest.

αi Individual fixed effects. Treati Dummy variable that equals one for individuals in the treatment

group and zero otherwise. Postt Dummy variable that equals one for the post-treatment period and

is zero for the pre-treatment period. Treati * Postt Dummy variable that can be described as the interaction between

a dummy variable that indicates whether individual i is in the con-trol or treatment group (Treati) and a dummy variable that indi-cates whether the observation corresponds to the period after the implementation of a more aligned IT system (Postt,).

CONTROLSit Control variables, including traditional demographic variables and organizational capital factors.

ϵit A random, unobserved ‘error’ term which contains all determinants of Yit which our model omits.

In our case, DID analysis is appropriate, since DID estimators can be

obtained by using OLS regression. In turn, regression procedure allows us to include other important variables and obtain more precise DID estimators. With a DID approach, it is easy to include multiple time periods and groups as it is in our research design and to calculate standard errors. Therefore, DID analysis is considered as appropriate for answering the research question based on the developed research design.

Therefore, in this section we have summarized research settings, conceptual set-up, complementarity configurations, and measurement of the main variables investigated in the Study 1. We also described how the data were accessed and collected over a period of study. Finally, we described the general analytical strategy of data analysis. In the next section, we present how the study of software programmer productivity was conducted in a well-controlled online experiment.

130

4.3 Study 2: Software programmer productivity

An online experiment (Study 2) was conducted to investigate the software programmer productivity with the above-given factors being tested in several configuration set-ups with reference to the developed hypotheses. For exam-ple, we expected that productivity of software programmers will increase if adaptive/innovative cognitive style is correctly matched to “stable”/”dy-namic” complementarity set-up respectively after a more aligned IT system was deployed and stabilized. An experiment was selected as the appropriate methodology to study complementary factors as it provides a clean test of hypotheses and causal relationships in well-controlled and carefully designed conditions (Camerer & Weber, 2013). Despite the fact that longitudinal field research overcomes many shortcomings of other research methods, other fac-tors, beside independent variables can affect the results in field experiments (Cardwell et al., 2001). Unlike laboratory and online experiments, replicating the context of longitudinal field experiments is difficult. Therefore, an online experiment was expected to provide a better control over extraneous variables and more accurate results. The way in which the experiment was organized, including its design and procedure, is presented below. We also describe how the main variables were operationalized, how the data were collected, and what analytical strategy was applied to derive the research findings.

4.3.1 Experimental design, procedure and pilot testing The purpose of this experimental study was to test whether a set of matched complementary factors can indeed affect productivity in relation to adap-tive/innovative cognitive style when a more aligned IT system was imple-mented. According to the developed hypotheses, we expected that adaptors rather than innovators have productivity advantages when a more aligned IT system is used together with a “stable”complementarity set-up. In contrast to adaptors, we expected that innovators would win productivity advantages when a more aligned IT system is used together with a “dynamic”comple-mentarity set-up (Table 4.7).

131

Table 4.7: Research design for the experiment and expected results

Cognitive style versus complementarity set-up Stable Dynamic

Adaptive Productivity gain Productivity loss Innovative Productivity loss Productivity gain

The experiment was designed to meet the following criteria: (a) individuals

with both adaptive and innovative cognitive styles are proportionally involved in stable and dynamic complementarity set-ups during all sessions, (b) com-plementary factors proposed in our research model attempt to reflect, as closely as possible, the real world of software development, (c) subjects cho-sen for the experiment possess relevant software development skills and are appropriate for the experiment (Camerer & Weber, 2013), and (d) sample size has to be large enough to provide statistical power of the test.

It was expected that both adaptors and innovators would be proportionally involved in “stable” and “dynamic” complementary operational set-ups. This comparison group design allows the control and establishment of evidence that a specific complementarity set-up when matched with a particular cogni-tive style indeed increases individual IT-enabled productivity. The partici-pants were randomly assigned to one of the complementarity set-ups. Yet, the mode of data collection allowed us to obtain relevant data only to partially test the research model and identify how productivity changed for innovators in-volved either in “stable” or “dynamic” complementarity set-ups.

The experiment consisted of the following main steps:

1. Initial instruction; 2. Pre-experimental questionnaire to identify adaptive/innovative

cognitive style; 3. The experiment (three sessions/assignments); 4. Data analysis.

The experimental procedure began with an initial instruction about the pur-

pose and structure of the experiment (Appendix D). The participants were in-formed that the focus of the assignment is on logic and basic programming concepts. The participants were assured about the confidentiality of their per-sonal information. Then the participants completed a questionnaire that al-lowed us to identify adaptive/innovative cognitive style and collect necessary demographic control variables (Appendix E). After identification of adap-tive/innovative cognitive style, the experimental sessions were conducted (Table 4.8).

132

Table 4.8: Experimental sessions

Factor Session 1 (benchmark with existing IT system)

Session 2 (introduction of the new IT system)

Session 3 (learning effect)

Treatment condition Stable Dynamic Stable Dynamic Stable Dynamic

Cognitive style 25% adaptors & 25% innovators

25% adaptors & 25% innovators

25% adaptors & 25% innovators

25% adaptors & 25% innovators

25% adaptors & 25% innovators

25% adaptors & 25% innovators

IT system Less aligned Less aligned More aligned More aligned More aligned More aligned

Operational pro-cess Structured Flexible Structured Flexible Structured Flexible

Training and ed-ucation No No Mandatory On-demand No No

Incentives External Internal External Internal External Internal

Decision-making structure Centralized Decentralized Centralized Decentralized Centralized Decentralized

Output metrics Time Quality

Time Quality

Time Quality

Time Quality

Time Quality

Time Quality

133

The productivity of software programmers was measured at three points in time. In the first session, in order to establish a benchmark, participants developed a software application using a text-editor, which accounts for an existing IT system in an organization. In the second session, an advanced IT system that provides comprehensive facilities for software development was introduced in a synchronized manner with both cognitive styles and complementarity set-ups. The third session was designed to take into account a ”learning curve” effect (Womer, 1984; McLeod et al., 2008). This session also included the advanced IT tool and complementarities as in the second session. Each session had identical time frames (approximately 20 min. – 1 hour) and a slight variation of assignments, yet with an equal level of complexity. Therefore, this study was a comparison group controlled evaluation of the effective IT use before, after and at follow-up of a new IT system use.

The online experiment was pilot tested in two phases. A lab-based version of the experiment was designed first, and it was subsequently transferred to an online version. The lab-based version was tested in two rounds. First, two professional software programmers performed the experiment in order to identify whether initial information, purpose, and sessions of the experiment were described clearly. The results of this pre-pilot study indicated that more detailed information about the purpose and structure of the experiment were required. The participants were also asked about the complexity of assign-ments in all three sessions. It was reported that the assignments are equally complex, yet some JavaScript help is necessary since in reality software pro-grammers need to look for reference material in order to perform their job. After the initial instruction of the experiment had been improved, it was pilot tested with five undergraduate students studying courses within the field of software development and computer applications software engineering. The results of this test indicated that the students failed to perform one of three assignments. This was explained by their difficulties in concentrating within given time frames and short breaks. Following this feedback, we decided to develop an online version of the experiment since online experiments allow 24-h access and uniformity of the procedures across participants (Reips, 2002). Four participants tested content and functionality of the online version of the experiment. Therefore, the online version was essentially the same as the lab version of the experiment. Yet, participants in the online experiment received participation compensation aimed to attract participants to partake in the study. We also provided performance rewards for those participants who were involved in a ”stable” complementarity set-up with exogenous incen-tives.

134

4.3.2 Participants and experimental material The participants in this experiment were professional software programmers. They were recruited by using two crowdsourcing platforms that provide on-demand participants: Microworkers (URL: https://microworkers.com) and Prolific.ac services (URL: https://prolificac.zendesk.com/hc/en-us). The main requirement for the participants was knowledge and experience of HTML/Ja-vaScript programming. The experiment was also open to participants familiar with similar languages as JavaScript, including Java, C and Python. There was no restriction in age, nationality, gender, education or prior programming ex-perience. Participation in the experiment was completely anonymous, volun-tary and rewarded with 30 USD (10 USD for each session/assignment). This payment was provided for two reasons. The first was to increase the partici-pation attractiveness while the second reason for payment was to create real work settings where employees earn their salaries for the job performed. In order to receive full data, participants were paid only when all three assign-ments had been completed within a reasonable time with the accepted quality level.

Sample data was collected in the following way. The online experiment was available online for 4 months. During that time, there were 2192 visitors. Of these, 1196 (55%) completed the questionnaire but did not attempt to begin the experiment. A total of 398 (18%) of visitors completed assignment 1 and 71 (3%) completed assignment 2. A total of 527 (24%) participants submitted all assignments as completed, and, of these 414 were rejected because their submissions did not meet the minimum requirements for the submission to be accepted (unreasonable time and poor quality of the assignments). Thus, 5% of the initial 2192 visitors completed the experiment. Table 4.9 summarizes these participation outcomes.

Table 4.9: Participation outcomes for four months of online availability Participation outcome Count Percentage

Total visitors 2192 100

Questionnaire completed 1196 55

One assignment completed 398 18

Two assignments completed 71 3

Submitted all assignments as completed 527 24

From submitted assignments as completed were rejected 414 19

From submitted assignments as completed were completed 113 5

135

Three participants of the 113 have had adaptive cognitive style and were excluded from the analysis because of an insufficient number of participants that fell into this category of individuals. Therefore, the initial sample size consisted of 110 participants that had innovative cognitive style. The geo-graphical areas from which participants logged into the experimental website are shown in Figure 4.7.

Figure 4.7: Percentage of participants from different geographic areas

52 (47%) participants were from Europe. 25 (23%) participants were located in Asia and 16 (14%) in North America respectively. The minority of participants logged in from South America, Africa and Australia.

Below, we demonstrate that a repeated-measures ANOVA is an appropriate statistical approach for the conducted experiment. Yet, in the repeated measures ANOVA, unequal sample sizes can affect the homogeneity of variance assumption when the difference between sample sizes is large (Keppel, 1991). In our case, we had 44 innovators involved in a “stable” complementarity set-up and 66 innovators involved in a “dynamic” complementarity set-up. Indeed, Levene Statistic in unbalanced sample sizes was significant for the dependent variable – quality in assignment 1 (p=0.016). To provide an accurate p-value, we balanced sample sizes by randomly selecting 44 observations out of 66 individuals with innovative cognitive style that were involved in a “dynamic” complementarity set-up. Therefore, the final sample size consisted of 88 observations, including 44 innovators involved in “stable” and “dynamic” complementarity set-ups respectively. An a priori power analysis with a power of 0.80 (Cohen, 1992) and an alpha of 0.05 indicated that a sample size of 28 observations is needed to detect a medium effect size (0.25). Therefore, we expected to obtain a significant difference between groups with the final sample size of 88 (44 in

136

each group) observations. The descriptive statistics are presented in Table 4.10. Table 4.10: Descriptive statistics Factors № Min. Ma. Mean Std. Dev.

Adaptive/innovative cognitive style

88 97 141 111 9.807

Gender 88 0 1 0.1 0.305

Year born 88 1967 1997 1988 6.944

Marital status 88 1 3 1.27 0.473

Education 88 1 5 2.7 1.176

Experience in programming 88 1 4 4.01 1.246

Experience in JavaScript 88 1 4 2.49 0.661

Number of programming lan-guages

88 1 5 3.7 0.899

Time spent with IT 88 1 4 3.01 0.795

Effective IT use 88 2 5 4.03 0.651

Routineness 88 1 5 2.98 1.083

The sample consisted of participants between 19 and 49 years of age, with

an average age of 28 years. Of the total 88 participants 79 (89.8%) were male and 9 (10.2%) female. The largest number of participants had a Bachelor degree (35 - 39.8%), whereas 23 (26.1%) had a High School education, and 22 (25.0%) a Master degree. Participants also reported their programming experience. For example, 30 (34.1%) participants reported up to five and 27 (30.7%) up to ten years of programming experience. Most participants (39 - 44%) used the JavaScript language on the regular basis and 4 (4.5%) were experts in this programming language. Participants had experience in three 40 (45.5%), four 24 (27.3%), and more than five 21 (23.9%) programming languages. A total of 54 (61.4%) participants reported that they are highly effective at using new technological tools. Preliminary analysis for both dependent variables (time and quality) found no significant main effects or interactions with control variables.

On the participant side, there was no required material. Participants developed software applications in natural settings using their own equipment – a computer and a browser with internet connection. All steps of the

137

experiment were automatized. Participants could optionally leave their comments about the experiment and request further information about the experiment. A document with information about the study results was sent to interested participants. Some participants reported enjoying the task and wishing to conduct similar experiments.

4.3.3 Operationalization of the main constructs The primary independent constructs of the experiment were an IT system, the structural complexity of the operational process, type of training and educa-tion, incentives and decision-making structure (Table 4.11). During the first session, participants have used a less aligned IT system. To perform their as-signments, the participants used only a generic text-editor (Notepad) to change the relevant file, and a web browser to run the code. During the second and third sessions, a more aligned IT system was deployed and used. This IT system was a Cloud9, which is an online IDE (Integrated Development Envi-ronment) with many of the features of a full-blown desktop IDE, such as code completion and integrated debugging tools. We were in no way affiliated with Cloud9. Our intention of using this tool was to ensure that the participants could work in a realistic professional environment. Table 4.11: Operationalization of the main constructs Construct Operationalization

IT system

Less aligned

A generic text-editor (Notepad) with basic text manipula-tion functions to change the relevant file. This text-editor allows: • creation of simple text documents; • editing files.

More aligned

Cloud9 – an online IDE that provides the following addi-tional facilities for software development: • pre-set-up workspaces; • code completion (participants can discover things by

typing code) and color highlighting; • automatic code reformatting, navigation and debug-

ging; • real-time language analysis for JavaScript;

138

Table 4.11: Continued from previous page Construct Operationalization

Opera-tional

process

Structured • Fully comprehensive – the requirement specifica-tion is perfectly specified with regard to all key as-pects of the software to be developed;

• Without any modifications being made to the re-quirement specification during the process of soft-ware development.

Flexible • Only partially completed requirement specification, implying that the software programmer needs either to get back and ask for more information for the re-quirement specification and/or make assumptions on her own;

• With some modifications and/or additions to the re-quirements specifications being given.

Training and

education

Mandatory Basic information about Cloud9 IDE and its main fea-tures and two introductory videos about how to get started and create a workspace.

On- demand

Basic information about Cloud9 IDE and its main fea-tures and a recommendation to watch short videos about Cloud9 if needed to make the most of using the IDE.

Incentives

External A possibility to receive rating and conclusions once the experiment is finished.

Internal A message with a positive feedback of the following content: “You performed well and keep going!”

Decision-making

structure

Centralized Message: “Do not improvise in the assignment and do not do anything different from the specification!”

De-centralized

Message: “Feel free to improvise in the assignment!”

Output metrics

Time Quality

Time – actual time subjects spent to develop a product, minutes. Quality – completeness and correctness, %.

The assignments for both structured and flexible processes were developed

based on two criteria (MacCormack et al., 2001). The first criterion was whether or not the requirement specification for the assignment is fully com-prehensive. The second criterion was whether any modification of the require-ment specification has been made during the process of software develop-ment. According to those two criteria, the process was defined as structured

139

when it was fully comprehensive (the requirement specification document was perfectly specified with regard to all key aspects of the software to be developed) and no modifications had to be made to the requirement specifi-cation during the process of software development. In contrast, in the flexible process, the requirement specification was partially completed, meaning that a software programmer needs to make assumptions on his or her own. In ad-dition, during the flexible process, some modifications to the requirements specifications were given. The formulated assignments for both structured and flexible processes are presented in Appendix F. Participants also received Ja-vaScript help in order to reduce the impact of the time they spend looking for reference material (Appendix G).

Before the second session, the participants received a specific type of training and education about Cloud9 IDE with regard to a particular complementarity set-up. For example, the participants involved in a “stable” complementarity set-up received information about the more aligned IT system and its main features. The participants were introduced to the main features of Cloud9, including code completion, automatic code formatting, code navigation and visualizing code documentation. It was also recommended to watch two introductory videos about how to get started and create a workspace with the new IT system. The participants involved in a “dynamic” complementarity set-up also received basic information about the main features of Cloud9 and a recommendation to watch short videos if needed to make the most of using the IDE.

All participants taking part in the experiment received 30 USD as participation incentives (10 USD for each session/assignment). These incentives were introduced to reflect working reality when employees receive a salary for performing their job and to limit dropout. To investigate the complementarity effect of incentives on individual productivity, we introduced external and internal incentives in the second and third sessions. For example, the participants involved in a “stable” complementarity set-up received messages that when they complete the assignments, they will be able to receive their rating and conclusions reached once the experiment is finished. These non-financial extrinsic rewards were introduced to demonstrate respect and recognition to the participants (Armstrong, 2010). The participants involved in a “dynamic” complementarity set-up during their performance of the second and the third assignments twice received a message with a positive feedback that they had performed well and a recommendation to keep going (Erasmus & Schenk, 2008).

Decision-making structure was manipulated based on a delivery of the most important decisions to a software programmer. To investigate the effect of centralized decision-making structure, the participants involved in a “stable” complementarity set-up in the beginning of the second and third assignments received the following message: “Do not improvise in the assignment and do not do anything different from the specification!” In

140

contrast, the participants involved in a “dynamic” complementarity set-up with a decentralized decision-making structure received the following message: “Feel free to improvise in the assignment!” in the beginning of the second and third assignments.

Among product-, process- and project-related factors, individual characteristics, including experience, knowledge and skills are reported as being among the most important factors of software development productivity (Walston & Felix, 1977; Lokan, 2001). For example, there is evidence that professionals with experience, knowledge, and expertise in a particular programming language are more productive compared with less experienced colleagues (Clincy, 2003; de Barros Sampaio et al., 2010). Therefore, we also controlled for demographic characteristics of participants together with expertise in JavaScript language and programming. Output metrics collected during the experiment are presented in the next sub-section.

4.3.4 Dependent measures In order to develop productivity measures, it is necessary to establish what productivity is and how it can be measured. In technical terms, productivity is a ratio of output units per units of input effort. This seemingly simple definition of productivity is not easily applicable in software development productivity measurement. For example, such output measures as the physical length of the written codes or functional points do not really indicate whether an individual is productive or not as these metrics depend on the coding style of the software programmer (Tomaszewski & Lundberg, 2005). Moreover, these metrics do not indicate any qualitative characteristics of the product developed.

The primary dependent variables in this study took into account two major dimensions of subjects’ performance such as quantity and quality. These two dimensions are equally important as “…To survive, IT firms must develop high quality software products on time and at low cost” (Harter et al., 2000, p. 451). In software programming productivity, a quantitative dimension is characterized by the change in the quantity of a developed product for a given period of time and a qualitative dimension refers to the quality of the software product (Duncan, 1988). Yet, as previously discussed the output metric (the unit of product) is a challenging construct in software development. Thus, our attention was devoted to the input part of the productivity ratio.

The input effort is usually a sum of resources that were used to produce output. In software development, the main part of resources is working hours spent on the software product development (Canfora et al., 2007). Therefore, time taken by the subjects to complete the assignment was used to characterize

141

a quantitative dimension of the productivity metric. Time values in the experiment were obtained after each assignment was performed by participants in every session.

Developed software products must also meet concrete qualitative requirements as the qualitative aspect may possibly explain the difference in time spent to complete a task. For example, to characterize the qualitative dimension of the developed software product, Tomaszewski and Lundberg (2005) consider design, final product and development process quality. In this experiment, we are interested in the final product quality, which is the quality of the developed application. Application quality can be examined from different perspectives. For example, how many defects the application has or how well the application satisfies the end-user (Edberg & Bowman, 1996). We used completeness (how many of the functional requirements were completed) and correctness (how well the functional requirements were implemented) of the application developed since these are the most significant and appropriate measures in the experimental settings (Appendix H). Thus, after each assignment, the data about productivity metrics (time and quality) was collected and analyzed based on the analytical strategy presented below.

4.3.5 General analytical strategy As mentioned before, we collected data from participants with only innovative cognitive style. Since we collected data only for innovators, 2x3 mixed design with repeated measures was used to examine the effect of complementary factors on software programmer’s productivity. The first factor in the experiment is the complementarity set-up and contains two levels: “stable” complementarity set-up and “dynamic” complementarity set-up. Time and quality of the developed products were measured at three time points: t1 – pre-test (baseline) before the introduction of a more aligned IT system, t2 – post-test after the implementation of a more aligned IT system and complementarity set-ups and t3 – follow-up to account for learning effect, i.e. the same configuration as in t2. Therefore, the second factor is time with three levels (t1, t2 and t3). Two separate groups have been tested at three different times, meaning that we had one between- and one within-subject factor.

According to the purpose of the experimental study, we were interested in how productivity of innovators involved in “stable” and “dynamic” complementarity set-ups differ over three testing time points and how factors under the investigation interact. At the same time, we were interested in whether the variance between conditions is larger than the variance within conditions. Since we had one between- and one within-subject, we used A General Linear Model for repeated measures. We studied one between-subject

142

factor with two levels and one within-subject factor with three levels. Time and quality of the application developed were not significantly correlated. Therefore, both dependent variables were analyzed with repeated measures ANOVAs.

A factor effects model for repeated measures ANOVA is denoted by the following formula:

Yijk=µ+τi+βj+(τβ)ij+εijk (7)

The notations and their interpretations from the above-presented formula are presented in Table 4.12. Table 4.12: Key components of the factor effect model for repeated measures ANOVA

Notation Interpretation Yijk Output measure of the observation taken at the kth time from jth group

and ith subject µ The overall (grand) mean τi The main effect of factor A (complementarity set-up); βj The main effect of factor B (time point). (τβ)ij The interaction effect between factor A and factor B. εijk Error term

Following this design, we test three possible hypotheses, one for each main

effect and one for the interaction:

1. H0: µstable complementarity set-up = µdynamic complementarity set-up H1: µstable complementarity set-up ≠ µdynamic complementarity set-up 2. H0: µtime1 = µtime2 = µtime3 H1: not all time means are equal 3. H0: an interaction is absent H1: an interaction is present

ANOVA was chosen as an appropriate and powerful technique to identify an effect of a combination of variables that respond to the developed research design. Moreover, ANOVA technique has a number of advantages (Field, 2011). For example, ANOVA (a) reduces the chance of making a Type I error in comparison with the t-test, (b) does not require a large total sample size, (c) removes some of the random variability, and (d) enables the examination of interactions between factors that is a particular interest of our study.

143

In general, the aim of Study 2 was to investigate whether software programmer productivity increases when a particular cognitive style is matched to appropriate complementarity set-up. After the data collection process, we expected that individuals with an innovative cognitive style would be more productive when involved in a “dynamic” rather than “stable” complementarity set-up. However, we also expected that productivity scores for participants involved in both complementarity set-ups may increase for time scores and decrease for quality scores during the second session while learning a new IT system. Eventually, it was expected that in the third session software programmers with innovative cognitive style would be more productive when matched with a “dynamic” rather than “stable” complementarity set-up.

Therefore, in this section we described how the online experiment was designed and performed. We also described study sample together with operationalization of the main constructs and analytical strategy. While both studies, Study 1 and Study 2 were carefully designed and performed, both are vulnerable to validity and reliability threats. The steps undertaken to reduce these threats are described below.

4.4 Validity and reliability

Two empirical studies, including field and online experiments, have been carried out to test previously developed hypotheses. To demonstrate the accuracy and credibility of the research conducted, we refer to validity and reliability concepts as main validation issues, primarily addressed in quantitative research (Cook & Campbell, 1979; Shadish et al., 2002). Validity (how accurately research represents the investigated world) refers to the best possible approximation of the relationships between the independent and dependent variables in the study in comparison with the true relationships for the entire population. Reliability characterizes consistency of the results when the experimental study is replicated. Onwuegbuzie (2003) suggests that more than fifty possible threats of validity can exist. In this research, a number of steps were undertaken to assure and assess conclusion, construct, internal and external validity of the main findings that are presented below.

4.4.1 Conclusion validity Conclusion validity refers to the statistical relationships between treatment and its effect on the dependent variables, given a certain significance level

144

(Wohlin et al., 2000). There are several threats that may affect conclusion validity, including unmet assumptions of statistical tests, low statistical power, low reliability of the measurements and treatment implementation. We assessed the degree to which conclusions about the statistical relationship between the treatments of the independent variables and their effect on the dependent variables are correct (Lipsey, 1990). Statistical conclusion validity can be assured by appropriate statistical tests, adequate sampling procedures, and reliable measurement procedures (Cook & Campbell, 1979). Below, statistical conclusion validity is assured by demonstrating the choice of the statistical tests, their statistical power, reliability of the dependent variable measurement and treatment implementation.

First, we checked that the main assumptions of the statistical tests presented in more detail in the next chapter were not violated in both studies. In Study 1, assumptions of independence, linearity, homoscedasticity, no perfect multicollinearity, normality and parallel trend were met. In Study 2, we checked that normality, homogeneity of between-group variance, homogeneity of variance-covariance matrices and sphericity of within-group variance assumptions were not violated. Therefore, the choice of statistical tests was justified.

Second, in order to assure that the obtained results are valid, the statistical power of the statistical tests was considered based on three main components, including a significance criterion; sample and effect size (Miller et al., 1997). In both tests (DID and ANOVA), the significance criterion was set at the level 0.05 to protect against Type I error. This means that there is less than 1 in 20 chance of incorrectly rejecting a null hypothesis (Cook & Campbell, 1979). The statistical power level was set at 0.8 which is a recommended value to avoid type II error (Cohen, 1992). An a priori power analysis with alpha 0.05. and statistical power 0.08 indicated that a sample size of 76 observations was needed to detect a medium effect size (0.15) in DID test. A sample size of 28 observations was also required to detect a medium effect size (0.25) in ANOVA test. Thus, we assured that the results of the statistical tests were not affected by low statistical power and the sample size was large enough to ensure confidence in meaningful results.

Conclusion validity also depends on the reliability of the dependent variable measurement. A measurement is considered reliable if it is repeated over time (Crano et al., 2014). There are two broad categories of measurements: objective and subjective. Objective measurements are more reliable than subjective since they do not include a response bias (Alexander & Wilkins, 1982). In Study 1, we used number of sales calls and products sold to operationalize dependent variables. These variables were received from the company during multiple time points. They are objective and do not include response bias. In Study 2, we used time and quality of the developed application as dependent variables that are also objective. Hence, we

145

minimized this potential reliability problem of the dependent variable measurement in both studies.

Reliability of treatment implementation can also affect conclusion validity. This reliability concerns how the treatments of the independent variables are applied to the participants (Cook & Campbell, 1979). To assure that treat-ments were similarly applied to participants, we conducted the following steps. In Study 1, we developed identical and detailed instructions of how complementarities have to be introduced in each business unit according to the developed design (treatment). The instructions were delivered to business unit managers at the same time. In Study 2, reliability of treatment implemen-tation was fixed on the developed website and treatment was not changed over the whole period of study. Therefore, we ensured that treatments were simi-larly applied to all participants in both studies.

4.4.2 Construct validity Construct validity assesses the extent to which the independent and dependent variables reflect the constructs of interest (Judd et al., 1991). Operationaliza-tion of independent and dependent variables is clearly defined in this chapter with instruments to allow replication by the others. For example, in both stud-ies, we used the Kirton Adaption-Innovation Inventory (Kirton, 1976) to as-sess individuals’ cognitive style. Over the last decades, the KAI has been found to demonstrate high levels of test-retest reliability, ranging from 0.82 to 0.86 (Duff, 2004), internal consistency reliability, ranging from 0.79 to 0.91 (Brown, 2001), and high construct validity (Kirton, 1994). In the field and online experiments, we used two different process-enabling IT systems. We provided a detailed description of both systems, including new functions and features. Complementary factors, including the structural complexity of the operational process, training and education, incentives, and decision-making structure were operationalized based on criteria from theoretical sources, yet were closely related to occupational requirements.

In order to determine that we had successfully configured the complemen-tary factors, in Study 1 we conducted a validation check to assess that the treatment functioned as expected (Appendix I). We compared the responses from the participants with different cognitive styles. The analysis confirmed the validity of the experimental manipulations. For example, adaptors re-ported more encouragement when they were involved in a “stable” comple-mentarity set-up and innovators reported more encouragement when they were involved in a “dynamic” complementarity set-up. Specifically, t-test confirmed that in contrast to adaptors, innovators demonstrated more encour-agement to work with key opinion leaders (M = 6.29; SD = 0.756), t(28) = 13.013, p < 0.001 in the operational process where almost all steps are non-

146

mandatory (M = 5.57; SD = 0.535), t(28) = 11.997, p < 0.001. Innovators also reported that they feel more encouraged when they receive on-demand train-ing and education in a new IT system (M = 5.57; SD = 0.787), t(28) = 8.289, p < 0.001 and constant feedback from managers (M = 6.43; SD = 0.535), t(28) = 11.089, p < 0.001. In a situation when most of the important decisions such as deviation from the budget have to be made without contacting the main manager, innovators also felt more encouraged than adaptors (M = 6.14; SD = 0.690), t(28) = 9.952, p < 0.001. Innovators, in contrast to adaptors, also reported that they feel more encouraged when they use a new version of an IT system which is more functional (M = 5.86; SD = 0.900), t(28) = 9.504, p < 0.001. Therefore, we have construct validity evidence that our manipulation by complementarities indeed affected the behavior of adaptors and innovators. In Study 2, we could not conduct a validation check, since most of the partic-ipants lacked adaptive cognitive style and we could not compare their re-sponses with innovative cognitive style individuals.

There also exist a number of construct validity threats. Some attempts that were made to minimize them are discussed below. For example, one of the construct validity threats is hypotheses guessing, meaning that participants can guess the aim of the study and behave in the way they guess expects (Shadish et al., 2002). Therefore, the behavior of participants can be affected by their guesses and not by treatment. In order to minimize the threat of hy-potheses guessing, participants in Study 1 were not informed about the inde-pendent and dependent variables. In Study 2, participants were informed that the required proof of a completed job is reasonable time spent and a satisfac-tory quality level for each developed product. Yet, they were not informed about complementary factors under investigation. Participants also did not know about their cognitive styles and to which complementarity set-up they were allocated.

Another threat of construct validity is the Hawthorne effect. This effect can occur when participants know that their performance is being observed and try to perform better (Campbell & Stanley, 1967). We believe that this effect was insignificant in the field experiment. Although participants were informed that they were participating in the experimental study, they were not told what data was being collected. The field study was conducted over a pe-riod of nice quarters after the intervention. This may possibly limit the Haw-thorne effect. In the online experiment, the participants knew that they have up to one hour to perform each assignment and if the time is reasonable they would be paid. This could limit construct validity, yet the way the online ex-periment was developed well reflects reality when software programmers re-ceive salaries after the job is completed. We minimized the Hawthorne effect by using voluntary and anonymous participation.

Experimenter expectancies are also one of the construct validity threats. This is a potential threat to construct validity from experimenters’ bias about what they expect from the experiments (Cook & Campbell, 1979). In order to

147

minimize this threat, the research designs of both studies were created in such a way so that the results would not be biased to experimenter expectations.

4.4.3 Internal validity Internal validity is the extent to which a particular variable has caused effect in the study (Robson, 1993). Internal validity ensures that the causal relation-ship between the independent and dependent variables is not the result of other factors that were not measured (Shadish et al., 2002). Internal validity is sub-ject to a number of threats, including history effect, maturation and selection effects, testing threat, mortality, diffusion of treatment effects, compulsory rivalry and resentful demoralization effects (Cook & Campbell, 1979). Though both studies had robust designs, the potential threats and preventive procedures for assuring the internal validity are addressed below.

History and maturation effects can significantly jeopardize the internal va-lidity over time the study is conducted. History effects are caused by un-planned external events that take place during experimental studies. These un-expected events can influence the participants, resulting in changes in their behavior, attitude or knowledge. Study 1 could be subject to history effect threat since it was conducted over a long time period. Yet, we minimize this threat by the use of a control group which experienced the same history as the experimental groups. We also monitored the organization’s changes in their operation and structure over a period of nine quarters before the introduction of the new IT system, as relevant for the study. Study 2 was conducted over a period of four months and there was little risk that other external events could affect the dependent variables. Maturation effect may affect internal validity by changes in participants’ reaction as time passes. For example, participants can become older, less excited about the study or simply tired. In Study 1, we believe that maturation over a period of study was common to all participants. In Study 2, we minimized this threat by a short duration of the experiment.

Another threat to internal validity is a non-random assignment of treatments. In Study 1, we were given sixteen business units by the company and used four-by-four operational structure (four market and four products) when assigning treatments (designs). We assigned treatments to eliminate the impact of a certain market and product on the dependent variables. This assignment was done to provide appropriateness with the research model investigated and opportunity to access the impact of a certain design on the dependent variables. We also collected additional information about comparability between participants. Participants in the control group did not differ significantly from participants in the experimental groups. In Study 2, we assigned treatments in an alternating sequence for the purpose of convenience to run an online experiment. Yet, we did not assign treatments

148

based on a cognitive style. We believed that since participants with different cognitive styles took part in the online experiment in a random order, an alternating assignment of treatments allows us to avoid selection bias.

The testing threat can affect the internal validity when participants scores are measured a number of times. For example, participants’ scores may change not because of treatment, but because participants become more fa-miliar with the testing instrument. In Study 1, we eliminated this threat by collecting historical (pre-test) data before a more aligned IT system was in-troduced in the company. Longitudinal data collected after the implementa-tion of a new IT system allowed us to take into account learning effect of a more aligned IT system. In Study 2, we eliminated this threat by the different nature of the assignments, yet equivalent in their complexity. Moreover, par-ticipants were not aware of the full set of complementarities that were used in the experiment.

Mortality as a threat to internal validity is concerned with the number of participants who do not complete experiments. Indeed, in Study 1, we removed seven participants from the analysis, because some of them had left their jobs or moved to another business unit. However, given the total sample size, mortality had minimal effect on internal validity. In Study 2, the experiment was conducted online. This type of experiment is often characterized by the high dropout rate for online participants (Dandurand et al., 2008). We minimized mortality threat by providing monetary incentives to the participants that allowed us to collect a sufficiently large number of participants to provide a powerful statistical test.

Diffusion or treatment effects can occur when participants from non-inter-vention groups learn about the treatment and try to imitate behavior from the intervention group. In Study 1, this threat was minimized in the following way. Even though participants knew that they were taking part in a field ex-periment related to the use of a new IT system, they did not know about the exact set of complementarities. Therefore, they could not imitate the behavior of their colleagues. In Study 2, our design was developed with no non-inter-vention group, therefore, this threat was minimized. The same strategy al-lowed us to minimize rivalry and resentful demoralization effects.

4.4.4 External validity External validity is the ability of the investigator to generalize the causal relationship beyond the conducted study (Davis & Buskist, 2008). It is recommended that we consider the external validity through its subtypes, including population, ecological, temporal validity and type of sampling (Bracht & Glass, 1968). Population validity refers to the extent to which findings can be generalized to the population at large. We tested our

149

theoretical model in two different information-intensive occupations. However, the model was developed in such a way as to provide its application in different information-intensive settings, thus further providing evidence about the population validity in relation to information-intensive occupations. Study 1 was conducted in the Nordic affiliation of a global pharmaceutical company and data was collected across four countries, including Sweden, Denmark, Norway, and Finland. Population validity, in this case, can be constrained to a particular company and Scandinavian market, since any claim for generalizability to other organizations, job types, and countries with their institutions, settings can be challenged. However, even though the study was conducted in one company where one IT system was deployed, it focused on a number of business units (products and markets) and factor configurations to reduce the significance of IT system and specific contexts. The focus on one company increased a control for extraneous factors that could potentially provide another explanation for the findings. Furthermore, by focusing on a single company, we were able to unfold a rich view of complementary factors associated with individual productivity. The ideal way to ensure population validity would be a replication of the study among different populations and at different times. The website of the online experiment in Study 2 was internationally available online. Participants taking part in the experiment were of different ages, countries of origin, educational background and programming language experience. Therefore, population validity in this particular study can be considered as high.

Ecological validity is concerned with the extent to which findings can be generalized across real-world settings. On a general level, ecological and internal validity was assured by conducting both a longitudinal quasi-randomized field experiment and online experiment. For example, a longitudinal quasi-randomized field experiment has high ecological validity. In this study, we reached participants in real working conditions and tested more context-specific hypotheses. Disadvantages of the field experiments such as poor control over the environment and a need to follow organizational restrictions are difficult to overcome. We also made an attempt to conduct an online experiment with software programmers by creating sessions which are closely related to real work settings, yet under control. Materials, tasks and time availability used in the experiment were similar to the target environment. Yet, real scenarios can be more complex due to the diversity of information work.

Temporal validity entails the extent to which the obtained findings can be generalized across time. Data in the field experiment was collected at multiple points in time. The developed design increased our understanding of how the relationships between complementary factors and individual IT-enabled productivity unfold over time. Therefore, we can claim that the temporal va-lidity of the results is high. Data in the online experiment was collected over three-time points, mostly during the same day. The results were collected over

150

a period of four months (October – January). There is no evidence that produc-tivity of software programmers can vary according to season. Even though such evidence would exist, the participants were from all over the world where seasons differ. Participants could also perform assignments twenty-four hours a day. Therefore, temporal validity in this study is also assured.

We made attempts to ensure that our sample groups are representative. In Study 1 we had access to 91 participants, from which only 27 were involved in conditions where all complementary factors were studied together. In Study 2, 113 participants completed the online experiment. Although a-priori sam-ple sizes were identified, further studies with more participants involved in complementarity set-ups could enhance population validity. We have to acknowledge that in Study 1 we used convenience sampling rather than ran-dom sampling. In Study 2, we ensured enhanced external validity by provid-ing international access to the experiment. Professionals who took part in the experiment enforced realism and representativeness of the target population.

4.5 Methodological limitations The intention of this research was to identify configurations of complemen-tary factors that affect IT-enabled productivity based on the complementarity theory by conducting both a longitudinal quasi-randomized field experiment and online experiment. Two studies were performed in order to compensate the weaknesses inherent in the use of any single research design. For example, Camerer and Weber (2013, p. 213) claim that: “Relative to field studies using empirical data, experiments often have obvious advantages, especially that of control and randomized assignment to implement theoretical assumptions that can only be imperfectly measured and controlled econometrically when using field data”. Yet, McGrath (1981) emphasizes that all empirical designs are subject to inherent limitations and it is always desirable: (i) to increase generalizability of the target population, (ii) to improve precision in measure-ment of variables and control of the behavior variables, and (iii) to provide existential realism for participants. Therefore, methodological limitations of these studies are discussed below.

The first limitation is evident from the complementary approach that itself has some limitations (Brynjolfsson & Milgrom, 2013). These include the need for a suitably homogenous population, the narrowness of the production function and the impact of unobservable factors that may have a critical role. We addressed those limitations in the following way. In both studies, we included controls for traditional demographic variables (e.g. age, gender, a level of education and experience) in order to take into account heterogeneity among the participants. By studying two different information-intensive

151

occupations in two different studies, we were able to specify precise production functions and measure the effect of complementary factors on individual productivity of sales representatives and software programmers. This narrow focus allowed us to eliminate many sources of heterogeneity that could influence productivity estimation.

The second limitation relates to the number of participants involved in both studies. The data collected during the field and online experiments may be representative for generalization, yet, increased number of participants, types of information-intensive occupations and organizations may be advantageous. Our field and online experiments were designed in a way that the approximately equal number of participants was involved in each treatment condition. However, the field experiments are often criticized for the lack of randomization to treatment conditions (Harrison, 2011). As a result, observed differences in the dependent variables between individuals may be due to differences in the types of participants and groups rather to the effect of the independent variable. To avoid this limitation, we controlled for demographic variables and allocated four different design patterns into four-by-four operational structure of the company. In the online experiment to provide generalizability across people and situations (Aronson et al., 2007) we controlled for demographic variables and designed the experiment in such a way as to provide realistic working conditions in software construction.

The third limitation is evident from the precision of variables’ measure-ment and the extent of a model. In this research, a parsimonious model with realistically developed measurements was tested. We believe that measures developed for both studies are sufficient and rigorous in the research context since they are based on previous theoretical sources and reflect, as closely as possible, the real world. However, other potential avenues for future research might include an examination of a set of additional individual characteristics and complementarities that hypothetically may affect individual productivity of information workers. In order to avoid the aforementioned limitations, we used complementarity factors that were identified in previous research as the most relevant for productive IT use. Future research can explore whether the inclusion of additional complementary factors may increase individual productivity of information workers. For example, according to the person-environment fit theory (Edwards, 1991), there are different dimensions of fit between individual and working environment. For example, person-vacation fit, person-job fit, person-group fit, person-person fit may be potential com-plementarities that can affect individual performance of information workers.

The fourth limitation relates to the approach of the data analysis, including DID analysis and a repeated measures ANOVA. Although DID analysis is recognized as a powerful statistical technique (Angrist & Pischke, 2008; Blundell & Dias, 2009), it has some drawbacks. For example, if the parallel trend assumption does not hold for pre-treatment outcomes in control and treatment groups, the DID estimator is inaccurate (Angrist & Pischke, 2008).

152

The data collected for both the number of sales calls and products sold support the parallel trend assumption before the introduction of a new IT system. Therefore, we may expect that both dependent variables are affected by the intervention. The repeated measures ANOVA also has some limitations, including carry-over, practice and fatigue effects (Smith et al., 2008). Yet, the repeated measures ANOVA remains a powerful technique for examining changes over time.

It is well-known that the choice of the research approach affects data collection, data analysis and threats to validity (Field, 2011). In this research, we have endeavored to avoid methodological monism and provide rigorous operationalization to enhance the validity of the investigation and its analysis. A combination of both methods provides us with a high level of control under-researched variables, helps us evaluate the accuracy of the research model and validation of the obtained results. Other approaches may possibly develop and/or refine existing complementarity factors and their interrelationships to enhance the potential of managerial work to drive individual productivity of information workers.

4.6 Research ethics One of the sensitive dilemmas in research arises in establishing a connection and receiving consent from research subjects and companies to conduct research. Usually, this dilemma is based on how to protect research subjects from disclosure of information and to maintain confidentiality. This dilemma has been considered carefully in this research. First of all, connection with the potential company of the research interest was established. A special nondisclosure agreement has been signed between the researcher and the company. The company was ensured in confidentiality of received data. At the same time, the company was guaranteed that all obtained data will be stored in a safe place free from disclosure. Second of all, subjects involved in an online study were informed that each subject’s privacy and confidentiality will be protected. Moreover, subjects were guaranteed that the risk of access by unauthorized persons to research materials will be minimized.

1. Ethical considerations in data generating

The company under investigation was guaranteed that the data collection procedure in the research process is completely confidential and employees involved in a longitudinal quasi-randomized field experiment are completely anonymous. No company and personal names were used in any material and

153

discussions regarding the research. Instead, specific identification numbers were developed by the company. This means that the researcher did not know the identity of the participants, i.e. a survey questionnaire was returned with no names on it. Subjects in an online experiment were informed about the purpose, methods and future use of the research and what their participation in research entails. The results from this research were not produced in such a manner that they can be used to control and command participants.

2. Ethical considerations in data analyzing and interpreting

One of the ethical issues which deserves particular attention is data analysis and interpretation. The researcher of this study was responsible for ensuring that the results of the research are based on accurate representation of all available and relevant information collected during the period of study. Data interpreting fully protects the anonymity of participants and the confidentiality of the company. Research has been conducted without putting subjects in a compromising position. The results of the studies are interpreted on the independent basis.

3. Ethical considerations in research disseminating:

The primary focus of any research is to protect subjects. Yet, the researcher must inform the scientific community and other groups of interest about the methods and results of the study. In order to assess the validity of research conclusions and ethical procedures, information how the research was conducted is needed. Therefore, this research includes a detailed methodology section supplemented by additional available information, containing the research instrument and research procedures. Moreover, the act of publishing is an important element in maintaining honesty and openness, since the research community can review research procedures, discuss unclear points, and therefore be informed about research results. Therefore, the act of reporting and publishing of research results is one more component of social value. Some chapters of this dissertation have been reported and published in conference proceedings, and can be accessed by the public at any time.

4. Ethical considerations in reflecting upon research conducted

In general, being self-critical and honest about own research and its findings potentially contributes to the advancement of knowledge. The intention of this study was to produce useful results and increase knowledge in the researched phenomenon. Moreover, the research was designed, conducted and undertaken to guarantee quality and transparency. Given the above, the research has been conducted carefully, independently and without any conflicts of interest.

154

155

5. Data analysis and research findings

In this chapter, we present results from the data analysis of the two studies conducted. The chapter starts with a presentation of the analysis of the findings from a longitudinal quasi-randomized field experiment of sales representative productivity followed by the findings obtained from the online experiment of software programmer productivity. We conclude with the research findings from both studies.

5.1 Study 1: Sales representative productivity

In this section, we present the results of the DID analysis in the following order. First, the underlying assumptions of DID analysis are tested and dis-cussed. Second, the results of the DID analysis are presented and interpreted. We conclude with the results of a robustness check and demonstrate the per-sistence of the intervention effect.

5.1.1 Assessment of main assumptions In order to conduct the DID analysis, a number of criteria (assumptions and restrictions) needed to be taken into consideration. A summary of assumptions and restrictions for DID technique (Angrist & Pischke, 2008; Field, 2011) is presented in Table 5.1 with regard to the available data. First, dependent variables must be measured on a continuous scale. Second, two or more independent variables must be either continuous or categorical. Third, linearity, independence, no perfect multicollinearity, normality, and homoscedasticity are all assumptions that cannot be violated if a reliable DID estimator is to be obtained. Last but not least, DID analysis requires a parallel trend assumption, meaning that the outcome trend is similar in treatment and control groups prior to the intervention.

156

Table 5.1: Criteria (assumptions and restrictions) for performing DID analysis

Assumptions and restrictions

Assessment Criteria for the valida-tion of assumptions

Two or more independent variables are either continuous or categorical

Assumption is valid

Research design

Dependent variable measured on a continuous scale14

Assumption is valid

Research design

A linear relationship between the de-pendent variable and independent variables

Assumption is valid

Research design

Independence of observations Assumption is valid

Research design

No perfect multicollinearity Assumption is valid

Correlation matrix and VIF diagnostics

Normal distribution of residuals Assumption is valid

Skewness and kurtosis of residuals, normal distribu-tion histograms

Homoscedasticity Assumption is valid

Scatterplots of standard-ized residuals

Parallel trend Assumption is valid

Graphical visualization

The first two assumptions for the DID analysis have been met according

to the research design with four different groups (designs/configuration set-ups) and two dependent variables (a number of calls and products sold) that are measured on a continuous scale. Linearity assumption was automatically satisfied since the main independent variable (the interaction term between treatment group and a period before and after the intervention) is a dummy variable. The assumption of independence was satisfied at the stage of study design development. We ensured that different participants were assigned to different treatment groups. No perfect multicollinearity assumption was examined by two diagnostics, including a correlation between independent variables and the variance inflation factor (VIF). Table 5.2 represents a bivariate correlation between independent variables and VIF diagnostic.

14 The first-order dependent variable – the number of sales calls includes non-integer values and is calculated as the sum of conducted face-to-face meetings by an individual divided by a number of working days in a certain period of time (in our case a quarter). The second-order dependent variable – the number of products sold is presented by reasonably large numbers (mean = 957). Neither variable includes zero-values. Therefore, we estimate all regressions using ordinary least squares (OLS) regressions.

157

Table 5.2: Correlation matrix and VIF diagnostics

2 3 4 5 6 7 8 9 10 11

1.Treat2*Post -0.12** -0.12** 0.12 -0.04 -0.03 0.05 0.04 0.05 0.02 -0.02 2.Treat3*Post 1 -0.14** 0.02 -0.03 -0.04 0.04 0.01 -0.05 0.01 0.05

3.Treat4*Post 1 -0.03 0.05 0.06* -0.08** -0.08** -0.01 0.01 -0.07**

4.Gender 1 -0.10** 0.13** 0.11** 0.06* -0.02 -0.01 -0.26** 5.Year born 1 -0.43** 0.34** -0.88** -0.71** -0.68** 0.08** 6.Marital status 1 -0.29** 0.39** 0.35** 0.24** -0.07** 7.Education 1 -0.39** -0.49** -0.33** 0.20**

8.Exp.Ind 1 0.72** 0.71** -0.02 9.Exp.Sales 1 0.75** -0.08** 10.Exp.Comp 1 -0.02 11.KAI 1 VIF 1.09 1.13 1.19 5.49 1.14 1.54 8.88 4.76 2.82 1.33

Note: **Correlation is significant at the 0.01 level *Correlation is significant at the 0.05 level

158

The results presented in Table 5.2 demonstrate that most of the correlation coefficients lie below a critical value of 0.8. However, year born is highly correlated with the experience in the industry. This correlation leads to VIF values that are above a critical value of 5 (Oyana & Margai, 2015). To overcome the problem of multicollinearity it is recommended that one of the highly correlated variables be removed from the analysis (Field, 2011). Indeed, when the year born is removed from the analysis, correlation coefficients and VIF value lie below critical values. Therefore, without year born as a control variable, the no perfect multicollinearity assumption was met.

The normal distribution of residuals assumption was examined based on skewness and kurtosis statistics (Table 5.3).

Table 5.3: Skewness and kurtosis of residuals

Statistic Zre_1 (Calls) Mean Minimum Maximum Range Skewness Kurtosis

0.0000000 -2.44801 3.91887 6.36687 0.417 0.051

Zre_1 (Sales) Mean Minimum Maximum Range Skewness Kurtosis

0.0000000 -2.15997 3.43518 5.59514 0.375 0.210

The obtained results demonstrate that the skewness and kurtosis statistics

of the residuals lie within the range between – 2 and + 2 (Westfall & Henning, 2013). This suggests that the residuals from the regression appear to conform to the assumption of being normally distributed. Figure 5.1 also demonstrates no serious departure from the bell–shaped curve. Therefore, the assumption of normality has been met.

159

Figure 5.1: Normal distribution histograms

The homoscedasticity assumption implies that the dependent variable

demonstrates similar amounts of variance across the range of independent variables (Field, 2011). This assumption is checked using a scatterplot between residuals and independent variables (Figure 5.2).

Figure 5.2: Scatterplots of standardized residuals The above scatterplots of standardized residuals against predicted values

demonstrate a pattern centered around the line of zero standard residual value. The points have the same dispersion around the line over the predicted value range. There is no clear relationship between the residuals and the predicted values which is consistent with the assumption of homoscedasticity.

Finally, DID estimators critically rely on the parallel trend assumption. This assumption states that the control and treatment groups should follow the

160

same trend in outcome before the intervention (Angrist & Pischke, 2008). The credibility of this assumption is maintained by visual inspection. We assess this assumption graphically by investigating pre-intervention data from the years 2012 till 2014. For each dependent variable, a graph was constructed by taking the mean across nine quarters for control and treatment groups. Figure 5.3 presents pre-intervention curves for control and treatment groups and both dependent variables15.

Figure 5.3: Parallel trend assumption for a number of calls and products sold before intervention

Our data demonstrate that pre-intervention outcome trends are parallel for

control and treatment groups, maintaining the same differences between groups over time. Control and treatment groups appear to follow common trends. Since both graphs demonstrate support for the parallel trend assumption, there is no reason to suspect that groups under investigation can be affected differently by the intervention. Therefore, we can expect that DID analysis will show a direct impact of the intervention on different treatment groups. The results of the DID test are presented below.

5.1.2 The intervention (operational changes) effect We used DID design to estimate the impact of the introduction of a more aligned IT system together with complementarities on a number of sales calls and products sold. We used the number of sales calls (face-to-face meetings) 15 Design 1: no change; Design 2: structured partial change; Design 3: semi-structured partial change; Design 4: full change.

161

in relation to the duration of time worked and the number of products sold as dependent variables. Our variables of interest, Treati * Postt, are represented by a dummy variable that can be described as the interaction between a dummy variable that indicates whether individual i is in the control or treat-ment group (Treati) and a dummy variable that indicates whether the observa-tion corresponds to the period after the introduction of the more aligned IT system (Postt,). We conducted DID analysis in three phases. First, we analyzed the impact of the intervention on the dependent variables in all four treatment (design) groups. Second, we compared individual outcome changes in treat-ment (design) groups two and four. Third, we conducted the same analysis, but we compared individual outcome changes in treatment (design) groups three and four. Fourth, we compared individual outcome changes in treatment (design) groups two and three. All phases of analysis have been conducted to ensure the emergence of complementary patterns.

We carried out DID analysis in all four treatment groups for the number of sales calls. Since we have four treatment groups, we created three dummy variables using Design one (less aligned IT system) as a reference category. The results are presented in Table 5.4. For the number of sales calls, we first tested the model (Model 1) without control variables. We then added a vector of individual demographic variables (Model 2). For the third model, we in-cluded the year the product was launched on the market and the market mean of sales calls instead of demographic variables (Model 3). Finally, we tested the model with all control variables (Model 4). The models include country dummies as well as product dummies. If the effect of the treatment is not im-pacted by the introduction of the control variables, we can conclude that the results of the study are due to the effect of the treatment being tested. Table 5.4: DID results demonstrating the impact of the intervention (treatment) on sales calls, controlling selected covariates (Design 1, 2, 3 and 4)

Model 1 Model 2 Model 3 Model 4 Constant 1.641***

(0.055) 4.211*** (0.170)

-81.175** (30.148)

14.741 (27.940)

Treat2*Post -0.369*** (0.092)

-0.369*** (0.084)

-0.218** (0.085)

-0.217** (0.079)

Treat3*Post -0.393*** (0.087)

-0.393*** (0.080)

-0.345** (0.082)

-0.183** (0.075)

162

Table 5.4: Continued from previous page

Model 1 Model 2 Model 3 Model 4 Treat4*Post 0.649***

(0.085) 0.649*** (0.077)

0.557*** (0.082)

0.538*** (0.072)

Gender -0.022 (0.031)

-0.004 (0.029)

Marital status -0.070** (0.034)

-0.105** (0.032)

Level of education -0.162*** (0.030)

-0.124*** (0.028)

Years of experience in sales -0.002 (0.009)

0.003 (0.008)

Years of experience in the company

-0.028** (0.013)

-0.035** (0.013)

Years of experience in indus-try

0.006 (0.006)

0.010* (0.005)

KAI -0.023*** (0.002)

-0.026*** (0.001)

Year when the product was launched

0.040** (0.015)

-0.006 (0.014)

Market mean 0.949*** (0.080)

1.142*** (0.072)

Country dummies Included Included Included Included Product dummies Included Included Included Included Observations 91 91 91 91 Adjusted R2 0.284 0.401 0.340 0.481

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level; Standard errors are in parentheses.

As depicted in Table 5.4, our results show that the number of sales calls for individuals involved in Design 4 remains unaffected by individual control variables and statistically significant 0.649 (p < 0.001). Year of product launch and market mean made an impact on the DID estimator by reducing its value to 0.557 (p < 0.001). This implies that not only operational changes but also external forces in the market place have an effect on sales calls. However, the external forces did not change the DID estimator significantly. These findings lead to the conclusion that the introduction of the more aligned IT system, together with a “full” set of complementarities, has had a positive and statis-tically significant effect on the number of sales calls compared to the control group (Design 1). In contrast to these findings, the number of sales calls in Design 2 and 3 has decreased compared to the control group -0.369 (p < 0.001) and -0.393 (p < 0.001) respectively.

163

In Figure 5.4 we visually illustrate the effect of the intervention on a num-ber of sales calls in all four designs (complementarity set-ups). The aim of the figure is to visually demonstrate the more robust regression analysis presented in Table 5.4. Each point in the graphs represents the average number of calls. 95% confidence intervals are represented by the dotted lines.

Figure 5.4: Intervention effect on a number of sales calls

164

The impact of the intervention on the number of sales calls represented in Figure 5.4 suggests a persistent negative effect in Design 2 and 3. In contrast, a persistent positive effect is observed in Design 4. Since a less aligned IT system has been used in Design 1, there is no persistent effect of the interven-tion on the number of sales calls.

More detailed DID analysis between pairs of designs is presented in Table 5.5. First, we compared the impact of the intervention on the sales calls of sales representatives involved in Design 2 and 4. Then we carried out the same analysis for sales representatives involved in Design 3 and 4 and finally we conducted DID analysis to compare the impact of the intervention in Design 2 and 3.

Table 5.5: DID results demonstrating the impact of the intervention (treatment) on sales calls between pairs of designs

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level; Standard errors are in parentheses.

The results in Table 5.5 demonstrate that the introduction of the more aligned IT system, together with a “full” set of complementarities, has had a positive and statistically significant effect on the number of sales calls 1.018 (p < 0.001) for sales representatives involved in Design 4 compared to Design 2. The introduction of the intervention has also had a positive and statistically significant effect on the number of sales calls 1.042 (p < 0.001) for sales representatives involved in Design 4 compared to Design 3. There was no statistically significant effect of the intervention on the number of sales calls between sales representatives involved in Design 2 and 3. These results remain unaffected by control variables and therefore demonstrate that sales representatives involved in Design 4 with a “full” complementarity set-up performed better in comparison to sales representatives involved in Design 1, 2, and 3 after the introduction of operational changes.

In general, the number of sales calls has decreased by 0.2, or approximately 10% after the intervention for sales representatives involved in Design 1. The

Design 2 and 4 Design 3 and 4 Design 2 and 3

Constant 1.504*** (0.060)

1.361*** (0.054)

1.881*** (0.095)

Treat*Post 1.018*** (0.084)

1.042*** (0.073)

-0.025 (0.085)

Country dummies Included Included Included Product dummies Included Included Included Observations 91 91 91 Adjusted R2 0.382 0.414 0.223

165

number of sales calls has reduced by 0.5 or approximately 23% after the in-tervention in Design 2 (the use of the more aligned IT system without com-plementarities). There was a 0.6 or 29% decrease in the number of sales calls in Design 3 (the use of partial IT use complementarities). In Design 4, the use of the more aligned IT system together with matched complementarities has increased the number of sales calls by 0.5 or approximately 31% of the pre-intervention average.

To summarize, the obtained results suggest that the productivity improve-ment related to the number of sales calls occurred mostly in Design 4 where the “full” set of complementarities has been introduced. Our results also demonstrate that when the more aligned IT system is used without any com-plementarities the opposite (negative) effect is produced. Moreover, the re-sults show that the more aligned IT system use with a limited number of com-plementarities can also have a negative effect on productivity.

For the analysis of the intervention effect on the number of products sold, we use the same logic as for the number of sales calls. By applying DID analysis, we analyzed four models with and without control variables. The results are presented in Table 5.6.

Table 5.6: DID results, demonstrating the relationship between the number of products sold and treatments, controlling selected covariates (Design 1, 2, 3 and 4)

Model 1 Model 2 Model 3 Model 4

Constant 604.573*** (28.476)

-283.838** (89.041)

-27666.026** (11504.245)

-2970.765 (11191.771)

Treat2*Post -17.367 (47.780)

-17.367 (44.235)

-17.367 (47.780)

-17.367 (44.235)

Treat3*Post -94.181** (45.091)

-94.181** (41.746)

-94.181** (45.091)

-94.181** (41.746)

Treat4*Post 87.235** (43.907)

87.235** (40.649)

87.235** (43.907)

87.235** (40.649)

Gender 56.756*** (16.110)

56.756*** (16.110)

Marital status 118.390*** (17.912)

118.390*** (17.912)

Level of educa-tion

-134.257*** (15.493)

-134.257*** (15.493)

166

Table 5.6: Continued from previous page

Model 1 Model 2 Model 3 Model 4 Years of experi-ence in sales

-9.442** (4.693)

-9.442** (4.693)

Years of experi-ence in the com-pany

38.923*** (7.080)

38.923*** (7.080)

Years of experi-ence in industry

-13.849*** (2.891)

-13.849*** (2.891)

KAI 10.648*** (0.823)

10.648*** (0.823)

Year the product was launched

-13.484** (5.727)

1.339 (5.566)

Country dummies Included Included Included Included Product dummies Included Included Included Included Observations 91 91 91 91 Adjusted R2 0.835 0.858 0.835 0.858

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level; Standard errors are in parentheses.

As depicted in Table 5.6, our results show that the number of products sold

for individuals involved in Design 4 remains unaffected by individual control variables and statistically significant 87.235 (p < 0.05). Year of product launch did not make an impact on DID estimator. These findings lead to the conclu-sion that the introduction of the more aligned IT system, together with a “full” set of complementarities, has had a positive and statistically significant effect on the number of products sold compared to the control group (Design 1). In contrast to these findings, the number of products sold in Design 3 has de-creased compared to the control group -94.181 (p < 0.05).

On average, the number of products sold has increased by 81 units or ap-proximately 14% after the intervention for sales representatives involved in Design 1. The number of products sold has increased by 64 units or approxi-mately 5% after the intervention in Design 2 (the use of the more aligned IT system without complementarities). There was a 12 unit or 1% decrease in the number of products sold in Design 3 (the use of partial IT use complementa-rities). In Design 4, the use of the more aligned IT system together with matched complementarities has increased the number of products sold by 168 units or approximately 30% of the pre-intervention average. The intervention effect on the number of products sold in all four designs (complementarity set-ups) is illustrated in Figure 5.5.

167

Figure 5.5: Intervention effect on a number of products sold

168

Overall, the impact of the intervention on the number of products sold as represented in Figure 5.5 suggests a persistent negative tendency in Design 2 and 3. In contrast, there is a positive tendency in Design 4. There is no persis-tent effect of the intervention on the number of products sold in Design 1 since a less aligned IT system has been used. Although DID analysis provides sig-nificant results demonstrating the complementarity effect of individual productivity, we also conducted a robustness check to ensure that differences in the number of sales calls and products sold occurred due to the actual inter-vention. Persistence of the intervention effect will be demonstrated in the next sub-section.

5.1.3 Robustness check and persistence of the intervention effect We conducted a robustness check by using a methodology that aims to ensure that pre-treatment time trends for the number of calls and products sold were similar for control and treatment groups. This robustness check is similar to the parallel trend assumption which determines whether or not differences exist in the pre-intervention trends. This robustness check helps us ensure that differences in the dependent variables across treatment and control groups before the actual intervention are non-existent or small, decreasing the likelihood that the estimated treatments are biased. Time placebo regressions (Winship & Morgan, 1999; Bertrand et al., 2004) are estimated with and without covariates using data from control and treatments groups. We randomly assigned actual intervention dates (quarter 3, year 2012; quarter 1, year 2013; and quarter 3, year 2013) and repeated our primary model for each dependent variable. The results of the robustness check are presented in Table 5.7. Table 5.7: Robustness check

Time placebo

Without individual covariates

With individual covariates

Calls Sales Calls Sales

2012Q3

Treat2*Post 0.075 (0.158)

181.032 (72.743)

0.075 (0.146)

181.032 (66.759)

Treat3*Post 0.043 (0.149)

26.469 (68.649)

0.043 (0.138)

26.469 (63.003)

Treat4*Post 0.067 (0.145)

-0.002 (66.846)

0.067 (0.134)

-0.002 (61348)

Adjusted R2 0.276 0.869 0.383 0.889

169

Table 5.7: Continued from previous page

Time placebo

Without individual covariates

With individual covariates

Calls Sales Calls Sales

2013Q1

Treat2*Post 0.038 (0.133)

155.928 (60.433)

0.038 (0.122)

155.928 (55.384)

Treat3*Post -0.057 (0.125)

33.853 (57.032)

-0.057 (0.116)

33.853 (52.268)

Treat4*Post 0.055 (0.122)

-9.828 (55.534)

0.055 (0.113)

-9.828 (50.894)

Adjusted R2 0.272 0.870 0.378 0.891

2013Q3

Treat2*Post 0.023 (0.139)

123.518 (64.307)

0.023 (0.129)

123.518 (59.045)

Treat3*Post -0.078 (0.131)

37.443 (60.688)

-0.078 (0.121)

37.443 (55.723)

Treat4*Post 0.013 (0.128)

-3.924 (59.094)

0.013 (0.118)

-3.924 (54.259)

Adjusted R2 0.276 0.868 0.383 0.889 Country dummies

Included Included Included Included

Product dummies

Included Included Included Included

Observa-tions

91 91 91 91

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level; Standard errors are in parentheses.

The results from the time placebo regression demonstrate that, because no

treatment was actually applied, there is no significant effect of the false time treatment (set at quarter 3, year 2012; quarter 1, year 2013; and quarter 3, year 2013) between control and treatment groups, supporting the primary results for both the number of sales calls and products sold. Following the robustness check, we established evidence that a change in outcome variables is not an artifact of seasonal or any other trends. Therefore, we can claim that the inter-vention caused changes in outcome variables.

We also tested the DID specification to estimate the effect of the interven-tion at different time points after the more aligned IT system together with complementarities was implemented, to demonstrate the persistence of the in-tervention impact. Specifically, we used time placebo assigned in the quarter 4, year 2014 and quarter 2, year 2015. The results are presented in Table 5.8.

170

Table 5.8: Persistence of the intervention impact (time placebo set in the quarter 4, year 2014 and quarter 2, year 2015)

Time placebo

Without individual covariates

With individual covariates

Calls Sales Calls Sales

2014Q4

Treat2*Post -0.508*** (0.092)

-103.405** (48.891)

-0.508*** (0.083)

-103.405** (45.246)

Treat3*Post -0.549*** (0.086)

-172.161*** (46.140)

-0.549*** (0.078)

-172.161*** (42.700)

Treat4*Post 0.781*** (0.084)

108.463** (44.928)

0.781*** (0.076)

108.463** (41.578)

Adjusted R2 0.328 0.836 0.446 0.859

2015Q2

Treat2*Post -0.501*** (0.100)

-179.692** (52.747)

-0.501*** (0.091)

-179.692** * (48.739)

Treat3*Post -0.523*** (0.094)

-251.837** *(49.778)

-0.523*** (0.086)

-251.837** * (45.996)

Treat4*Post 0.886*** (0.092)

133.818** (48.471)

0.886*** (0.084)

133.818** (44.788)

Adjusted R2 0.323 0.838 0.441 0.862

Country dummies

Included Included Included Included

Product dummies

Included Included Included Included

Observa-tions

91 91 91 91

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level; Standard errors are in parentheses.

The DID results from post-treatment quarters demonstrate changes in the size treatment effects across time. Treatment effects are consistent and significant for the number of sales calls. For example, when the time placebo is assigned to the quarter 4 year 2014 our results remain unaffected and statistically significant -0.508 (p < 0.001) for individuals involved in Design 2, -0.549 (p < 0.001) for individuals involved in Design 3, and 0.781 (p < 0.001) for individuals involved in Design 4 when compared with individuals involved in Design 1. When the time placebo is assigned to quarter 2 year 2015, the results remain unaffected and statistically significant -0.501 (p < 0.001) for individuals involved in Design 2, -0.523 (p < 0.001) for individuals involved in Design 3, and 0.886 (p < 0.001) for individuals involved in Design 4 in comparison with individuals involved in Design 1.

171

Treatment effects are also consistent and significant for the number of sales calls. For example, when the time placebo is assigned to the quarter 4 year 2014 our results remain unaffected and statistically significant -103.405 (p < 0.05) for individuals involved in Design 2, -172.161 (p < 0.001) for indi-viduals involved in Design 3, and 108.463 (p < 0.05) for individuals involved in Design 4 when compared with individuals involved in Design 1. When the time placebo is assigned to quarter 2 year 2015, the results remain unaffected and statistically significant -179.692 (p < 0.001) for individuals involved in Design 2, -251.837 (p < 0.001) for individuals involved in Design 3, and 133.818 (p < 0.05) for individuals involved in Design 4 in comparison with individuals involved in Design 1.

5.2 Study 2: Software programmer productivity In the online experiment, we analyzed how ”stable” and ”dynamic” comple-mentarity set-ups affect productivity of individuals with innovative cognitive style when a more aligned IT system is used. In this section, a chain of com-plementarity evidence has been developed based on a repeated measures ANOVA technique. The choice of the repeated measures ANOVA test is ex-plained by the research interest in group differences (intervention), rather than by the description (association) and explanation (prediction) over time (Ran-dolph & Myers, 2013). In the online experiment, we conducted a more de-tailed comparison of how the performance of software programmers involved in different complementarity set-ups has changed during three sessions. The repeated measures ANOVA makes this comparison possible since this tech-nique demonstrates how the dependent variables change over time across treatment groups (Field, 2011). A significant influence of a particular comple-mentarity set-up denotes the emergence of a potentially effective IT use pat-tern. Second, since the quality of the developed software and time were not statistically correlated, the repeated measures ANOVA is recommended for evaluation of the main hypotheses (Mayer, 2013). The underlying assump-tions of the repeated measures ANOVA and an interpretation of the obtained results are presented below.

172

5.2.1 Assessment of main assumptions In order to conduct the repeated measures ANOVA, a number of criteria (assumptions and restrictions) need to be analyzed in advance. A summary of the assumptions and restrictions necessary to make the repeated measures ANOVA valid (Field, 2011) is presented in Table 5.9 with regard to the available data. Table 5.9: Criteria (assumptions and restrictions) for the repeated measures ANOVA Assumptions and restrictions Assessment Criteria for the

validation of assumptions

At least two categorical independent variables, one of which is between-group and one within-group, at least two levels

Assumption is valid Research design

The dependent variable(s) must be continuous

Assumption is valid Research design

The dependent variable(s) must be normally distributed across the independent groups and over the within-group conditions

The distribution is slightly abnormal

Skewness and kurtosis

No significant outliers Assumption is valid Boxplots

Homogeneity of between group variance

Assumption is valid Levene’s test

Homogeneity of variance-covariance matrices

Assumption is valid Box’s M test

Sphericity of within-group variance Assumption is valid for the first dependent variable (time)

Mauchly’s test

First, the independent variables must be categorical with at least two

groups and the dependent variables must be continuous. These first two as-sumptions for the repeated measures ANOVA have been met according to the research design with two groups (complementarity set-ups), three sessions (tasks) and two continuous dependent variables (time and quality).

173

Second, the dependent variables must be reasonably normally distributed across the independent groups and over the within-group conditions. We ap-pear to have a slightly abnormal distribution in both dependent variables ac-cording to the data skewness and kurtosis (Table 5.10). Table 5.10: Skewness and kurtosis of the dependent variables

Time

Assignment 1 Stable Skewness -0.523

Kurtosis -0.778

Dynamic Skewness 0.622 Kurtosis -0.657

Assignment 2 Stable Skewness -0.239

Kurtosis -0.865

Dynamic Skewness -0.688 Kurtosis -0.616

Assignment 3 Stable

Skewness -0.532 Kurtosis -0.857

Dynamic Skewness -0.395 Kurtosis -0.985

Quality

Assignment 1 Stable

Skewness -0.925

Kurtosis -0.211

Dynamic Skewness -0.114 Kurtosis -1.418

Assignment 2 Stable Skewness -0.557

Kurtosis -0.842

Dynamic Skewness -0.004 Kurtosis 0.800

Assignment 3 Stable Skewness -0.552

Kurtosis -1.242

Dynamic Skewness -0.100 Kurtosis -1.517

The values for skewness and kurtosis between – 2 and + 2 are considered

acceptable to establish normal distribution (Westfall & Henning, 2013). Moreover, the ANOVA test with balanced sample sizes is ‘robust’ to moderate departure from normality (Gray & Kinnear, 2012). No significant outliers have been identified in the data set by inspecting boxplots. Therefore, we may expect that the repeated measures ANOVA results are valid.

Third, groups in the ANOVA analysis must have the same or similar variances. The most common assessment of the homogeneity of variance assumption is Levene’s test (Table 5.11). In our case, Levene’s test is insignificant and, therefore, our data meet the assumption of variance equality, meaning that there is no difference between the variances in the population.

174

Table 5.11: Test of homogeneity of variances

Measure Assignment Levene Statis-tic

df1 df2 Sig.

Time Assignment 1 3.465 1 86 0.066 Assignment 2 0.282 1 86 0.597 Assignment 3 0.028 1 86 0.868

Quality Assignment 1 1.224 1 86 0.272 Assignment 2 2.204 1 86 0.141 Assignment 3 0.151 1 86 0.698

Fourth, homogeneity of variance-covariance matrices is measured with

Box’s M test, which demonstrates whether the correlation between the de-pendent variables is significantly different between groups (Mayers, 2013). The Box’s M test of equality of covariance matrices was insignificant, p (0.358) > (0.001) and p (0.672) > (0.001) for time and quality respectively. Therefore, the assumption of homogeneity of covariance across groups is not violated.

The last assumption known as sphericity has been analyzed by using a Mauchly’s test of sphericity. This test indicated that the assumption of sphe-ricity had not been violated, for the dependent variable – time χ2(2) = 3.224, p = 0.200. However, the assumption of sphericity had been violated for the de-pendent variable – quality χ2(2) = 9.010, p = 0.011. This implies that in order to interpret the output, the Greenhouse-Geisser correction is appropriate (Field, 2011).

Therefore, our data set met the main assumptions required to perform the repeated measures ANOVA for receiving valid results. The statistical output received from this test is presented and analyzed below.

5.2.2 Output interpretation A two-way repeated measure ANOVA (2x3) was conducted to assess the in-fluence of the complementarity set-ups on individual productivity (time and quality) of software programmers (alpha was set at p ˂ 0.05). Below, the re-sults are interpreted in the following way. First, we interpret the main effect for complementarity set-up (between-group independent variable), the main effect for assignment (within-group independent variable) and within-between interaction (complementarity set-up versus assignment). Second, we locate the source of interaction. Third, we analyze effect, size and power of the sta-tistical tests.

175

§ Main effect for complementarity set-up (between-group inde-pendent variable)

First, we identify the main effect for complementarity set-up (between-

group independent variable). Table 5.12 presents data on estimated marginal means for complementarity set-up. Table 5.12: Main effect for complementarity set-up

Dependent variable

Complementa-rity set-up Mean Std.

Error

95% confidence intervals Lower bound

Upper bound

Time Stable 46.197 1.652 42.913 49.481 Dynamic 43.212 1.652 39.928 46.496

Quality Stable 75.227 2.632 69.995 80.460 Dynamic 62.348 2.632 57.116 67.581

The data in the table suggests that innovators involved in a “stable” com-

plementarity set-up across all assignments spent slightly more time than inno-vators involved in a “dynamic” complementarity set-up. Innovators involved in a “stable” complementarity set-up across all assignments demonstrate higher quality than innovators involved in a “dynamic” complementarity set-up. Table 5.13 demonstrates whether there is a significant between-group dif-ference in productivity metrics. Table 5.13: ANOVA test for between-group differences Source Measure Type III

sum of squares

df Mean square

F Sig. Partial eta sq.

Intercept Time 527603.045 1 527603.045 1464.542 0.000 0.945

Quality 1249187.88 1 1249187.88 1365.953 0.000 0.941

Compl. set-up

Time 588.015 1 588.015 1.632 0.205 0.019

Quality 10946.970 1 10946.970 11.970 0.001 0.122

Error Time 30981.606 86 360.251

Quality 78648.485 86 914.517

The obtained results in Table 5.13 demonstrate that there is no significant

between-group difference, F (1, 86), = 1.632, p > 0.001 for time, but there is

176

a significant between-group difference F (1, 86), = 11.970, p = 0.001 for quality. This outcome demonstrates that there is a significant main effect of complementarity set-up on quality scores across all assignments.

§ Main effect for assignment (within-group independent variable)

After the main effect for complementarity set-up was established, we ana-

lyzed the main effect for assignment (Table 5.14).

Table 5.14: Assignment means

Measure Assignment Mean Std. error

95% confidence intervals Lower bound

Upper bound

Time 1 37.966 1.711 34.564 41.368 2 50.295 1.239 47.832 52.759 3 45.852 1.443 42.984 48.721

Quality 1 70.114 2.534 65.077 75.151 2 68.807 1.994 64.842 72.771 3 67.443 2.691 62.095 72.792

The results show that time scores appear to be highest for the second as-

signment (to-do list) followed by the third assignment (text analysis tool) and the first assignment (animation). The highest time scores for the second as-signment can be explained by the time lag which makes an impact on perfor-mance after the introduction of a new IT tool (Brynjolfsson & Hitt, 1993). Quality scores appear to be relatively stable across all assignments. However, before drawing any conclusions we need to analyze whether the observed dif-ference is significant by using the ANOVA test for within-group differences (Table 5.15).

177

Table 5.15: ANOVA test for within-group differences and interaction Source DV Type III sum

of squares df Mean

square F Sig. Partial eta

squared

Assignment Time Sphericity Assumed 6862.659 2 3431.330 31.802 0.000 0.270

Greenhouse-Geisser 6862.659 1.928 3559.024 31.802 0.000 0.270

Huynh-Feldt 6862.659 1.995 3440.706 31.802 0.000 0.270

Lower-bound 6862.659 1.000 6862.659 31.802 0.000 0.270

Quality Sphericity Assumed 313.826 2 156.913 0.492 0.612 0.006

Greenhouse-Geisser 313.826 1.817 172.694 0.492 0.594 0.006

Huynh-Feldt 313.826 1.876 167.296 0.492 0.600 0.006

Lower-bound 313.826 1.000 313.826 0.492 0.485 0.006

Assignment * Comple-mentarity set-up

Time Sphericity Assumed 1394.508 2 697.254 6.462 0.020 0.070

Greenhouse-Geisser 1394.508 1.928 723.202 6.462 0.020 0.070

Huynh-Feldt 1394.508 1.995 699.159 6.462 0.020 0.070

Lower-bound 1394.508 1.000 1394.508 6.462 0.013 0.070

178

Table 5.15: Continued from previous page

Source DV Type III sum of squares

df Mean square

F Sig. Partial eta squared

Quality Sphericity Assumed 77.462 2 38.731 0.122 0.886 0.001

Greenhouse-Geisser 77.462 1.817 42.626 0.122 0.867 0.001

Huynh-Feldt 77.462 1.876 41.294 0.122 0.874 0.001

Lower-bound 77.462 1.000 77.462 0.122 0.728 0.001

Error (as-signment)

Time Sphericity Assumed 18558.167 172 107.896

Greenhouse-Geisser 18558.167 165.829 111.912

Huynh-Feldt 18558.167 171.531 108.912

Lower-bound 18558.167 86.000 215.793

Quality Sphericity Assumed 54825.379 172 318.752

Greenhouse-Geisser 54825.379 156.282 350.811

Huynh-Feldt 54825.379 161.325 339.844

Lower-bound 54825.379 86.000 637.504

179

Table 5.15 shows that there is a significant main effect of assignment on time F (2, 172) = 31.802, p < 0.05. There is also the interaction effect on time between complementarity set-up and assignment F (2, 172) = 6.462, p < 0.05. In contrast, there is no significant within-group main effect or interaction main effect on quality. Table 5.16 demonstrates where the difference lies within the main effect.

Table 5.16: Post hoc test for within-group differences

(I) Assign-ment

(J) Assign-ment

Mean difference (I-J)

Std. error Sig.a

95% confidence intervals for differencea Lower bound

Upper bound

Tim

e

1 2 3

-12.330* -7.886*

1.617 1.660

0.000 0.000

-16.278 -11.939

-8.381 -3.833

2 1 3

12.330* 4.443*

1.617 1.409

0.000 0.007

8.381 1.002

16.278 7.884

3 1 2

7.886* -4.443*

1.660 1.409

0.000 0.007

3.833 -7.884

11.939 -1.002

Qua

lity

1 2 3

1.307 2.670

2.435 3.088

1.000 1.000

-4.638 -4.869

7.252 10.209

2 1 3

-1.307 1.364

2.435 2.504

1.000 1.000

-7.252 -4.751

4.638 7.479

3 1 2

-2.670 -1.364

3.088 2.504

0.072 1.000

-10.209 -7.479

4.869 4.751

a. Adjustment for multiply comparison: Bonferroni

The results in Table 5.16 demonstrate that time to perform assignment 1 is significantly lower than for assignment 2 (p ˂ 0.001) and assignment 3 (p ˂ 0.001). Time to perform assignment 2 is significantly higher than time to per-form assignment 3 (p ˂ 0.05). There is no significant difference in quality scores between the first, second and third assignments.

§ Within-between interaction (complementarity set-up versus as-

signment) After establishing main effects for complementarity set-ups and assign-

ments, we further demonstrate interaction means for within-between interac-tion (Table 5.17).

180

Table 5.17: Interaction means

Comple-mentarity set-up

Assignment Mean Std. error

95% confidence interval

Lower bound

Upper bound

Tim

e Stable

1 2 3

42.545 49.364 46.682

2.420 1.753 2.041

37.735 45.879 42.625

47.356 52.848 50.738

Dynamic 1 2 3

33.386 51.227 45.023

2.420 1.753 2.041

28.576 47.743 40.966

38.197 54.717 49.079

Qua

lity Stable

1 2 3

77.273 75.114 73.295

3.583 2.820 3.805

70.149 69.507 65.731

84.396 80.720 80.860

Dynamic 1 2 3

62.955 62.500 61.591

3.583 2.820 3.805

55.831 56.893 54.027

70.078 68.107 69.155

Table 5.17 shows that the difference in time scores across assignment 1

and 2 is much greater for innovators involved in a “dynamic” complementarity set-up than for innovators involved in a “stable” complementarity set-up. The same holds for assignment 2 and 3. This suggests that there is a within-be-tween interaction as it previously shown in Table 5.15: F (2, 172) = 6.462, p < 0.05. However, quality scores remain mostly flat across assignments and complementarity set-ups.

In order to illustrate the obtained results, line graphs are presented in Figure 5.6 below.

Figure 5.6: Estimated marginal means of time and quality scores

181

The obtained results demonstrate that, following the introduction of the more aligned IT system, time scores increased more for innovators involved in a “dynamic” complementarity set-up than for innovators involved in a “sta-ble” complementarity set-up. However, during the third assignment, innova-tors involved in a “dynamic” complementarity set-up appear to be more pro-ductive than innovators involved in a “stable” complementarity set-up. Quite different results are observed for quality scores. Overall, changes in quality scores were insignificant across assignments. However, according to esti-mated marginal means, quality drops were less “dramatic” for innovators in-volved in a “dynamic” complementarity set-up than for innovators involved in a “stable” complementarity set-up.

5.2.3 The source of interaction Since we found a significant interaction for time scores in the repeated measures ANOVA analysis of the data obtained from the online experiment, we explored the data further to illustrate the source of the interaction (Mayers, 2013). Follow-up tests also helped us determine whether any changes in the dependent variables occur due to one of the factors even when no interaction was identified before as it is in quality scores (Tybout et al., 2001). In order to establish the identified interaction, we explored the dependent variables with respect to complementarity set-up, using three independent t-tests: one for each assignment. We also examined the dependent variables across three as-signments, using a repeated measures one-way ANOVA both for “stable” and “dynamic” complementarity set-ups.

First, three independent t-tests (one for each assignment) were performed for both time and quality scores and the obtained results are presented in Table 5.18 below.

182

Table 5.18: Independent samples test for time and quality scores

Lev

ene’

s te

st

for

equa

lity

of v

aria

nces

t-test for equality of means

F Sig. t df Sig. (2-tailed)

Mean difference

Std. error difference

95% confidence interval of the dif-ference Lower Upper

Tim

e

A 1 Equal vari-ances assumed 3.465 0.066 2.676 86 0.009 9.159 3.422 2.356 15.962

Equal vari-ances not as-sumed

2.676 80.952 0.009 9.159 3.422 2.350 15.968

A 2 Equal vari-ances assumed 0.282 0.597 -0.752 86 0.454 -1.864 2.479 -6.791 3.064

Equal vari-ances not as-sumed

-0.752 85.762 0.454 -1.864 2.479 -6.791 3.064

A 3 Equal vari-ances assumed 0.028 0.868 0.575 86 0.567 1.659 2.886 -4.078 7.396

Equal vari-ances not as-sumed

0.575 85.881 0.567 1.659 2.886 -4.078 7.396

183

Table 5.18: Continued from previous page

Lev

ene’

s te

st

for

equa

lity

of v

aria

nces

t-test for equality of means

95% confidence interval of the dif-ference

F Sig. t df Sig. (2-tailed)

Mean difference

Std. error difference

Lower Upper

Qua

lity

A 1 Equal vari-ances assumed 1.224 0.272 2.825 86 0.006 14.318 5.068 4.244 24.392

Equal vari-ances not as-sumed

2.825 85.707 0.006 14.318 5.068 4.244 24.392

A 2 Equal vari-ances assumed 2.204 0.141 3.162 86 0.002 12.614 3.989 4.685 20.543

Equal vari-ances not as-sumed

3.162 82.336 0.002 12.614 3.989 4.685 20.543

A 3 Equal vari-ances assumed 0.151 0.698 2.175 86 0.032 11.705 5.381 1.007 22.402

Equal vari-ances not as-sumed

2.175 85.790 0.032 11.705 5.381 1.007 22.402

184

The results of the t-tests (adjusted p = 0.016) demonstrate that there is a significant difference in time scores in respect to complementarity set-up, for the first assignment (p = 0.009). Yet, for the second and third assignments differences are insignificant (p = 0.454 and (p = 0.567 respectively), meaning that the means in times scores between two groups of innovators involved in different complementarity set-ups do not differ statistically from each other. Yet, there is a significant difference in quality scores in respect to comple-mentarity set-up, for the first assignment (p = 0.006), the second assignment (p = 0.002), and the third assignment (p = 0.032), meaning that the means in quality scores between two groups of innovators involved in different com-plementarity set-ups are statistically different from each other.

Second, two one-way repeated measures ANOVA for both “stable” and “dynamic” complementarity set-ups were performed for time and quality scores and the results obtained are presented in Table 5.19.

185

Table 5.19: Repeated measures one-way ANOVA for assignment (reported by complementarity set-up)

DV Complementarity set-up

Source Type III sum of squares

df Mean square

F Sig. Partial eta squared

Time Structured Assignment Sphericity Assumed 1038.242 2 519.121 5.570 0.005 0.115

Greenhouse-Geisser 1038.242 1.961 529.406 5.570 0.006 0.115

Huynh-Feldt 1038.242 2.000 519.121 5.570 0.005 0.115

Lower-bound 1038.242 1.000 1038.242 5.570 0.023 0.115

Error (assignment) Sphericity Assumed 8015.091 86 93.199

Greenhouse-Geisser 8015.091 84.329 95.045

Huynh-Feldt 8015.091 86.000 93.199

Lower-bound 8015.091 43.000 186.397

Flexible Assignment Sphericity Assumed 7218.924 2 3609.462 29.442 0.000 0.406

Greenhouse-Geisser 7218.924 1.776 4063.934 29.442 0.000 0.406

Huynh-Feldt 7218.924 1.847 3907.494 29.442 0.000 0.406

Lower-bound 7218.924 1.000 7218.924 29.442 0.000 0.406

Error (assignment) Sphericity Assumed 10543.076 86 122.594

Greenhouse-Geisser 10543.076 76.383 138.030

Huynh-Feldt 10543.076 79.441 132.716

Lower-bound 10543.076 43.000 245.188

186

Table 5.19: Continued from previous page

DV Complementarity set-up

Source Type III sum of squares

df Mean square

F Sig. Partial eta squared

Quality Structured Assignment Sphericity Assumed 348.864 2 174.432 0.602 0.550 0.014

Greenhouse-Geisser 348.864 1.813 192.456 0.602 0.534 0.014

Huynh-Feldt 348.864 1.888 184.787 0.602 0.541 0.014

Lower-bound 348.864 1.000 348.864 0.602 0.442 0.014

Error (assignment) Sphericity Assumed 24917.803 86 289.742

Greenhouse-Geisser 24917.803 77.946 319.681

Huynh-Feldt 24917.803 81.181 306.942

Lower-bound 24917.803 43.000 579.484

Flexible Assignment Sphericity Assumed 42.424 2 21.212 0.061 0.941 0.001

Greenhouse-Geisser 42.424 1.783 23.797 0.061 0.924 0.001

Huynh-Feldt 42.424 1.855 22.876 0.061 0.930 0.001

Lower-bound 42.424 1.000 42.424 0.061 0.806 0.001

Error (assignment) Sphericity Assumed 29907.576 86 347.763

Greenhouse-Geisser 29907.576 76.657 390.146

Huynh-Feldt 29907.576 79.746 375.034

Lower-bound 29907.576 43.000 695.525

187

The results from a one-way repeated measures ANOVA demonstrate that there is a significant difference in time scores across the assignments for “sta-ble” F (2, 86) =5.570 p ˂ 0.05 and “dynamic” F (2, 86) =29.442 p ˂ 0.05 complementarity set-ups. For quality scores, no significant differences were established. Although data analysis from a repeated measures ANOVA does provide important results for complementarities’ effect on individual IT-ena-bled productivity, the next sub-section shows the effect size and power for the time and quality scores that helped us interpret the obtained results.

5.2.4 Effect size, power and final results Effect size and achieved power have been calculated by using G*Power (http://www.gpower.hhu.de) software (version 3.1) (Faul et al., 2009). The re-sults for time and quality scores are presented in Table 5.20. Table 5.20: Effect size and power for time and quality scores Effect size Power

Time

Between group 0.14 (small) 0.800 (strong)

Within group 0.61 (large) 1.000 (perfect)

Interaction 0.27 (medium) 0.999 (strong)

Quality

Between group 0.37 (medium) 1.000 (perfect)

Within group 0.10 (small) 0.450 (medium)

Interaction 0.03 (small) 0.850 (weak)

Based on the effect size and achieved power we can summarize the

research findings from Study 2 in the following way. A repeated measures 2 x 3 ANOVA for time scores indicated insignificant between-group difference for complementarity set-up F (1, 86) = 1.632, p = 0.205, d = 0.14. There was a significant within-group difference for the assignment performed F (2, 172) = 0.492, p = 0.000, d = 0.61. Bonferroni post hoc tests indicated that individuals spent significantly more time on the second assignment than on the first assignment (p < 0.001) and on the third assignment than on the first assignment (p < 0.001). Time to perform the third assignment was

188

significantly lower than the time spent on the second assignment (p < 0.001), yet significantly higher than the time spent on the first assignment (p < 0.001). There was a significant interaction between complementarity set-up and assignment F (2, 172) = 6.462, p = 0.020, d = 0.27. Further examination of the sources of interaction (using t-tests and repeated measures one-way ANOVAs) suggests that an interaction occurred because the main effect of the assignments was apparent for both “stable” and “dynamic” complementarity set-ups.

A repeated measures 2 x 3 ANOVA for quality scores indicated a signifi-cant between-group difference for complementarity set-up F (1, 86) = 11.970, p = 0.001, d = 0.37. Innovators involved in a “stable” complementarity set-up demonstrated a higher quality than innovators involved in a “dynamic” com-plementarity set-up. There was insignificant within-group difference for the assignment performed F (2, 172) = 31.802, p = 0.594, d = 0.10. There was also an insignificant interaction between complementarity set-ups and assignments performed F (2, 172) = 0.122, p = 0.867, d = 0.03. Additional t-test analysis demonstrates that there was a significant difference in t scores in respect to the first t (86) = 2.825 (p = 0.006) and second assignments t (86) = 3.162 (p = 0.002), but there was no significant difference in t scores with respect to the third assignment (p =0.032). Additional repeated measures one-way ANOVA demonstrates no significant difference in equality scores across the assign-ments for a “stable” F (2, 86) = 0.602 p = 0.55 and “dynamic” F (2, 86) = 0.061 p = 0.94 complementarity set-up. A minimal increase in effect size and power for within group and interaction effect could be brought about if the treatment effect was tested over a longer period of time (Stevens, 2012), since a-priori sample size was appropriate and the test conducted with a more lenient α = 0.10 demonstrated the same results.

5.3 Summary of the research findings By conducting both studies, we expected that individuals with adaptive cognitive style involved in a “stable” complementarity set-up and individuals with innovative cognitive style involved in a “dynamic” complementarity set-up would generate higher productivity compared to other configurations of complementary factors. To test our hypotheses, we conducted two empirical studies, a longitudinal quasi-randomized field experiment and an online experiment. A summary of the results obtained is presented below.

In Study 1, we expected that sales representatives would be more produc-tive when their cognitive styles are matched with appropriate complementarity set-ups in comparison with other configurations of these factors. A summary of the expected versus obtained results is presented in Table 5.21 below.

189

Table 5.21: Summary of expected versus obtained results: Study 1

Variable DID estimator

Expected results Obtained results

Number of sales calls

-0.369*** Design 1 £ Design2 Design 1 > Design 2 -0.393*** Design 1 £ Design3 Design 1 > Design 3 0.649*** Design 1 < Design 4 Design 1 < Design 4

Number of products sold

-17.367 Design 1 £ Design 2 Design 1 > Design 2 -94.181** Design 1 £ Design 3 Design 1 > Design 3 87.235** Design 1 < Design 4 Design 1 < Design 4

Note: * significant at 10%; ** significant at 5%; *** significant at 1% confidence level;

We expected that sales representatives involved in Design 4 with a “full” set of complementarities would generate greater productivity than sales repre-sentatives involved in Design 1 without operational change, Design 2 with structured partial change, and Design 3 with semi-structured partial change. Consistent with this expectation, the obtained results show a positive and sta-tistically significant effect of complementarities on the number of sales calls. Overall, the number of sales calls has increased in Design 4 by 31% after a “full” set of complementarities was implemented. Moreover, our findings demonstrate that the number of sales calls has decreased significantly in De-sign 2 (by 23%) and Design 3 (by 29%) compared to Design 1 as a control group. Thus, our expectations were supported.

In order to support the obtained results, we conducted more detailed DID analysis between pairs of designs. This analysis confirmed previously ob-tained results. For example, the results demonstrated that sales representatives involved in Design 4 were significantly more productive after the introduction of operational changes in comparison to sales representatives involved in De-sign 1, 2 and 3. The robustness check and analysis of the intervention effect persistence confirmed the obtained results.

In general, the number of products sold has increased by 30% after the operational changes have been implemented in Design 4. These results give support to the research model and formulated hypotheses. Moreover, our findings demonstrate that the number of products sold has decreased in Design 3 (by 1%).

To summarize, the obtained results demonstrate that productivity increase related to the number of sales calls occurred in Design 4 where the “full” set of complementarities was introduced. Our results also show that when the more aligned IT system is used without complementarities the opposite (negative) effect can occur. Moreover, the results show that limited or wrongly assumed complementary factors may negatively affect individual IT-enabled productivity. Therefore, our expectations about the impact of a “full” complementarity set-up were supported in Study 1.

190

In Study 2, we tested the research model partially since we collected data largely for software programmers with innovative cognitive style. To summa-rize the results, the assignment completion times and quality of the developed application for innovators involved in ”stable” and ”dynamic” complementa-rity set-ups were statistically different before the introduction of a more aligned IT system. For example, innovators involved in a “dynamic” comple-mentarity set-up spent, on average, much less time performing the first assign-ment. However, on average, the quality was higher for applications developed by innovators involved in a ”stable” complementarity set-up. The manner in which both groups of participants learned a more aligned IT system was also quite different. Innovators involved in a “dynamic” complementarity set-up had a greater change in completion time when learning to use the more aligned IT system first. However, learning pattern was lower than that of innovators involved in a ”stable” complementarity set-up. In Table 5.22, we provide a summary of expected versus obtained results. Table 5.22: Summary of expected versus obtained results: Study 2

Session/Assignment Scores in complementarity set-ups

p value

Exp

ecte

d re

sults

Obt

aine

d re

sults

Stable (S)

Dynamic (D)

Time, min. 1. Less aligned IT system 42 33 0.009 S>D S>D 2. More aligned IT system 49 51 0.454 S>D S<D 3. Learning effect 47 45 0.567 S>D S>D

Quality, % 1. Less aligned IT system 77 63 0.006 S<D S>D 2. More aligned IT system 75 62 0.002 S<D S>D 3. Learning effect 73 61 0.032 S<D S>D

Quality per time (% / min.) 1. Less aligned IT system 1.83 1.91 0.094 S<D S<D 2. More aligned IT system 1.53 1.22 0.026 S<D S>D 3. Learning effect 1.55 1.36 0.156 S<D S>D

We hypothesized that individuals with innovative cognitive style will gen-

erate significantly greater productivity when matched with a “dynamic” com-plementarity set-up. Since the participants were learning a more aligned IT system, it was expected that innovators in both “stable” and “dynamic” com-plementarity set-ups would initially take longer to complete the assignment, but that assignment completion time would decrease over time, especially for

191

innovators involved in a “dynamic” complementarity set-up. It was also ex-pected that since innovators dislike structure and prefer changes, assignment time completion of innovators involved in a “dynamic” complementarity set-up will be significantly lower when the participants first learn the use of a more aligned IT system.

The results from Study 2 demonstrate that, when completing the first assignment with a less aligned IT system, time scores were significantly different for innovators who were later involved in “stable” and “dynamic” complementarity set-ups (42 versus 33 minutes). Quality scores were significantly different for innovators involved in “stable” and “dynamic” complementarity set-ups (77 versus 63%). Yet, quality per time was statistically insignificant for both groups of participants (1.83 versus 1.91 % / min.). Therefore, both groups performed well during the period preceding the more aligned IT system introduction.

As we expected when completing the second assignment with a more aligned IT system, time scores increased for both groups. However, average assignment completion time increased by 7 minutes (16%) for innovators in-volved in a “stable” complementarity set-up and by 18 (54%) minutes for in-novators involved in a “dynamic” complementarity set-up compared to base-line. Yet, the difference between time scores for both groups of participants became insignificant. Quality scores remained similar to the first assignment and the difference between these scores was statistically significant (75 versus 62% respectively). Quality per time unit became significantly higher for inno-vators involved in a “stable” complementarity set-up rather than in a ”dy-namic” complementarity set-up (1.53 versus 1.22 % / min. respectively).

Since the participants were mastering a more aligned IT system, it was ex-pected that cognitive style matched with the appropriate complementarity set-up might affect productivity and that innovators matched with “dynamic” complementarity set-up might perform the third assignment more quickly. The results demonstrate that in comparison to the second assignment, time scores decreased for the participants involved in both complementarity set-ups. However, average assignment completion time increased by 2 minutes (4%) for innovators involved in a “stable” complementarity set-up and by 6 minutes (12%) for innovators involved in a “dynamic” complementarity set-up compared to the second assignment. Quality scores did not change signifi-cantly in comparison to the second assignment (73 versus 61% respectively). The difference between quality per time became statistically insignificant for both groups of participants.

To summarize the results, our empirical investigations in Study 1 demon-strate support for the formulated configurations of complementary factors and provide some new insights of their impact on individual IT-enabled produc-tivity. In principle, our results provide strong support for the systems approach of the complementarity theory. The results show that productivity of infor-mation workers can hardly be increased only by the use of a new more aligned

192

IT system. The results also demonstrate that limited and wrongly assumed complementarities may negatively affect individual productivity. The ob-tained findings also support the time lag hypothesis for productivity growth after the deployment of an IT system. The results obtained in Study 2 do not offer support for our major expectation due to technical limitations such as completeness (we only obtained data for innovators and not for adaptors) and a limited number of runs before and after a more aligned IT system was intro-duced. Yet, this study provides us with a number of suggestions that will be discussed in the next chapter for improvements of the experimental set-up to study complementarities and their impact on IT-enabled productivity which is also one of the contributions of this research.

193

6. Discussion

In this chapter, we discuss the obtained findings, the value of the research model, theoretical and practical implications of this research as well as some inherent limitations and avenues for future research. In section 6.1, we discuss the findings obtained in two empirical studies and in section 6.2, we discuss the value of the formulated theoretical model as such and how it advances current knowledge. In sections 6.3 and 6.4, we present theoretical contribu-tions and managerial implications of this research. This chapter concludes with a discussion of limitations that present several avenues for future re-search.

6.1 Discussion of empirical results In this section, we discuss the obtained findings from the two empirical studies that were conducted to test the research model. In general, our findings pro-vide not only interesting insights into complementarities of productive IT use but also help us develop suggestions for future research design that can be further used to study the impact of complementary factors on individual IT-enabled productivity with higher accuracy. The obtained findings are reported first separately and then together to integrate these findings into the results of previous studies.

6.1.1 Discussion of empirical results: Study 1 The first study was designed to test the formulated research model of complementary factors and identify their impact on sales representatives’ productivity in a longitudinal perspective when a new, more aligned IT system was used. Consistent with our expectation, the obtained results show a positive and statistically significant effect of complementarities on individual IT-enabled productivity of sales representatives (Design 4). Particularly, the obtained results indicated that productivity of sales representatives that were

194

involved in the design with a “full” complementarity set-up has increased significantly after the implementation of changes when compared to designs with no or only partial complementarity set-ups. These results showed support for the research model, that is complementarities introduced together with a new, more aligned IT system positively affected individual productivity of sales representatives. The findings were robust to pre-treatment time placebo and persistent to post-treatment time placebo analysis. Several interesting insights can be derived from these findings that can help address the relationships between complementary factors and individual IT-enabled productivity of sales representatives in particular and information workers in general.

First, this study successfully identified a configuration of complementary factors for individual IT-enabled productivity, a new contribution that adds to previous complementarity studies that focus on individual level (Athey & Stern, 2002; Autor et al., 2003) and studies that focus on the identification of complementarity patterns (Brynjolfsson & Milgrom, 2013). The complex configuration of complementary factors represents the most significant contribution of this study because this configuration constitutes a unique synthesis of previous studies and offers a more comprehensive account of individual productivity factors. Moreover, these findings add a unique configuration of complementary factors based on the systems approach of the complementarity theory (Ennen & Richter, 2010). To our knowledge, this is the first study to investigate and find support for the impact of a system of complementary factors on individual IT-enabled productivity.

Second, we found that productivity of individuals who used a new IT system without complementarities (Design 2) has decreased, meaning that it is not sufficient to merely use a new, more aligned IT system and expect direct effects on individual productivity. These findings support the results reported in previous studies on information worker productivity as well as sales representative’s productivity that IT use is necessary, yet not sufficient alone to increase IT-enabled productivity (Jain & Kanungo, 2005, 2013; Becker et al., 2009). These studies reported that besides IT system use, other factors, including people, operations, and managerial support can play no less an important role in productivity increase. Another plausible explanation for these results is the existence of a time lag for the emergence of IT benefits (Brynjolfsson, 1993). Indeed, previous studies investigating the relationships between IT use and productivity found that there is a time lag between the introduction of a technology and its impact on productivity (Hu & Plant, 2001; Devaraj & Kohli, 2003). For example, some studies found that there is a significant time lag (from five to six years) in terms of productivity gains from new IT system adoption (Hu & Plant, 2001). Other studies report a one to two-year time lag before a new IT system’s effects on productivity become considerable (Dong & Zhu, 2006). Therefore, future data may help to

195

encounter time lag that is necessary before productivity can increase after the use of a new IT system.

Third, we found that productivity of individuals involved in a partial complementarity set-up (the introduction of a more aligned IT system together with non-mandatory obligation to follow each step in the operational process) has decreased significantly (Design 3). One possible reason for these results can be explained by difficulties in showing the complementary impact when a limited number of complementarities is used (Ennen & Richter, 2010). For example, Athey and Stern (2002) could not demonstrate the complementary relationship between the use of a new IT system and the operational process. Therefore, our results also demonstrate difficulties in demonstrating complementarity impact by applying the interaction approach (the analysis of the interaction between a limited number of identified factors) of the complementarity theory, yet further confirmation of the obtained results is needed. Another plausible reason for these findings is in line with previous studies (Roberts, 2007; Poon et al., 2009) that demonstrated that some configurations of factors may generate positive performance while others negative performance. This implies that on the one hand, properly understood and matched complementary configurations of factors may generate an important productivity increase. On the other hand, wrongly assumed configurations of factors may also generate negative productivity impact. Thus, knowledge of these factors and their match is extremely important when the undesirable impact on individual productivity is to be avoided.

Fourth, somewhat contrary to our expectations that productivity would increase gradually through Design 2 (structured partial change), Design 3 (semi-structured partial change), and Design 4 (full change) when compared with Design 1 (no change); our findings demonstrate that productivity has decreased significantly in Design 2 and 3. These results propose that managers should neither underestimate the influence of complementary factors on individual IT-enabled productivity, nor overestimate the use of new, more aligned IT systems alone. Hence, the obtained knowledge can be used to advance the new IT system use – not only with regard to individual productivity but also considering corresponding complementarities. This is important because, as the obtained findings demonstrate, IT-enabled productivity differs substantially with the presence of the “full” set of complementarities.

Fifth, these findings provide empirical confirmation of previously formulated more general propositions that a person, technological tool, task and contextual settings together influence individual productivity (Mason & Mitroff, 1973; Yaverbaum, 1988; Kraemer & Danziger, 1990). In the context of sales operations, the findings also confirm previous studies that proposed to study individual productivity as a function of an individual, IT tool, task/process and managerial settings (Mullins, 2007; Hopp et al., 2009). Particularly, this study highlights the importance of a match between

196

adaptive/innovative cognitive style, the structural complexity of the work process, specific training and education activities, incentives’ mode, and decision-making structure for individual IT-enabled productivity in a situation where a new, more aligned IT system is used. Yet, further research can uncover new combinations of complementary factors that also can provide interesting insights into individual IT-enabled productivity of information workers.

Taken together, the findings obtained from the first study provide support to the notion that the productivity of an information worker who uses an IT system is positively influenced by complementary factors versus an approach where an IT system is used without any synchronized complementary factors. The study also provides empirical support to the notion that assembling a set of a limited and wrongly assumed factors is not enough, but the right factors need to be synchronized in the right manner to increase IT-enabled productivity.

6.1.2 Discussion of empirical results: Study 2 The second study was designed to test the formulated research model in experimental settings in another information-intensive environment – software construction. Particularly, this study was conducted to validate the formulated research model and identify cross-study complementary patterns of productive IT use. In the online experiment of software programmer productivity, we made an attempt to compare how adaptors and innovators perform in “stable” (structured operating process, push mode training in work technology, exogenous incentives, and centralized decision-making) and “dynamic” (flexible operating process, a combination of minor upfront training with optional on-demand training in work technology, endogenous incentives, and decentralized decision-making) complementarity set-ups. Yet, we were only able to collect data for innovators, and therefore, only partially test the formulated research model. Therefore, the completeness of the experiment was disrupted, which is a key limitation to this study. Hence, we were unable to compare whether adaptors perform better or worse in any of the two contexts (“stable” versus “dynamic”) than innovators. In general, the results of this study demonstrate that innovators involved in different complementarity set-ups (“stable” versus “dynamic”) appear to have quite different results from what we expected and thus help us develop suggestions for improvements of the experimental set-up for future research. We discuss these results and suggestions in more detail below.

First, we found that the time needed to perform the first assignment with an old, less aligned IT system was significantly different for innovators involved in different complementarity set-ups. We found that innovators spent

197

much less time when they were involved in a “dynamic” rather than a “stable” complementarity set-up, yet with lower quality. On the one hand, these results are somewhat counter-intuitive since a “stable” context is usually more productive when compared with a “dynamic” one for all individuals as a “dynamic” context by its definition imposes more time (Rigby et al., 2016). Moreover, innovators in task performance try many completion paths (Alter, 2001) which also takes time. On the other hand, significantly lower quality of the developed application for innovators involved in a “dynamic” complementarity set-up could explain this difference in time scores. It could also be that in a “dynamic” set-up the operational process was designed in a way that provided fewer guidelines for programmers, and therefore shortened the time needed to perform the assignment, yet generated poorer quality. The difference in quality could also be caused by an innovators’ tendency to avoid details and ignore the rules (Kirton et al., 1991). For example, in a “stable” complementarity set-up participants were explicitly informed of application requirements, yet in a “dynamic” complementarity set-up, innovators were given additional requirements and details about the application they were developing during the work process that could be ignored, and thus lowered quality. Taking these results together, we calculated the quality of the developed application divided by time unit. The results were statistically insignificant for the first assignment, meaning no difference in this indicator between innovators involved in different complementarity set-ups. This implies that future experiments have to be designed with multiple runs when an old, less aligned IT system is used in order to obtain saturation16 and establish a clear benchmark.

Second, we found that time to perform the second assignment with a new, more aligned IT system significantly increased for innovators involved in both complementarity set-ups. This noticeable difference in time scores between the first and the second assignments is consistent with the results obtained in previous studies that innovators spend quite a considerable amount of time when they master a new IT system (McLeod et al., 2008). However, on average, innovators involved in a “stable” complementarity set-up spent less time on performing the second assignment than innovators involved in a “dynamic” complementarity set-up (7 min. versus 18 min.). These results indeed support the current wisdom that “dynamic” contexts are more time consuming (Rigby et al., 2016). For example, in a “stable” complementarity set-up, the operational process did not require that participants form their own assumptions and judgment to perform assignments (Abdolmohammadi & Wright, 1987). In comparison, in a “dynamic” complementarity set-up, alternative solutions require considerable judgment from participants that can increase time scores significantly (ibid.). This time difference can also be explained by the differences in learning strategies for “stable” and “dynamic” 16 Saturation is meant here as the state when no changes of pattern occur.

198

complementarity set-ups. For example, in a “stable” complementarity set-up participants could learn about a new IT system prior to performing the second assignment. In a “dynamic” complementarity set-up, participants most probably explored a new IT system while performing the assignment which also required additional time. Therefore, these results imply that individual productivity of innovators mastering a new, more aligned IT system depends on a complementarity set-up in which they are involved. Quality scores did not change significantly in the second assignment and were statistically different between both groups of participants. Due to significant time changes, quality per time unit became significantly different. For example, quality per time unit was significantly higher for innovators involved in a “stable” complementarity set-up. Taken together, these results also support the obtained results from previous studies on the time lag required to observe productivity increase from the use of new IT systems (Hu & Plant, 2001; Devaraj & Kohli, 2003).

Third, the results also demonstrate that there are differences in time scores between different complementarity set-ups in the third assignment compared to the second assignment. On average, the time to perform the third assignment in comparison to the second assignment was less for innovators involved in a “dynamic” complementarity set-up than for innovators involved in a “stable” complementarity set-up (3 min. versus 6 min.). These results confirm the existing wisdom of learning effect from repeated work functions (Womer, 1984), new IT system use (Waldman et al., 2003; McLeod et al., 2008) and its effect on productivity. Yet, time score differences were still statistically insignificant between both groups of participants. Moreover, time score differences in the third assignment could occur because on average innovators involved in a “dynamic” complementarity set-up demonstrated better time scores than innovators involved in a “stable” complementarity set-up during the first assignment. Quality score in the third assignment remained statistically different between groups, yet did not demonstrate significant changes compared to the first and second assignment. Quality per time unit became statistically insignificant between both groups of participants. All this implies that more runs of the experiment are needed to achieve saturation in productivity scores after a new, more aligned IT system is stabilized.

Fourth, in general, from the experimental results it is noticeable that while in the first and third assignments innovators involved in a “dynamic” complementarity set-up worked faster, yet with significantly lower quality than innovators involved in a “stable” complementarity set-up. One possible explanation for these results can be that a dynamic context offers more positive cognitive stimulation and therefore creates appropriate conditions for innovators to work faster, yet not necessarily with better quality. This implies that both performance metrics (time and quality) have to be closely monitored in future research to understand the impact of complementarity set-ups. Based on these metrics, the optimal trade-off between time and quality can further

199

be established. This may help managers to implement corresponding complementarities with a desired level of outcome.

Fifth, the results of the online experiment could also be affected by operationalization of the main constructs and control settings. For example, in the online experiment we made an attempt to operationalize the main constructs in close relation to real work settings. Yet, constructs such as training and education, incentives and decision-making structure were operationalized so as to satisfy the requirements of the online experiment. In reality, these constructs could be more complex and diverse. In addition, we could not fully control how IT systems were designed and how participants learnt a new IT system. For example, it has previously been established that IT systems can be created to be more appropriate for the preferences of one or another cognitive style, and hence positively affect their performance (Chen & Popovich, 2003; Coskun & Grabowski, 2005). It can also be so that IT system design could affect learning patterns in relation to cognitive style (Chilton et al., 2005). Therefore, future experiments can be improved in terms of reflecting actual real world practices together with a more detailed consideration of the complementarities between cognitive style and IT system characteristics.

Sixth, we have already mentioned that we were able to collect data only for individuals with innovative cognitive style. One interesting explanation for this finding might be that an online environment (internet-based jobs) as a working environment mostly attracts individuals with innovative cognitive style, rather than individuals with adaptive cognitive style. For example, internet-based jobs are characterized as temporary and rapidly changing (Sadler et al., 2009) which is more suitable for innovative individuals. This fit between cognitive style and working environment seems to match findings from previous studies (Kirton, 2003; Chilton et al., 2005), suggesting that a rapidly changing environment requires individuals with innovative cognitive style. Yet, more research is needed since an online work environment (Sadler et al., 2009) and its fit with a particular cognitive style is little studied in current literature, and future research has to take these findings into consideration before shifting from laboratory to online experimental settings.

Therefore, by conducting this experiment, we were able to show empirically that when individuals with the same type of cognitive style operate in two different contexts (complementarity set-ups), this gives rise to two different performance outcomes. We were also able to demonstrate that learning effect from repeated work functions and new IT system use (Womer, 1984; Waldman et al., 2003; McLeod et al., 2008) is relevant for IT-enabled productivity. Yet, by addressing technical and operational limitations of the conducted online experiment to study complementarities of productive IT use, future experiments can be created to overcome these limitations in the following way. Since more runs are required to establish saturation with both old, less aligned and new, more aligned IT systems, time series design of the

200

experiment would be appropriate. This design would allow the assessment of productivity prior to and after the introduction of a new IT system and identify the existence of complementarity effects on IT-enabled productivity within a temporal sequence of events (McLeod et al., 2008). The online experiment demonstrated that collection of data in these settings can be questioned in relation to a particular cognitive style. Therefore, field or laboratory experiments could overcome this uncertainty. A field experiment would overcome limitations of the main complementarities’ operationalization. Yet, a laboratory experiment would help to address control over main complementarities of productive IT use. Hence, both strategies can be taken into account when designing further experimental research of complementarities and its impact on IT-enabled productivity.

6.1.3 Discussion of empirical results: Summary

Taken together, the obtained findings from both studies demonstrate that com-plementarities in the context of an individual information worker IT-enabled productivity matter (Roberts, 2007; Brynjolfsson & Milgrom, 2013). These findings extend the complementarity theory (Milgrom & Robers, 1990, 1995) by applying the systems approach (Ennen & Richter, 2010) and revealing two complementarity set-ups centered at individual cognitive style: ”stable” (com-plementarity set-up which includes adaptive cognitive style, structured oper-ating process, upfront comprehensive mandatory training in work technology, exogenous incentives, and centralized decision-making) and ”dynamic” (in-novative cognitive style, flexible operating process, a combination of minor upfront mandatory training with optional on-demand training in work tech-nology, endogenous incentives, and decentralized decision-making). These findings point to complementary set-ups that are necessary and sufficient con-ditions associated with increased individual information worker IT-enabled productivity.

Our findings also suggest that in order to study complementarities of productive IT use in experimental settings, learning effect has to be taken into account (Waldman et al., 2003; McLeod et al., 2008) and multiple runs are needed before and after a new, more aligned IT system is introduced and used. This implies that experiments on the impact of complementarities on IT-enabled productivity require more complex research design, and studies on before-, after- and follow-up strategy alone may not lead to consistent productivity results. Therefore, further experimental research with additional sessions is needed to provide us with more detailed statistical analysis of IT-enabled productivity. Moreover, both field and laboratory experiments can help overcome limitations of an online experiment when studying complementarities and their impact on IT-enabled productivity.

201

6.2 Value of the research model In this research, we formulated and tested a new research model of the con-figurations of complementary factors, considered at the individual level in the context of an information-intensive environment, when a more aligned IT sys-tem is used. Specifically, we applied the systems approach of the complemen-tarity theory when formulating the model. The choice of the theoretical ap-proach was based on a practical organizational problem which faces managers when they decide to purchase a new, more functional IT system to increase individual productivity. Namely, what complementary factors should be con-sidered in advance to ensure that using a new, more aligned IT system will increase individual productivity.

The proposed model manifests several contributions which are listed here. First, in the model development, we made an attempt to advance long-time efforts to conceive information worker productivity. Current literature shows that we have only a partial understanding of the factors affecting individual IT-enabled productivity in post adoption context and that it is difficult to generalize the majority of studies. Moreover, the reviewed studies mainly assume a non-contingency notion that one model fits all situations. However, the complementarity theory as a theory of the modern firm is a further extension of contingency thinking (Roberts, 2007) states that there exist a number of complementary factors that, when correctly matched, can increase performance outcome. Therefore, in the model development key challenges identified in the literature were addressed by means of an integration of the various productivity factors identified independently of each other into a system of complementary factors with binary value ranges, respectively, determining information worker productivity.

Second, the model also contributes to the recent and growing literature on complementarities and their fit, particularly as there are very few conceptions of the individual level and none that adopt a systems approach (Brynjolfsson & Milgrom, 2013). For example, over the last decades, only a limited number of studies have focused on the identification of complementarities that affect individual IT-enabled productivity (Athey & Stern, 2002; Autor et al., 2003). The study by Athey and Stern (2002) found it difficult to demonstrate empirical support for proposed complementary factors, including IT use and operational process. Another study by Autor et al. (2003) demonstrated that in contrast to routine work, work technologies are complementary for non-routine, yet do not show the complementarity effect on productivity outcomes. These difficulties were explained by the existence of a limited number of complementary factors for a synergy fit to emerge (Ennen & Richter, 2010). Therefore, to overcome limitations from previous studies we choose to apply the systems approach and hereby contribute to the literature on the complementarities of productive IT use.

202

Third, in our research model construction, we followed proposals from management science disciplines and organization economics which recommend considering information worker productivity as a function of an individual, task/process, technological tool and organizational settings (Mason & Mitroff, 1973; Sonnentag, 2003; Hopp et al., 2009). In order to develop our model, we combined sources from different disciplines. Particularly, Kirton’s adaption-innovation theory (Kirton, 1976), literature on the operational process structure (Keen & Scott-Morton, 1978; MacCormack et al., 2001; Weber & Wild, 2005) and human resource management sources (Amabile, 1996; Hayes & Allinson, 1997; Ee, 1998; Ryan & Deci, 2000; Baer et al., 2003; Skinner & Drake, 2003; Ahearne et al., 2005; Sense, 2007) formed the basis of the research model. As the formulated research model draws on factors developed in various disciplinary quarters, such as psychology, managerial economics, organization studies, operations management, and information systems, this model contributes to each of the sourced theoretical bodies with regard to how each adopted factor may be complemented with a unique set of other factors.

Fourth, previous studies indicate that the impact of IT use on productivity can be measured more accurately by examining its contribution based on individual IT systems and applications at individual/process level (Mukhopadhyay et al., 1997; McCullough et al., 2013). For example, previously studied aggregate measures of IT use made it difficult to interpret the results and draw conclusions (Athey & Stern, 2002). To overcome this limitation, we investigated two specific process-enabling IT systems that were designed to support the individual productivity of employees, specifically sales practices in the first empirical study and code completion in the second empirical study. Our research, thus, contributes to the IT use impact literature by looking at the impact of specific IT systems in two different information-intensive environments, not yet sufficiently examined in the existing organizational economics and information systems literature. The proposed strategy of studying specific applications of IT can further be applied to advance our knowledge of complementarities in different information-intensive occupations. At the same time, we considered a specific situation where a more aligned IT system was used by information workers, and therefore, increased consistency and comparability between the two conducted studies, and also increased meaningful research replication. Therefore, the research model can be further tested in different information-intensive occupations and contribute to the understanding of complementary factors and IT-enabled productivity of information workers.

Fifth, in contrast to previous studies on individual information worker productivity (Czerwinski et al., 2004; Jain & Kanungo, 2005; Aral et al., 2006; Sundaram et al., 2007; Aral & Van Alstyne, 2011; Wu et al., 2009; Webster, 2012) and complementary factors (Athey & Stern, 2002; Autor et al., 2003), our research model contributes to the understanding of information worker

203

productivity by “humanizing” an individual in terms of his or her individual cognitive style (adaptive versus innovative). To date, some studies have investigated the relationships between individual cognitive style, task, features of decision support systems and individual performance (Dickson et al., 1977; Benbasat & Taylor, 1978; Benbasat & Dexter, 1982; Blaylock & Rees, 1984). Yet, these studies demonstrated that although there is a potential for the joint impact of individual attributes, information system characteristics and task design on individual performance, the results were inconclusive due to the lack of underlying theories. In our research model formulation, we linked individual cognitive style with other complementary factors based on the complementarity theory and the proposed fit between the factors, and, hereby contribute to the studies listed above.

Sixth, previous studies on technology use have been criticized for a narrow artifact-actor relational approach (Faraj & Azad, 2012; McCullough et al., 2013). These studies recommended exploring the potential of technology in a given individual and organizational context. By including different contextual factors that previously demonstrated double or triple interrelationships, such as the operational process, training and education, incentives and decision-making structure and matching them to particular cognitive style, the research model provides a novel and more comprehensive conception of the productivity of an information worker.

Therefore, the research model formulated in this research can be used to further our knowledge on complementarities and their impact on individual information worker productivity. This model is novel as it enriches the literature and practice with new configurations of complementarities (factors that are linked together and not from each factor being new as such) that increase IT-enabled productivity of information workers. While a number of factors have been explored to enrich the research model, such as adaptive/innovative cognitive style, operational process, training and education, incentives and decision-making structure; the systems approach can further be used to uncover complementary factors that might increase individual IT-enabled productivity of information workers. Below, we discuss implications of the obtained results for both theory and practice. We also acknowledge the limitations of the conducted studies and propose avenues for future research into complementarities of productive IT use at the individual level.

6.3 Theoretical contributions Previously, it was widely discussed that theories applied to the study of the relationships between technology, productivity, and organization were highly

204

deterministic in their nature, giving technology the central role (Orlikowski & Robey, 1991; Walsham, 1993). Further contradictory empirical findings across studies, for example, contradictory results from the same IT use in a single organization, different results from the use of identical IT in comparable settings, and expected consequences of IT use did not occur, led to the emergence of theories with the opposite logic, placing technology in a social context and treating it as one of the elements in complex social processes (Robey & Boudreau, 1999). Although deterministic logic is still used in current investigations, it is highly criticized for omitting essential, non-technical aspects that are integral parts of the system and that together can contribute to IT-enabled productivity and profitability (Alter, 2013). In this research, we support this kind of systems thinking and recognize that an understanding of the conditions of an information worker’s productivity must include a number of specific and diverse factors that interact in a specific manner, thus giving rise to a system of complementarities that determine the output of his or her work. Therefore, the results from the two studies conducted here provide important implications for the theory by addressing several important issues (Table 6.1).

First, this research contributes to the literature that calls for the need to investigate the underlying factors of information worker productivity at individual level (e.g. Drucker, 1999; Lord, 2002; Kraiger, 2003; Morgeson & Humphrey, 2006; Hopp et al., 2009; Cequea et al., 2011; Singh & Mohanty, 2012). Particularly, these studies reported a growing need for information worker productivity research since knowledge of the underlying factors of information worker productivity is limited. Although these studies highlighted that productivity of the individual information worker is a function of such factors as an individual, task/process, IT tool and contextual settings, they do not provide an explanation for how these factors can interact with each other, so that productivity of an information worker can be increased. Therefore, we contribute to this literature by demonstrating how these factors can complement each other in order to increase individual information worker productivity.

Second, independent studies of information worker productivity also showed that individual productivity can also be affected by its individual differences, work process, motivation, training for skills, decision-making structures or degree of autonomy, and the use of IT systems (e.g. Zmud, 1979; Agarwal & Prasad, 1999; Sokoya, 2000; Kessels, 2001; Parker et al., 2001; Frey & Osterloh, 2002; Kozlowski & Bell, 2003; Kraiger, 2003; Jain & Kanungo, 2005; Aral et al., 2012a). In this research, we overcome a key challenge of this literature – partiality (focus on one or a few factors only) – and address individual information worker productivity by an integration of the various productivity factors identified independently into a system of complementarities with binary value rages.

205

Table 6.1: Major theoretical contributions

№ Literature Studies Limitations Contributions

1 Literature that addresses the need to study factors affecting individual in-formation worker productivity

Drucker (1999); Lord (2002); Kraiger (2003); Morgeson & Humphrey (2006); Hopp et al. (2009); Cequea et al. (2011); Singh & Mohanty (2012)

No explanation about how in-dividual productivity factors can interact with each other, so that productivity of an in-formation worker can be in-creased.

We contribute by demonstrating how these factors can complement each other in order to increase individual information worker productivity.

2 Independent studies that investigate factors affect-ing information worker productivity

Zmud (1979); Agarwal & Prasad (1999); Sokoya (2000); Kessels (2001); Parker et al. (2001); Kozlowski & Bell (2003); Kraiger (2003); Jain & Kanungo (2005); Aral et al. (2012b)

Partiality We overcome partiality by addressing individual information worker produc-tivity by an integration of the various productivity factors into a system of complementarities with binary value ranges.

206

Table 6.1: Continued from previous page № Literature Studies Limitations Contributions

3 Earlier studies that inves-tigated the impact of a number of factors, in-cluding individual cogni-tive style, task/process and IT use on individual productivity

Dickson et al. (1977); Benbasat & Taylor (1978); Benbasat & Dexter (1982); Blaylock & Rees (1984)

The lack of underlying theory; Inconsistent results

We demonstrate that individual productivity of information workers can be investigated based on the com-plementary approach. We also empiri-cally validate the proposed configura-tions of complementary factors.

4 Studies that applied a non-complementary ap-proach to study the im-pact of IT use on individ-ual productivity of infor-mation workers

Czerwinski et al. (2004); Kvassov (2004); Jain & Kanungo (2005); Aral et al. (2006); Sundaram et al. (2007); Aral & Van Alstyne (2011); Chung & Hossain (2009); Wu et al. (2009)

Partiality; Generalizability

To overcome partiality limitation, we studied the system of complementary factors of productive IT use. To overcome generalizability limita-tion, we developed a research model in a way to ensure its applicability in different information-intensive occu-pations where it is decided to use new, more aligned IT system with the aim to increase individual productivity.

5 Literature on organiza-tional economics that rests on the complemen-tarity theory

Athey & Stern (2002); Autor et al. (2003); Bloom et al. (2007); Tambe et al. (2012); Aral et al. (2012b); Brynjolfsson & Milgrom (2013)

Studies mostly focused on firm level rather than individ-ual level assuming that com-plementary configuration fits all firms. A limited number of factors were investigated at individual level.

We contribute by applying the com-plementarity theory at individual level. We also applied a systems approach of the complementarity theory to demon-strate the impact of a system of com-plementarities on IT-enabled produc-tivity at individual level.

207

Table 6.1: Continued from previous page № Literature Studies Limitations Contributions

6 Studies that partly ad-dressed the relationships between such factors as operational process, training and education, incentives, decision-making structure, and adaptive/innovative cog-nitive style

Kirton (1976); Amabile (1996); Hayes & Allinson (1997); Ryan & Deci (2000); Baer et al. (2003); Ahearne et al. (2005); Sense (2007)

Partiality We add complementary evidence by integrating these previously studied factors in a system of complementary factors.

7 Previous studies of com-plementarities and IT-en-abled productivity at firm level, that reported the need to decompose the firm with the aim to un-derstand its internal mechanisms in relation to IT use

Mukhopadhyay et al. (1997); Melville et al. (2004); Kohli & Grover (2008); Schryen (2013); McCullough et al. (2013)

Top-down approach We offer a contribution in investigat-ing the internal work of a firm by fol-lowing the bottom-up approach and complementing the top down ap-proach.

208

Third, we also contribute to earlier studies in information systems discipline that made attempts to investigate the impact of a number of factors, including individual cognitive style, task/process and IT use, in IT-enabled productivity (Dickson et al., 1977; Benbasat & Taylor, 1978; Benbasat & Dexter, 1982; Blaylock & Rees, 1984). While these studies already assumed that a more detailed attention has to be given to individual IT-enabled productivity factors, they were mostly criticized for their lack of underlying theory and inconsistent results. In this research, theoretical efforts are made to demonstrate that individual productivity of information workers can be investigated based on the complementary approach. By applying the complementarity theory, we demonstrate that individual IT-enabled productivity can be considered as a function of complementary factors and, if matched correctly, can increase individual productivity. This research extends knowledge of information worker productivity by identifying specific and complex configurations, out of various testable complementary factors that drive their IT-enabled productivity. We also empirically validate the proposed configurations of complementary factors that condition individual IT-enabled productivity.

Fourth, this research also complements studies that mostly applied non-complementary approach to the study into the impact of IT use on individual productivity of information workers in a post-adoption context (Czerwinski et al., 2004; Kvassov, 2004; Jain & Kanungo, 2005; Aral et al., 2006; Sundaram et al., 2007; Aral & Van Alstyne, 2011; Chung & Hossain, 2009; Wu et al., 2009). In this research, we address two key limitations of recent IT-enabled productivity studies – partiality and generalizability. On the one hand, we overcome partiality by studying the system of complementary factors of productive IT use. We identified and integrated these factors based on various productivity factors, double or triple interrelations between which were identified independently in previous studies. On the other hand, we overcome generalizability by developing a research model in such a way as to assure its applicability for different information-intensive occupations where the decision has been taken to use a new, more aligned IT system with the aim of increasing individual productivity. By studying configurations of complementary factors, we provide a richer and broader understanding of how a more aligned IT system can be used to drive individual productivity of an information worker. Our findings suggest that complementarities of productive IT use can play an important role in increasing individual IT-enabled productivity.

Fifth, the research results also contribute to the emerging literature on organizational economics that rests on the complementarity theory (Brynjolfsson & Milgrom, 2013), and where new configurations of complementarities need to be added. This research extends to studies (Bloom et al., 2007; Tambe et al., 2012; Aral et al., 2012b) that used the complementarity theory by applying it at the level of an individual. Prior work

209

on IT complementarities and individual productivity has not been precise about whether complementarities increase productivity at individual level, since only a limited number of factors have been investigated (Athey & Stern, 2002; Autor et al., 2003). To overcome this limitation and materialize IT complementarities, we applied a systems approach of the complementarity theory (Ennen & Richter, 2010) to explore the integrative effect of a system of complements when a more aligned IT system is deployed and used. Therefore, this research adds to the complementarity literature in terms of the systems approach by providing new configurations of complementarities that, if matched correctly, can increase individual IT-enabled productivity.

Sixth, and more specifically, we formulated the research model based on prior work that partly addressed the relationships between such factors as operational process, training and education, incentives, decision-making structure, and adaptive/innovative cognitive style (Kirton, 1976; Amabile, 1996; Hayes & Allinson, 1997; Ryan & Deci, 2000; Baer et al., 2003; Ahearne et al., 2005; Sense, 2007). We add complementary evidence by integrating these previously studied factors into a system of complementary factors with binary value rages which is the most significant contribution of this research. Therefore, this research has implications for studies that partially addressed complementary relationships between the aforementioned factors by providing compatibility evidence.

Seventh, previous studies on the impact of IT on productivity, particularly at firm level, reported that a decomposition of the firm is needed in order to understand its internal mechanisms in relation to IT use (Mukhopadhyay et al., 1997; Melville et al., 2004; Kohli & Grover, 2008; Schryen, 2013; McCullough et al., 2013). However, this top-down approach to understanding a complex phenomenon requires a holistic view to explain the whole before investigating any specific question. The research described here offers a contribution to investigating the internal work of a firm by following the bottom-up approach in such a way as to complement the top-down approach and thereby account for anything the latter may miss.

Eventually, research on computer-supported environments involves different disciplines such as information economics, information systems, operations management, organization economics, business administration and work psychology (Wellman et al., 1996). Following this idea, this study is based on several theoretical foundations. First, sources of information economics allow us to acquire necessary knowledge in order to understand the variety of information workers’ occupations, functions of information workers and characteristics of information work. Second, a theoretical understanding of the productivity concept requires insight into labor economics, specifically the labor production function. Third, theories from information system literature enable us to understand how IT adoption and its interaction with organization tasks affect individual productivity. Fourth, operations management literature helps us address the design of work processes. Fifth,

210

work psychology contributes with the individual’s cognitive style identification. Sixth, organizational studies and human resource management literature enable us to identify complementary organizational factors of effective IT use. Together, these fields provide us with new ideas and tools to understand how IT use can make information workers more productive. Therefore, this research has implications for operations and productions management literature, production and information economics, strategic management, organization and human resource management, literature on change management and information system research that focuses on individual productivity of information workers and the role of complementary factors of effective IT use.

6.4 Managerial implications In today’s highly competitive environment, there is an increased emphasis on individual achievement and IT-enabled productivity in a post-adoption con-text (Aral et al., 2012a). The main concern for managers in an information-intensive environment is to be not only ensured that IT systems are accepted and used, but also that their usage provides a productivity increase (Jain & Kanungo, 2005, 2013). The high rate of IT project failure to enhance produc-tivity is often explained by the managers’ inability to understand the relation-ships between technical, human and organizational factors (Brynjolfsson & Milgrom, 2013). In the current environment, indeed, practitioners require a deeper knowledge of intangible, not directly measurable resources and assets which, together with IT systems’ use, are of critical importance.

The results of the present research are important for managers who plan to introduce a new, more aligned IT system with the aim to increase productivity of information workers. In general, this research demonstrates that several factors, of various kinds, should be elaborated in a deliberately synchronized manner in order to increase individual IT-enabled productivity of information workers. As careful consideration must be given to planning and introducing new IT systems in information-intensive organizations, this research empirically demonstrates that additive value synergies arising from the complementarity of the introduction of a more aligned IT system, particular individual characteristics and management processes have significant positive effects on individual information worker productivity. This research also guides practitioners by informing them that addressing only one or more factors in an isolated manner may not have much influence on the productivity of an information worker. For increasing the individual IT-enabled productivity of information workers, the results of this research have the following practical implications for managers who decide upon IT system

211

acquisition, implementation, and use. First, in terms of practical implications, our findings suggest that when

managers decide upon complementarities of productive IT use by information workers, a systems approach would be most beneficial (Ennen & Richter, 2010). Although previous studies highlight some important factors, including people, work processes and particular human resource management practices in order to increase IT-enabled performance (Becker et al., 2009; Bloom & Van Reenen, 2011), there is to our knowledge no practical evidence to show a particular system of complementarities that can lead to individual IT-enabled productivity increase. Our results suggest that it would be worthwhile for managers to assess a set of complementary factors in order to ensure that the use of a new, more aligned IT system will lead to individual information worker productivity increase.

Second, even though there are many factors that require managers’ atten-tion when using new IT systems, the results of this research suggest that par-ticular individual and organizational factors have to be addressed in a synchro-nized manner. To begin with, managers need to clearly understand which characteristics of employees are relevant in a situation when a new, more aligned IT system is used. One of them, as proposed in this research, is the adaptive/innovative cognitive style. The cognitive style then requires careful matching with specific contextual settings of productive IT use, including the structure of the operating process, training and education mode of new tech-nology, incentive mode, and decision-making structure. The two sets of com-plementarity configurations studied in this research suggest that managers need to design appropriate configurations of complementarity factors that match the individual cognitive style adopted.

Third, we demonstrate which factors need to be synchronized with particular cognitive style when a new, more aligned IT system is introduced and used with the aim to increase the individual productivity of an information worker. Specifically, the results of this research suggest that adaptive cognitive style should be matched with structured operating process, push mode training in new technology, exogenous incentives, and centralized decision-making structure. In contrast, innovative cognitive style requires flexible operating process, a combination of pull mode training with optional on-demand training in work technology, endogenous incentives and decentralized decision-making structure to be more productive. The present research demonstrates that these complementary configurations lead to individual IT-enabled productivity increase when matched correctly. Thus, the results from this research provide managers with two configurations of complementary factors that may be used in order to increase individual IT-enabled productivity.

Fourth, our results demonstrate that the use of a new, more aligned IT system, irrespective of complementarities, or the use of a limited number of complementarities can be ineffective in increasing individual IT-enabled

212

productivity. Therefore, in general, we recommend that managers should not underestimate complementarities and avoid situations where IT systems are introduced without other factors being addressed. Moreover, specific complementary factors have to be taken into account to take advantage of a new, more aligned IT system. Managers should first be secure that such a system manifests a potential to increase productivity, by offering significant new features and also a good fit with the complementary factors at hand.

Fifth, our results also suggest that managers should be aware that IT-enabled productivity is sensitive to complementary factors when a new, more aligned IT system is used. When limited or wrongly assumed complementary factors (in our case, the introduction of a more aligned IT system together with the non-mandatory obligation to follow each step in the operational process) are introduced, this can lead to a decrease in individual productivity. Therefore, our findings also point to challenges that must be encountered by managers when considering complementary factors of productive IT use.

In general, knowing potential complementary factors enables managers to acquire a better understanding of where additional resources are needed and which factors require more attention. Managers should also consider a range of other factors that may also influence productivity in the given situation, including how all these factors may interact with each other, both positively and negatively. This would enable organizations to receive sustainable benefits from IT use. Clear understanding of complementarities ensures optimal workload, prevents undesirable workforce burnout, may assist professionals in analyzing and taking proactive measures for reducing costs and may contribute towards better control, and enable professionals to take advantages of IT use.

6.5 Limitations and avenues for future research Despite the contributions we make, this research is subject to certain inherent limitations that may be addressed in future research (Table 6.2). Table 6.2: Limitations and avenues for future research № Limitations Avenues for future research

1 Particular focus on a single infor-mation worker

To investigate group level and its com-plementarities.

2 Application of the systems ap-proach of the complementarity the-ory

To apply the interaction approach and explore the nature of particular comple-mentary factors.

213

Table 6.2: Continued from previous page № Limitations Avenues for future research

3 IT system construct was black-boxed

To open up IT system construct with re-gard to: • the degree of support, or fit, with

various kinds of work processes; • the cognitive style preferences in

representational mode of infor-mation.

4 One instrument for measuring indi-vidual cognitive style (adaptive/in-novative)

To explore other instruments for as-sessing cognitive style and formulate other complementarity set-ups.

5 The research model was formu-lated based on complementary fac-tors that were identified in previous research as the most relevant.

To uncover other complementary factors and study them in relation to individual IT-enabled productivity of information worker.

6 Study 1: The research model was tested in one empirical context of pharmaceutical sales operations.

To replicate the study across different sales companies as well as across differ-ent information-intensive occupations.

7 Study 1: Only the impact of com-plementarity set-ups on the out-come variables (sales calls and number of products sold) was stud-ied.

To analyze the differences between adaptors and innovators with regard to complementarity set-ups in learning pat-terns and times required to reach stabili-zation of IT system use.

8 Study 2: The online experiment and its design made it difficult to validate the research model.

To extend the experimental design to multiple runs before a new, more aligned IT system is used and after this system was introduced. To improve operationalization of the main constructs. To focus on longitudinal field experi-ments and randomized controlled exper-iments instead of online experiments.

9 The research model was tested in two empirical contexts.

The research model can further be tested in different information-intensive set-tings.

10 Quantitative approach to study complementarities of IT-enabled productivity was applied.

To apply a qualitative approach to un-derstand the nature of the relationships that drive complementarity effects.

11 Complementarity theory has been used as a meta-theory to combine complementary factors.

To investigate other theories that can help uncover new complementary fac-tors.

214

First, this research was focused only on a single information worker, and disregarded both the position in a network of workers and the dynamics of group work. We acknowledge that the complementarity approach can also be applied at group level to investigate how group level complementarities affect IT-enabled productivity. Yet, some studies suggest that a comprehension of the team level requires an understanding of the individual level (Ruch, 1994; Sonnentag & Volmer, 2009). Other potential avenues for future research might be an examination of complementarities and their impact on information worker productivity at group level. For example, many tasks in an information-intensive environment require collaboration (Hopp et al., 2009). Moreover, sometimes the tasks are so closely related that they are almost indistinguishable. Hence, studies at group level can provide new and important evidence of complementarities and their impact on IT-enabled productivity.

Second, this research and model formulation were based on the systems approach of the complementarity theory. Based on this approach, we developed a more elaborated understanding of the synergistic effect of a set of factors on individual IT-enabled productivity in an information-intensive environment. Nevertheless, there are several prospects for continuing on this matter. For example, future research may also explore the nature of the particular elements among which complementarity emerges by applying the interaction approach of the complementarity theory (Ennen & Richter, 2010). For example, to explore how much decentralization of decision-making is needed for a certain operational process. Such studies may provide greater specificity of the relationships between individual factors. At the same time, an understanding how exactly complementary factors may increase IT-enabled productivity will also advance work in this area. An understanding of complementary factors and their configurations may help determine where both research and managerial efforts need to be concentrated in order to enhance individual productivity of information workers.

Third, the IT system construct in this research was black-boxed and, thus, homogenous with regard to its functionality, usability, and usefulness (Davis, 1989). Previously, it was noticed that when studies aggregate productivity effects over IT applications, the results are difficult to interpret and draw conclusions (Athey & Stern, 2002). Therefore, we limited ourselves to a specific situation when a new mandatory IT system with functionality closely aligned with the work process is used by information workers to increase their productivity. The empirical investigations of this research, however, can be advanced in the following ways. For example, previous studies show that various IT systems may show a different degree of support, or fit, with various kinds of work processes (Mitchell & Zmud, 1999; Collins, 2003) and thus provide better opportunity for productivity increase, and suggest that different degrees of support or fit with the work process, provide greater opportunity for high productivity in relation to individual cognitive style (Chen &

215

Popovich, 2003). Previous studies also propose that various cognitive styles do have a preference with regard to the representational mode of information (Benbasat & Dexter, 1982; Riding & Douglas, 1993) – where a key distinction advanced is between graphical representation versus numerical representation modes, which also gives rise to differences in information worker productivity (Isaksen et al., 1992). These complementarities may also advance work in the area of complementarities and IT-enabled productivity.

Fourth, although the formulated research model is grounded on the complementarity theory that offers a broad perspective to understand complex relationships between complementary factors, we acknowledge that the research model is based on one instrument for measuring individual cognitive style, derived from Kirton’s adaption-innovation theory. However, there are other frameworks and instruments for assessing cognitive style and personality preferences such as intuitive, analytical, and integrated (Agor, 1984) or analytic, conceptual, directive, and behavioral (Rowe & Mason, 1987) that may help develop new complementary factors of productive IT use.

Fifth, the research model was formulated based on complementary factors that are identified in previous research as the most relevant for adaptive/innovative cognitive style. Yet, we are aware of the potentially existent unobservable factors which, given the right set of unobserved correlations, could mimic the correlation patterns resulting from the complements. For example, companies can have a different organizational culture (Roberts, 2007) or top management support (Shao et al., 2016) which, together with the use of a new IT system, can make an impact on individual productivity. Yet, due to empirical settings we limited ourselves to certain set-ups as it is impossible to test all possible combinations of complementary factors. We also decided to apply chosen factors and restrict some possible complementary relations in order to decrease model overload and simplify the research design. Although the formulated research model is grounded on the complementarity theory that offers a broad perspective to understanding complex relationships between complementary factors, the theoretical framework presented in this research can be further developed in many ways. For example, the chosen set of complementary factors satisfies the aim of the study and addresses a particular research problem of individual productivity increase when a more aligned IT system is used. Yet, any other complementary factors not included here should be regarded in further studies as subsequent to the present research. Further alternation of the work design may unearth new complementary factors that can increase information worker productivity.

Sixth, our first study was conducted in the given professional context of pharmaceutical sales operations in the Nordic region with a quasi-randomized assignment of the treatment that could bias our results. However, restricting our study to one company enabled us to control extraneous factors. Moreover, given the four-by-four design of this study, we regard this as reasonable. In

216

this study, we examined a relatively small sample size – 91 participants, yet reached rich and unique data. It is also possible that our study’s results are not generalizable across sales companies. However, the results of this study remain stable and confirm the interplay of complementarity factors. Therefore, there is a need to replicate this study across different sales companies as well as across different information-intensive occupations with different populations to investigate the generalizability of the findings.

Seventh, in the first study, we investigated the impact of complementarities of individual IT-enabled productivity by applying DID analysis. This analysis helped us demonstrate the impact of treatment on outcome variables by comparing the average change over time for the treatment groups and the control group. Yet, although Kirton (1976) emphasized that adaptors and innovators perform equally, it was established that they demonstrate different learning patterns before the use of a new IT system is stabilized (McLeod et al., 2008). Therefore, future research and analysis can uncover the difference between adaptors and innovators with regard to complementarity set-ups in learning patterns and times required to reach stabilization in IT system. This may help managers to understand the manner in which both adaptors and innovators learn and master a new IT system and how this affects their productivity.

Eighth, we also acknowledge some limitations of the second study where we investigated the productivity of a software programmer in the online experiment. We collected data mostly for individuals with innovative cognitive style. The online experiment as such and its design made it difficult to validate the research model. Yet, this study generated side-effects knowledge that can be used further to improve the research design and strategy to study complementarities of productive IT use. For example, our research design consisted of before-, after-, and follow-up sessions. Yet, this was not enough to achieve saturation in the use of both IT systems. Therefore, future experimental studies have to be based on multiple sessions to created meaningful results. This can be achieved by time series design of the experiment. It is also recommended to use a laboratory experiment to address control over main complementarities of productive IT use (Brynjolfsson & Milgrom, 2013). Therefore, a longitudinal field experiment and a randomized controlled experiment can be further used to demonstrate the impact of complementarities on individual information worker IT-enabled productivity.

Ninth, the formulated research model was tested in two information-intensive occupations; sales representatives and software programmers. Yet, in previous chapters, we mentioned that information-intensive occupations are characterized by high diversity and present sustainable growth. The formulated research model can be further tested in different information-intensive settings. For example, such professions as journalists, physicians, accountants, architects are all information-intensive occupations. This can

217

strengthen the model validity and provide stronger generalizability of the obtained results.

Tenth, most of the studies on the relationships between complementary factors and IT-enabled productivity applied quantitative approaches. Yet, this approach provides little knowledge on the nature of the relationships that drive complementarity effects. Further study with a qualitative approach may provide details of individual factors’ relationships. For example, case studies and action research at individual level can increase our understanding of complementary factors in more detail. It would also be informative to consider more complicated systems of components, with a greater number of complementarities and more complex interactions between them.

In summary, in this research we have used the complementarity theory as a basis for research model formulation. It is not, however, unlikely that other theories may contribute to the understanding of factors that affect IT-enabled individual productivity of information workers. For example, according to the person-environment fit theory (Edwards, 1991), there are different dimensions of fit between and individual and working environment. According to this theory; person-job fit, person-group fit and person-person fit may be potential complementarities that enhance individual performance of information workers. Future research may develop and/or refine existing complementary factors and their interrelationships to enhance the potential of managerial work in order to drive individual productivity of information workers. These theories, however, lie beyond of the scope of this research.

218

219

7. Summary and conclusions

In this chapter, we summarize and present the main findings of the undertaken research with regard to the research question. We also present general conclu-sions based on the obtained findings of this research and demonstrate how the research aim was achieved. This chapter ends with a discussion of major in-sights and a review of the research contributions to knowledge made by this research.

7.1 Summary of this research The aim of this research was to develop and test a new research model of complementary factors that affect individual information worker IT-enabled productivity. Achieving this aim was important because it constitutes an at-tempt to address the current lack of knowledge about complementarities that can increase individual IT-enabled productivity of an information worker. The research aim was formulated after an extensive literature review. This review demonstrated that while a number of factors can affect individual productivity, there is a dearth of knowledge of just which complementary factors are needed and how they should be synchronized to increase IT-enabled productivity of information worker at the individual level. In order to achieve the aim of this research, the following question was proposed:

What are the configurations of complementary factors that influence productivity, considered at nano-level (i.e. individual/process/task) in the con-text of an information-intensive environment, when a more aligned IT system is used?

This research question helped us formulate the research model that accounts for complementary individual and organizational factors in relation to individual productivity increase when a more aligned IT system is used. The research model was designed on the basis of the systems approach of the complementary theory, which states that the presence of one factor increases the value of others. Literature on cognitive style differences, the structure of the operating process and human resource management helped us identify

220

specific complementary factors which, when correctly matched, can increase individual IT-enabled productivity. Based on this literature two research hypotheses were formulated. Particularly, we hypothesized that individuals with adaptive cognitive style will generate higher productivity when matched with a structured operating process, push mode training in work technology, exogenous incentives, and centralized decision-making, compared to other configurations of these factors. In contrast, individuals with innovative cognitive style will generate greater productivity when matched with a flexible operating process, a combination of minor upfront mandatory training with optional on-demand training in work technology, endogenous incentives, and decentralized decision-making, compared to other configurations of these factors. We formulated both hypotheses based on earlier research that has already established double or triple interrelations between complementary factors.

To test the research model and developed hypotheses, we identified two information-intensive contexts with objective performance metrics; imple-mented changes based on the developed research model, and gathered data before and after a more aligned IT system was introduced. We investigated a detailed workflow for both sales representatives and software programmers and offered substantial insights into how information workers actually create value, which complementary factors can make them more productive when a more aligned IT system is used, and how the productivity of an information worker at individual level can be assessed. The complementary approach and the results of the subsequent empirical investigation have driven a unique per-spective to answer the research question that motivated this research.

The obtained findings, in particular, provide strong evidence that such fac-tors as operating process structure, training and education mode of new tech-nology, incentive mode, and decision-making structure represent a clear com-plementary configuration that has to match adaptive/innovative cognitive style in a situation when a more aligned IT system is used in order to drive individual information worker productivity. Our findings also suggest that in-dividual productivity of information workers can hardly be improved merely by the introduction of a more aligned IT system without matching individual and organizational complementarities. The findings also demonstrate that the introduction of a more aligned IT system with a limited and/or wrongly as-sumed complementarities can negatively affect individual information worker productivity. In principle, our empirical results give strong support to the statement that IT use is essential, but not sufficient to increase individual productivity of employees. Besides, the results support the time lag hypothesis due to learning for productivity growth to show up after the deployment of a more aligned IT system.

Our empirical investigation also demonstrates how the research design of future experimental studies that intend to test complementarities of productive IT use at individual level can be improved to demonstrate meaningful results.

221

First, we demonstrated that since more runs are required to establish saturation with both old, less aligned and new, more aligned IT systems; time series de-sign of the experiment is required. Second, since the online experiment showed that most participants had an innovative cognitive style, we recom-mend focus being placed either on field or on laboratory experiments to ensure sufficient productivity data for both adaptive and innovative cognitive style individuals. Although field experiments help to overcome operationalization limitations of complementary factors, laboratory experiments help to address control over main complementarities of productive IT use. While our online experiment had some technical and operationalization limitations, its findings support current wisdom of learning effect from repeated work functions, new IT system use and its effect on individual productivity.

7.2 Contributions in conclusion In general, the obtained findings of this research provide substantial support for the notion of complementarity in relation to IT-enabled productivity. The results of this research offer a potential theoretical contribution to individual information worker productivity literature, studies that investigate the impact of IT use on individual productivity as well as literature on managerial and organizational economics where complementarities are investigated in rela-tion to IT-enabled productivity. The results also offer managerial implications, calling for a contingent approach to achieve productivity increase from IT use.

First, to theory, the complementary approach applied in this research can be used to further our knowledge on complementary factors of productive IT use. Research to date that has applied this approach at individual level has not progressed sufficiently due to significant challenges in complementarities management and difficulties in finding complementary relationships between a limited number of factors. Yet, the results of this research demonstrate that the contribution of complementary factors in information-intensive work is considerable and can be estimated in terms of individual productivity with high accuracy. The established configurations of complementarity factors are novel and can further be used to study individual IT-enabled information worker productivity.

Second, in this research, we developed a complementarities-based model of IT use and individual productivity. A distinguishing feature of the research model development is the application of the systems approach of the comple-mentarity theory and the synthesis of a set of factors such as adaptive/innova-tive cognitive style, the structural complexity of the operational process, and most common human resource management practices of productive IT use. We built upon prior work that addressed in part the relationships between

222

those factors but added complementary evidence by studying them together. The investigated complementary factors are not unique, yet the proposed con-figurations are novel. This is another theoretical key contribution of this re-search.

Third, our model posits that individual productivity will increase when necessary complementary factors are introduced together. We illustrated the model application in two very different information-intensive environments: sales operations and software construction. This research extends the current knowledge of complementary factors of productive IT use. We demonstrate that operational process structure, training and education, incentives, and de-cision-making structure represent a clear complementary configuration when correctly matched to adaptive/innovative cognitive style in a situation when a new, more aligned IT system is used to increase individual productivity. Therefore, the obtained results contribute to the literature on the use of IT for productivity increase, supporting the argument that organizational and indi-vidual complementary factors lead to higher IT-enabled productivity.

Fourth, an earlier recommendation has been to conduct more longitudinal research and also provide greater control over investigated variables to iden-tify the exact impact of complementary factors on IT-enabled productivity (Brynjolfsson & Milgrom, 2013). By conducting the field and online experi-ments, we provide support for the causal relationship between complementa-rities and individual productivity. The results demonstrate that in a situation when a more aligned IT system is used to increase individual productivity complementarities matter. Yet, further research is required to identify differ-ent possible individual and organizational elements that complement IT use. Moreover, the recommendations developed in this research for the research design of experiments can be taken into account to ensure success of studies.

Therefore, to date, important advances in assessing the impact of comple-mentary factors on IT-enabled productivity have been made at firm and estab-lishment (plant) levels. In contrast, our research opens three new frontiers: (i) detailed process-level evidence of information worker output in different in-formation-intensive environments, (ii) detailed investigation of new IT sys-tems, operational processes, and human resource management practices, (iii) objective measures of complementary factors. This approach provides an ef-fective tool for studying individual productivity of information workers, re-vealing finer grained relationships between complementary factors that would be possible in any type of information work when a new IT system is used. Three practical contributions result from this approach.

First, the results suggest that managers who wish to increase productivity of their information workers should assume a comprehensive approach that accounts for a number of distinct factors, including cognitive style of workers, operational process structure, training and education mode in new technology, incentives, and decision-making structure when a new, more aligned IT sys-

223

tem is used. Particularly, the results of this research suggest two configura-tions of complementary factors that, when matched with the appropriate indi-vidual cognitive style, may increase individual IT-enabled productivity. The proposed configurations of complementary factors are specifically useful for managers who are restructuring their organization to take advantages of new IT systems use. Knowing complementary factors of productive IT use also enables managers, economists, and engineers to better understand where ad-ditional resources are needed and which processes require more attention in order to increase individual IT-enabled productivity.

Second, our results also suggest that managers should account for the time lags inherent in non-trivial productivity improvement programs, partly caused by necessary learning costs and effects of new IT system deployment and use. Particularly, our results demonstrate that managers should pay more detailed attention to the learning curve of a new, more aligned IT system as well as take into account time lags that are required for the emergence of an IT-ena-bled productivity increase. Therefore, managers should carefully consider the time that is needed for a productivity increase to show up after a new, more aligned IT system is deployed and used when they develop their evaluative strategies.

Third, our research findings send a warning signal to managers that installing a new IT system without complementarities is far from sufficient to increase productivity. Indeed, our results suggest that the introduction of a new IT system without complementary factors may generate a productivity decrease. Moreover, managers should adequately assume complementary factors, otherwise addressing only a few factors in isolated manner can also affect individual productivity negatively. This highlights the need for managers to acquire in-depth knowledge of complementary factors of productive IT use. Overall, our findings demonstrate that managers should neither overestimate the potential impact of introducing a new, more aligned IT system on individual productivity, nor underestimate the impact of complementary factors on individual IT-enabled information worker productivity.

In conclusion, our empirical results provide strong support that complementarities matter for individual information worker IT-enabled productivity. The results of this research also enrich our understanding of new configurations of complementarities that, when matched correctly, can increase individual information worker IT-enabled productivity. Yet, future research can advance the research model and empirical design presented in this research to uncover new or support established configurations of complementary factors that can positively affect information worker IT-enabled productivity at individual level.

224

225

Appendix A

Questionnaire Dear participant, please answer the following questions: What is your gender? Male Female In what year were you born? ____ What is your marital status? ____ What is the highest level of education you have completed? Bachelor degree Master degree PhD degree How many years of experience do you have in the sales industry? ___ How many years of experience do you have of the sales representative func-tion? ____ How many years of experience do you have in this particular company? _____ Please rate from 1 (not at all) till 7 (significantly): I am highly effective at using new technological tools. This means I know what information they con-tain and can easily find, add, and modify the records I need. I have control over the information I use; I can access and modify it at will. _____ What proportion of your working time do you spend using current (new) tech-nological system? ____ Please rate from 1 (not at all) till 7 (significantly): My job tasks are highly interdependent with other people’s tasks. I must often coordinate with other team members. ______ Please rate from 1 (not at all) till 7 (significantly): My data requirements are highly routine. I could specify all I need on a standard form. ______

226

How long have you been working on this particular market? Denmark _______ Sweden _______ Finland _______ Norway _______ Please, complete the following statements by choosing the option that most closely matches your opinion for each row: Please consider how you feel about the following statements. Circle your choice on the scale provided. Table 1: Self-assessment form

Questions I am a person who:

Strongly disagree

Disagree Agree Quite agree

Strongly agree

Has original ideas Proliferates ideas Is stimulating Copes with several new ideas at the same time

Will always think of something when stuck

Would sooner create than improve

Has fresh perspectives on old problems

Often risks doing things differently

Likes to vary set routines at a moment’s notice

Prefers to work on one problem at a time

Can stand out in disagree-ment against group

Needs the stimulation of frequent change

Prefers changes to occur gradually

Is thorough Masters all details pains-takingly

Is methodical and system-atic

Enjoys detailed work

227

Table 1: Continued from previous page

Questions I am a person who:

Strongly disagree

Disagree Agree Quite agree

Strongly agree

Is (not) a steady plodder Is consistent Imposes strict order on matters within own control

Fits readily into “the sys-tem”

Conforms Readily agrees with the team at work

Never seeks to bend or break the rules

Never acts without proper authority

Is prudent when dealing with authority

Likes the protection of precise instructions

Is predictable Prefers colleagues who never “rock the boat”

Likes bosses and work patterns which are con-sistent

Works without deviation in a prescribed way

Holds back ideas until ob-viously needed

Thank you for your participation!

228

Appendix B

Summary of the training and education procedures utilized during the launch of the new IT system

Overview of the training and education procedure

§ Pre-launch activities:

Activity 1: General and brief information provided (5-10 min.):

• to all employees in the Nordic affiliate; • in the form of a regular “town hall meeting”; • among other matters that meeting included information about a forth-

coming global salesforce efficiency program that is to be deployed also in the Nordic affiliate;

• this information was provided more than six months prior to the in-troduction of the change.

Activity 2: Dedicated meeting with specific information:

• half a day length; • details of the salesforce efficiency initiative were disclosed; • to all salesforce in the Nordic affiliate, one meeting per market; • content: what and why, how, when, who, including discussions and

expectations identification; • in the form of a physical meeting.

§ Launch activities:

Activity 0: Measurement of the usage of the IT system and how its use initi-ated; Activity 1: Detailed information in digital format (i.e. e-learning module) about the new operations canalised to each sales representative, adapted to the configuration of the business unit’s salesforce efficiency design;

229

Activity 2a: For the adaptors, upfront training and education session, one full day in the office; Activity 2b: For all sales representatives, an additional e-learning module cus-tomised to each sales business unit, sent out to the sales representatives, aimed for re-active self-learning.

§ Post-launch activities: Activity 1a: A follow up training and education session in the office, half a day, offered to all adaptors; Activity 1b: Introduction of a call-center, where all sales representatives can call for support (via phone and/or email), both operational (i.e. how to work) and technical (i.e. troubles with the new IT system); Activity 2: Best practice sharing:

• establishment of a best practice community where sales representa-tives can share ideas and suggestions for how to work across the Nor-dic organization;

• continued monitoring of the usage statistics of the system and also bug fixing.

230

Appendix C

Guidelines for changes in relation to the introduction of the new CRM-system

Study Design 1: No Change Study Design 2: Structured Partial Change Study Design 3: Semi-Structured Partial Change Study Design 4: Full Change

Study Design 1: No Change 1. IT system: Use the Old IT system. 2. Working Process (structured approach): - All / Most working-process steps are Mandatory (i.e. as it was).

Study Design 2: Structured Partial Change l. IT system: - Introduce the New IT system. 2. Working Process (structured approach): - All / Most working-process steps, for the new sales process, are Mandatory. 3. Organization Set-up: a) Training and Education provided:

231

- Provide both Push (sales representatives have to participate in the initial Training and Education) and Pull (offer sales representatives further Training and Education on request when needed). b) Sales Incentive: - Offer a Bonus System with a maximum up 6 months’ salary, on realized sales objectives (with a gradual adjustment for bonus versus Sales). c) Sales Decision-Making Centrality (as was): - Split decision-making authority; key decisions to the Sales Manager, minor decisions made by sales representatives (as it was).

Study Design 3: Semi-Structured Partial Change l. IT system: - Introduce the New IT system. 2. Working Process (semi-structured approach): - Most working-process steps, for the new sales process, are Not Mandatory (Only Reporting is Mandatory). 3. Organization Set-up: a) Training and Education provided: - Provide both Push (sales representatives have to participate in the initial Training and Education) and Pull (offer sales representatives further Training and Education on request when needed). b) Sales Incentive (as it was): - Offer a Bonus System with a maximum up 6 months’ salary, on realized sales objectives (with a gradual adjustment for bonus versus Sales). c) Sales Decision-Making Centrality (as was): - Split decision-making authority; key decisions to the Sales Manager, minor decisions made by sales representatives (as it was).

232

Study Design 4: Full Change For sales representatives that denoted as ‘Adaptors’ l. IT system: - Introduce the New IT system. 2. Working Process (structured approach): - All / Most working-process steps, for the new sales process, are Mandatory; - Assign these sales representatives to customers that are Easy-to-Access. 3. Organization Set-up: a) Training and Education provided: - Provide both Push (sales representatives have to participate in the initial Training and Education) and Pull (offer sales representatives further Training and Education on request when needed). b) Sales Incentive: - Offer a Bonus System with a maximum up 6 months’ salary, on realized sales objectives (with a gradual adjustment for bonus versus Sales). c) Sales Decision-Making Centrality: - Deliver most decisions to the Sales Manager, rather than to sales representa-tives. For sales representatives denoted as ‘Innovators’ l. IT system: - Introduce the New IT system. 2. Working Process (semi-structured approach): - Most working-process steps, for the new sales process, are Not Mandatory (Only Reporting is Mandatory); - Assign these sales representatives to customers that are Hard-to-Access.

233

3. Organization Set-up: a) Training and Education provided: - Provide only Pull Training and Education (offer sales representatives only further Training and Education on request when needed). b) Sales Incentive: - Offer a Bonus System with a maximum up 6 months’ salary, on realized sales objectives (with a gradual adjustment for bonus versus Sales); - Provide these sales representatives constant Feedback about their work. c) Sales Decision-Making Centrality: - Deliver most of decisions to sales representatives, rather than to Sales Man-agers.

234

Appendix D

Initial instructions

We invite you to participate in an experiment where we try to understand the conditions for productivity of software developers. Software developers who know how to program in general are particularly invited to participate in this exciting experiment! The programming language is JavaScript, but as long as you know some language that is at least similar, such as Java, C, Python, it will be good enough. Please do not participate in this campaign if you don’t know the programming languages described above.

The experiment contains one questionnaire and three sessions (assignments). Each assignment consists of a simple, incomplete HTML/JavaScript application. Your objective is to modify the JavaScript code in order to fulfill the requirements for the application. The questionnaire should take between 5 and 15 minutes to fill in. Each assignment has been designed to take between 20 minutes and one hour to complete. So you can expect to spend around one to three hours on this experiment, but you may of course take breaks between sessions (assignments). Preferably, all tasks should be completed on the same day. All tasks must be completed using the same machine. The faster and more accurately you can fulfill all of the requirements of each challenge, the better.

During the first session, you must animate a Shape in a HTML Canvas by changing the Shape’s size and position over time. You should not use the help of any tools like IDEs or even external text editors (you can write the code from the web page itself), though you should feel free to look up information online, of course.

In the second and third sessions, you will be invited to make use of an online development tool designed to make software developers more productive. One of our goals is to find out how the use of such a tool affects developers.

During the second session, you have to apply logical rules to sort items in a ToDo list by priority and due date.

During the third session, you need to implement a simple text analysis tool that helps find useful information about text using search expressions.

You participate anonymously and your results will not be communicated to anyone outside the research project.

235

Important! Required proof of the job finished is: 1. Three completed assignments and questionnaire; 2. Reasonable time for each assignment; 3. Satisfactory quality level of the developed products.

236

Appendix E

Questionnaire Personal What is your gender?____ In what year were you born?____ What is your marital status?____ Education and experience What is the highest level of education you have completed? ____ How much experience do you have as a programmer? ____ How much experience do you have using JavaScript?____ How many programming languages are you familiar with? _____ How much of your working time do you spend using IT systems?_____ Please rate from 1 (not at all) till 5 (significantly): I am highly effective at using new technological tools. This means I know what information they con-tain and can easily find, add, and modify the records I need______ What proportion of your working time do you spend using technological sys-tems?_____ Please rate from 1 (not at all) till 5 (significantly): My work is routine. I could specify all I need on a standard form.______

237

Please consider how you feel about the following statements. Circle your choice on the scale provided. Table 2: Self-assessment form

Questions I am a person who:

Strongly disagree

Disagree Agree Quite agree

Strongly agree

Has original ideas Proliferates ideas Is stimulating Copes with several new ideas at the same time

Will always think of something when stuck

Would sooner create than improve

Has fresh perspectives on old problems

Often risks doing things differently

Likes to vary set routines at a moment’s notice

Prefers to work on one problem at a time

Can stand out in disagree-ment against group

Needs the stimulation of frequent change

Prefers changes to occur gradually

Is thorough Masters all details pains-takingly

Is methodical and system-atic

Enjoys detailed work Is (not) a steady plodder Is consistent Imposes strict order on matters within own control

Fits readily into “the sys-tem”

Conforms Readily agrees with the team at work

238

Table 2: Continued from previous page

Questions I am a person who:

Strongly disagree

Disagree Agree Quite agree

Strongly agree

Never seeks to bend or break the rules

Never acts without proper authority

Is prudent when dealing with authority

Likes the protection of precise instructions

Is predictable Prefers colleagues who never “rock the boat”

Likes bosses and work patterns which are con-sistent

Works without deviation in a prescribed way

Holds back ideas until ob-viously needed

Thank you!

239

Appendix F

Description of the assignments

Assignment 1 – Animation problem

Your client has a simple HTML5/JavaScript web application. They would like to have a simple but attractive animation on a Shape object that will be used to animate Shapes while users wait for long-running processes to run inside the application. The animation should change the Shape object (a black rectangle) in a smooth way, using the whole available area (the rectangular area of the Canvas object) in as varied a manner as possible. All the existing code is contained in 2 files:

• index.html (the HTML container); • animation.js (the javaScript code that draws on the canvas element).

You do not need to modify the HTML file. You need to implement your

animation by modifying the Shape object and its prototype (you can add more properties and functions to the Shape objects). To run the application, simply open the HTML file with any modern browser. The URL will look like this: file:///home/it-exercises/animation/index.html

Refreshing the page in the browser will restart the animation from the be-ginning.

§ For structured process

The Shape should move using the following algorithm: • start by moving the Shape to the right and up simultaneously; • the shape should move at a constant speed of approximately 10 pixels

per second on the x-axis (left and right), and 5 pixels per second on the y-axis (up and down). Assume the tick function runs 60 times per second;

• if hitting the top edge, keep moving in the same direction on the x-axis, but instead of going up, start moving down at the same speed;

240

• if hitting the bottom edge, keep moving in the same direction on the x-axis, but instead of going down, start moving up at the same speed;

• if hitting the right edge, keep moving in the same direction on the y-axis, but instead of going right, start moving left at the same speed;

• if hitting the left edge, keep moving in the same direction on the y-axis, but instead of going left, start moving right at the same speed.

The Shape should change size using the following algorithm:

• start by growing the Shape; • the size should change by 1 pixel per second, approximately, in both

width and height, until the size has changed by 100 pixels in each direction;

• once the maximum size is reached, start shrinking it back to the initial size, starting again from step 7;

• the whole Shape should always be fully visible within the bounds of the canvas.

§ For flexible process

After 10 minutes working on the exercise, the following instruction was

given. The client has requested the following: • the animation should change both the position and size of the Shape; • the Shape should move around the whole available area.

Assignment 2 – To-Do List Problem

Your client is developing an application for managing To-Do lists. Users

can create lists of tasks using the JSON text format (the user interface for cre-ating and editing tasks is not ready yet) and then import the list into a web interface for processing. Tasks may have any attribute, but the following at-tributes are treated specially:

• done; • dueDate; • priority.

Assume that every task will have at least these 3 attributes and they will

always have valid values (i.e. no need to define default values for these or handle errors in case they are not present).

• “done” should be a boolean (true or false);

241

• “dueDate” should be a String representing a date, for example “1 Jan-uary 2015 09:00” (or anything accepted by Javascript’s Date.parse() function);

• “priority” is an integer between 1 and 10, where higher value means a higher priority.

Hint: JavaScript allows you to compare dates easily. Example: var af-ter20thCentry = new Date() >= Date.parse(“1 January 2001”).

All the existing code is contained in 3 files: • index.html (the HTML container); • todo.js (the javaScript code implementing the logic of the app); • todo.css (the stylesheet).

There is also a file called todoList.json which is just an example JSON file

that can be used as input for the program. IMPORTANT: to make this assignment predictable, pretend the current

date and time is: "01 May 2015 09:00" (just use the now variable in the todo.js file when you need the current date-time).

You do not need to modify the HTML and css files, but feel free to do it if you want. The JS file is the one you should modify (look for TODOs which help you locate where your modifications should go).

Hint: Notice that to change the background color of a task; you can add a

CSS class declared in the todo.css file to its HTML element. For example: “list-item late”. To run the application, simply open the HTML file with any modern browser. The URL will look like this: file:///home/it-exercises/to-do-list/index.html

§ For structured process Your mission is to improve the web interface so that users may see their

tasks in the order they should be completed. The algorithm used to decide which tasks have higher priority will be described below. If a task is done, then it should be displayed on a light green background in the User Interface and it should not be shown interleaved with not-yet-done tasks.

Tasks that are overdue should be shown on a red background. Use the fol-lowing algorithm to determine which task should be done first:

• if a task is overdue, its effective priority is given by the number of

hours it is overdue plus 10 times its priority value;

242

• if a task is not overdue, its effective priority is given by 10 times the priority value, divided by the result of multiplying 0.5 by the number of hours to the due date;

• the task that should be done first is the task with the highest effective priority.

For example: if a task was due 2 hours ago and its priority is 5, the algo-rithm should calculate an effective priority of 5 * 10 + 2 = 52 for the task. If a task is due in 12 hours from now and its priority is 5, its effective priority is 5 * 10 / (0.5 * 12) = 8.333.

§ For flexible process

Your mission is to improve the web interface so that users may see their

tasks in the order they should be completed. If a task is done or overdue, then it should be displayed differently in the User Interface so that the user can see that. When calculating which task should be done first, the algorithm should increase the priority of a task for each hour it is overdue, so that tasks get a higher priority the later they become.

After 10 minutes working on the exercise, the following instruction was given. The client has come up with their own algorithm that should be used for overdue tasks! It is the following: effective priority = number of hours task is overdue + (10 * task priority). They also said you should use an algorithm that “makes sense” for tasks that are NOT overdue.

Assignment 3 – Text Analysis Problem

Write a program that takes some text as input and allows the user to analyze it to find interesting information, such as word count, list of lines and words matching a certain expression, etc.

To help you get started, the user interface is already implemented for you. Users can enter a search expression on a text field, and the text to be analyzed in the text area below it. The search expression should be treated as a regular expression. To perform the analysis, users click on the “Analyze text” button. The analysis results are shown at the bottom of the screen.

Hint: You can use this website to test your regular expressions: http://re-

gexpal.com/

All the existing code is contained in 2 files:

• index.html (the HTML container); • textanalysis.js (the javaScript code implementing the logic of the

app).

243

You do not need to modify the HTML file. You should implement the

analysis logic in the function getAnalysisResults(), which should return the result of the analysis as plain text or HTML.

Do not worry about the format of the text being analyzed (assume it is plain text and never HTML or JSON). To run the application, simply open the HTML file with any modern browser. The URL will look like this: file:///home/it-exercises/text-analysis/index.html

§ For structured process

The analysis report should contain the following information: • lines – displays each line containing a String matching the given ex-

pression. If no expression is given, shows all lines; • lineCount – shows how many lines contain a String matching the

given expression. If no expression is given, shows how many lines the document contains;

• words – displays each word matching the given expression. If no ex-pression is given, shows the whole document. Notice that “word” here means whatever the expression matches;

• wordCount – shows how many times a String matching the given ex-pression is found. If no expression is given, shows how many words (space-delimited group of characters) the document contains;

• regular expressions should be supported for every option. For exam-ple, the user should be able to search for every line containing a String matching the expression “Book\s*\d+”.

§ For flexible process

The analysis report presented to the user should be simple but clear. Don’t

worry too much about the appearance of the report, try to focus on the con-tents. After 10 minutes working on the exercise, the following instruction was given:

The analysis report should contain, at least:

• each line matching the expression; • how many lines match the expression; • each word matching the expression; • how many words match the expression.

If no expression is given, use some default behavior that users might have

expected the system to use.

244

Appendix G

JavaScript Help

Assignment 1 – Animation problem

To animate the rectangle in the canvas is a matter of changing the rectan-gle’s width, height, x- and -y coordinates. These are all properties of the Shape object (w, h, x, y respectively). In JavaScript, you can set the property of an Object by assigning a value for it as shown below: var shape = new Shape(); // set the height of the shape to 100 shape.h = 100 // increase the width of the shape by 10 shape.w = shape.w + 10 You may use conditional blocks to decide what to do as in most languages: by using an “if” block. For example: if (shape.w > 500) { // what to do if the shape’s width is larger than 500 } else { // what to do otherwise }

Assignment 2 – To-Do List Problem

Function todoItemAsHtml takes a todoItem object as an argument. This object will have the same properties as the equivalent json. For example, give the following json item: { "description": "Study for the CS exam",

245

"done": false, "priority": 1, "dueDate": "10 May 2015 10:00" } You could check the properties of the todoItem like this: todoItem.description // results in “Study for the CS exam”; if (todoItem.done) { // do something if the item is done } To convert the date String to a Date object (so you can easily compare dates), you can do the following: var dueDate = Date.parse(todoItem.dueDate); To concatenate Strings together, you can use the + operator: “hello” + “ “ + “world” // results in “hello world”;

Assignment 3 – Text Analysis Problem

This problem is mostly about string and array processing. JavaScript’s strings come with several methods which can be helpful. For example: To split a String called text into an array of lines: var lines = text.split(‘\n’); Concrete example: “abc\ndef\nghi“.split(‘\n’) // results in [ ‘abc’, ‘def’, ‘ghi’ ]; To get just part of a String starting from a certain index: var subString = text.substring(firstIndex); To find the index of the first occurrence of an expression (which can be a regular expression): var firstIndex = text.search(expression); // if this returns -1, the expression was not found To get all the parts of a string that match an expression:

246

var allMatches = text.match(expression); Working with arrays (similar to List in other languages) is very simple: var myList = [10, 20, 30]; for (i in myList) { console.log(“Item at index ” + i + “ is “ + myList[i]); } You can add items to an array: myList.push(40); // now myList is [10, 20, 30, 40] If you know functional programming, you can use map/filter/reduce in Ja-vaScript! var nonEmptyLines = text.split(‘\n’).filter(function(t) { return t.trim().length > 0; }); var quote = ‘“‘; var quotedItems = myList.map(function(item) { return quote + item + quote; }); var summedItems = myList.reduce(function(a, b) { return a + b; });

247

Appendix H

Quality assessment

Assignment 1 – Animation problem

§ For structured process

Check how many of the animation requirements were correctly imple-

mented. Each requirement is worth 10 points. The Shape should move using the following algorithm: 1. start by moving the Shape to the right and up simultaneously; 2. the shape should move at a constant speed of approximately 10 pixels per second on the x-axis (left and right), and 5 pixels per second on the y-axis (up and down). Assume the tick function runs 60 times per second; 3. if hitting the top edge, keep moving in the same direction on the x-axis, but instead of going up, start moving down at the same speed; 4. if hitting the bottom edge, keep moving in the same direction on the x-axis, but instead of going down, start moving up at the same speed; 5. if hitting the right edge, keep moving in the same direction on the y-axis, but instead of going right, start moving left at the same speed; 6. if hitting the left edge, keep moving in the same direction on the y-axis, but instead of going left, start moving right at the same speed; The Shape should change size using the following algorithm: 7. start by growing the Shape; 8. the size should change by 1 pixel per second, approximately, in both width and height, until the size has changed by 100 pixels in each direction; 9. once the maximum size is reached, start shrinking it back to the initial size, starting again from step 7; 10. the whole Shape should always be fully visible within the bounds of the canvas.

§ For flexible process

A simple guideline for assessing unstructured assignment is how “rich” the animation is compared to the structured assignment animation, where

248

“rich” is defined as the number of different ways in which the shape changes and moves about: 1. if the Shape grows smoothly (15 points); 2. if the Shape shrinks smoothly (15 points); 3. if the Shape moves left and right (15 points); 4. if the Shape moves up and down (15 points); 5. if the Shape moves about approximately the whole available area (40 points).

Assignment 2 – To-Do List Problem

Solutions for all assignments should be evaluated according to how many of the stated goals were achieved. Not all goals were explicitly stated for par-ticipants doing the unstructured assignment and participants may have chosen to add some other features or ways of achieving similar goals which may be considered as achievements as well. The assessment of this assignment is done in the following way: 1. if all 7 issues show all required fields (done, dueDate, priority) (10 points); 2. dueDate field value is shown as a well formatted date (easy to know which date it is) (10 points); 3. extra fields are shown correctly; all items should have a description field (10 points); the item “Help Mary…” should have a field “reason” with value “Love!!!” (10 points); 4. overdue items are clearly shown (red background for structured assign-ment). Two issues should be overdue: “Fix the sink...” and “Go the the mar-keting workshop” (15 points); 5. done items are clearly shown (green background for structured assignment). Only one issue is done: “Finish the PL assignment” (15 points); 6. done items are not interleaved with others. This means the only done issue should be either the first item or the last item in the list (10 points); 7. the sorting algorithm is correctly implemented. For structured assignments, the correct order of items is the following (20 points): “Fix the sink...”; “Help Mary…”; “Go to the marketing workshop”; “Do the weekly shopping”; “Pay the electricity bill”; “Study for the CS exam”; “Finish the PL assignment”.

249

For unstructured assignments, it is considered ok if a maximum of one item is not in the correct order.

Assignment 3 – Text Analysis Problem

§ For structured process

§ For flexible process

With the given expression the following requirements should be imple-mented (each is worth 10 points): With the expression box empty, the results should be: 1. number of words = 25; 2. number of lines = 7. When you type abc in the expression box: 3. number of words = 5; 4. number of lines = 3; 5. list of lines: abc def ghi jkl mno pqr abc abc; this is abc again; this is another abc. 6. the list of words should contain 5 times “abc”;

Check how many of the requirements were correctly implemented. Each requirement is worth 20 points. 1. lines – displays each line containing a String matching the given expres-sion. If no expression is given, shows all lines; 2. lineCount – shows how many lines contain a String matching the given expression. If no expression is given, shows how many lines the document contains; 3. words – displays each word matching the given expression. If no expres-sion is given, shows the whole document. Notice that “word” here means whatever the expression matches; 4. wordCount – shows how many times a String matching the given expres-sion is found. If no expression is given, shows how many words (space-de-limited group of characters) the document contains; 5. regular expressions should be supported for every option. For example, the user should be able to search for every line containing a String matching the expression “Book\s*\d+”.

250

When you type t.+?a in the expression box: 7. number of words = 3; 8. number of lines = 2; 9. list of lines: this is abc again; this is another abc. 10. list of words (or matches): this is a; this is a; this is a.

If the participants fail to implement some of the requirements but manage to implement other, different information, points should be granted as if the participants had met one of the requirements (the more characteristics the re-port presents, the better the quality of the software).

251

Appendix I

Table 3: Validity assessment form

Extremely discouraged

Very discouraged

Somewhat discouraged

Neutral Somewhat encouraged

Very encouraged

Extremely encouraged

When I use a new version of CRM-system which is more functional, I feel

When almost all of my customers are standard prescribers, I feel

When almost all of my customers are KOL, I feel

When my manager asks me: “…all steps in your working process are mandatory, please follow them”, I feel

252

Table 3: Continued from previous page

Extremely discouraged

Very discouraged

Somewhat discouraged

Neutral Somewhat encouraged

Very encouraged

Extremely encouraged

When my manager asks me: “…do your work as you wish with only one mandatory step – reporting”, I feel

When I receive initial training and education in a new CRM-system, I feel

When I receive on-demand training and education in a new CRM-system, I feel

When I receive bonuses, I feel

When I receive constant feedback from managers, I feel

When I have to discuss almost all im-portant decisions such as deviation from the budget with a manager, I feel

When I have to make important decis-ions such as deviation from the budget on my own , I feel

253

References

Abdolmohammadi, M., & Wright, A. (1987). An examination of the effects of expe-rience and task complexity on audit judgments. The Accounting Review, 62(1), 1-13.

Adams, S. J. (1963). Towards an understanding of inequity. Journal of Abnormal and Social Psychology, 67, 422-436.

Adler, R. F., & Benbunan-Fich, R. (2012). Juggling on a high wire: Multitasking ef-fect on performance. International Journal of Human-Computer Studies, 70(2), 156-168.

Agarwal, R., & Prasad, J. (1999). Are individual differences germane to the ac-ceptance of new information technologies? Decision Sciences, 30(2), 361-391.

Agor, W. H. (1984). Intuitive management. Englewood Cliffs, NJ: Prentice Hall. Aguinis, H., & Kraiger, K. (2009). Benefits of training and development for individ-

uals and teams, organizations, and society. Annual Review of Psychology, 60, 451-474.

Ahadi, H. R. (2004). An examination of the role of organizational enablers in business process reengineering and the impact of information technology. Information Re-sources Management Journal, 17(4), 1-19.

Ahearne, M., Jelinek, R., & Rapp, A. (2005). Moving beyond the direct effect of SFA adoption on salesperson performance: Training and support as key moderating factors. Industrial Marketing Management, 34(4), 379-388.

Alexander, E. R., & Wilkins, R. D. (1982). Performance rating validity: The relation-ship of objective and subjective measures of performance. Group & Organization Management, 7(4), 485-496.

Alfes, K., Truss, C., Soane, E. C., Rees, C., & Gatenby, M. (2013). The relationship between line manager behavior, perceived HRM practices, and individual perfor-mance: Examining the mediating role of engagement. Human Resource Manage-ment, 52(6), 839-859.

Alter, C. E. (2001). Creativity styles and personality characteristics. Unpublished doctoral dissertation, Fordham University, New York.

Alter, S. (2013). Work system theory: Overview of core concepts, extensions, and challenges for the future. Journal of the Association for Information Systems, 14(2), 72.

Amabile, T. M. (1996). Creativity in context: Update to “the social psychology of creativity”. Boulder, CO: Westview Press.

Amabile, T. M., Hill, K. G., Hennessey, B. A., & Tighe, E. M. (1994). The Work Preference Inventory: Assessing intrinsic and extrinsic motivational orientations. Journal of Personality and Social Psychology, 66(5), 950.

American Psychological Association. (2009). Publication manual of the American psychological association. Washington: American Psychological Association.

254

Angrist, J. D., & Pischke, J-S. (2008). Mostly harmless econometrics: An empiricist’s companion. Princeton University Press, Princeton, NJ.

Antikainen, R., & Lönnqvist, A. (2006). Knowledge work productivity assessment. Institute of Industrial Management. Tampere University of Technology. PO Box, 541, 79-102.

Appelbaum, S. H., Hébert, D., & Leroux, S. (1999). Empowerment: Power, culture and leadership – A strategy or fad for the millennium?. Journal of Workplace Learning, 11(7), 233-254.

Appelbaum, S. H., Marchionni, A., & Fernandez, A. (2008). The multi-tasking para-dox: Perceptions, problems and strategies. Management Decisions, 46(9), 1313-1325.

Apte, U., Karmarkar, U., & Nath, H. (2012). The US information economy: Value, employment, industry structure, and trade. Foundations and Trends (R) in Tech-nology, Information and Operations Management, 6(1), 1-87.

Aral, S., & Van Alstyne, M. (2007). Network structure and information advantage. Proceedings of the Academy of Management Conference. Philadelphia, PA.

Aral, S., & Van Alstyne, M. (2011). The diversity-bandwidth tradeoff. American Journal of Sociology, 117(1), 90-171.

Aral, S., Brynjolfsson, E., & Van Alstyne, M. (2006). Information, technology and information worker productivity: Task level evidence. Proceedings of the 27th Annual ICIS. Milwaukee, Wisconsin, USA, 285-306.

Aral, S., Brynjolfsson, E., & Van Alstyne, M. (2011). Information, technology and information worker productivity. http://papers.ssrn.com/sol3/papers.cfm?ab-stract_id=942310. Date accessed 2012-01-05.

Aral, S., Brynjolfsson, E., & Van Alstyne, M. (2012a). Information, technology and information worker productivity. Information System Research, 23(3), 849-867.

Aral, S., Brynjolfsson, E., & Wu, L. (2012b). Assessing three-way complementarities: Performance pay, monitoring and information technology. Management Science, 58, 913-931.

Arbaugh, J. B. (2000). An exploratory study of the effects of gender on student learn-ing and class participation in an Internet-based MBA course. Management Learn-ing, 31(4), 503-519.

Armstrong, M. (2010). Armstrong’s handbook of reward management practice: Im-proving performance through reward. Kogan Page Publishers.

Armstrong, S. J., Cools, E., & Sadler-Smith, E. (2012). Role of cognitive styles in business and management: Reviewing 40 years of research. International Journal of Management Reviews, 14(3), 25-36.

Aronson, E., Wilson, T. D., Akert, R. M., & Fehr, B. (2007). Social psychology. To-ronto, ON: Pearson Education.

Arvanitis, S., & Loukis, E. N. (2009). Information and communication technologies, human capital, workplace organization and labor productivity: A comparative study based on firm-level data for Greece and Switzerland. Information Econom-ics and Policy, 21(1), 43-61.

Ashenfelter, O., & Card, D. (1985). Using the longitudinal structure of earnings to estimate the effects of training programs. Review of Economics and Statistics, 67, 648-660.

Athey, S., & Stern, S. (1998). An empirical framework for testing theories about com-plimentarity in organizational design (No. w6600). National Bureau of Economic Research.

255

Athey, S., & Stern, S. (2002). The impact of information technology on emergency health care outcomes. RAND Journal of Economics, 33, 339-432.

Atkinson, J., & Joel, O. R. (1978). Personality, motivation and achievement. Hemi-sphere Pub. Corp.

Auh, S., & Menguc, B. (2007). Performance implications of the direct and moderating effects of centralization and formalization on customer orientation. Industrial Marketing Management, 36(8), 1022-1034.

Autor, D. H., Levy, F., & Murnane, R. J. (2003). The skill content of recent techno-logical change: An empirical exploration. Quarterly Journal of Economics, 118(4), 1279-1333.

Axtell, C. M., Holman, D. J., Unsworth, K. L., Wall, T. D., Waterson, P. E., & Har-rington, E. (2000). Shopfloor innovation: Facilitating the suggestion and imple-mentation of ideas. Journal of Occupational and Organizational Psychology, 73(3), 265-285.

Baer, M., Oldham, G. R., & Cummings, A. (2003). Rewarding creativity: When does it really matter? Leadership Quarterly, 14, 569-586.

Bagozzi, R. P., & Foxall, G. R. (1995). Construct validity and generalizability of the Kirton adaption-innovation inventory. European Journal of Personality, 9(3), 185-206.

Bannister, F., & Remenyi, D. (2009). Multitasking: The uncertain impact of technol-ogy on knowledge workers and managers. Electronic Journal Information Systems Evaluation, 12(1), 1-12.

Barber, A. E., & Bretz, R. D. (2000). Compensation, attraction, and retention. Com-pensation in Organizations, 32-60.

Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44(1), 1-26.

Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9, 9-30.

Bartel, A., Ichniowski, C., & Shaw, K. (2007). How does information technology af-fect productivity? Plant-level comparisons of product innovation, process im-provement, and worker skills. Quarterly Journal of Economics, 122(4), 1721-1758.

Basoglu, K. A., Fuller, M. A., & Sweeney, J. T. (2009). Investigating the effects of computer mediated interruptions: An analysis of task characteristics and interrup-tion frequency on financial performance. International Journal of Accounting In-formation Systems, 10(4), 177-189.

Basu, S., & Fernald, G. F. (2007). Information and communications technology as a general-purpose technology: Evidence from US industry data. German Economic Review, 8(2), 146-173.

Bates, S. (2004). Getting engaged. HR Magazine, 49(2), 44-51. Becker, J. U., Greve, G., & Albers, S. (2009). The impact of technological and organ-

izational implementation of CRM on customer acquisition, maintenance, and re-tention. International Journal of Research in Marketing, 26(3), 207-215.

Beecham, S., Baddoo, N., Hall, T., Robinson, H., & Sharp, H. (2008). Motivation in software engineering: A systematic literature review. Information Software Tech-nology, 50(9-10), 860-878.

256

Bell, C. S., Compeau, D. R., & Olivera, F. (2005). Understanding the social implica-tions of technological multitasking: A conceptual model. Proceedings of the Fourth Annual Workshop on HCI Research in MIS, Las Vegas, NV, USA, 80-84.

Benbasat, I., & Dexter, A. S. (1982). Individual differences in the use of decision support aids. Journal of Accounting Research, 1-11.

Benbasat, I., & Taylor, R. N. (1978). The impact of cognitive styles on information systems design. Management Information Systems Quarterly, 2(2), 43-54.

Benbunan-Fich, R., Adler, R. F., & Mavlanova, T. (2011). Measuring multitasking behavior with activity-based metrics. ACM TOCHI, 18(2), 1-22.

Berings, M. G. M. C., Poell, R. F., & Simons, P. R. (2005). Conceptualizing on-the-job learning styles. Human Resource Development Review, 4, 373-400.

Bertrand, M., Duflo, E., & Mullainathan, S. (2004). How much should we trust dif-ferences-in-differences estimates?. Quarterly Journal of Economics, 119(1), 249-275.

Betsch, C., & Kunz, J. J. (2008). Individual strategy preferences and decisional fit. Journal of Behavioral Decision Making, 21(5), 532.

Biggs, J., & Moore, P. (1993). The process of learning. Australia: Prentice Hall. Birnbaum, M. H. (2004). Human research and data collection via the Internet. Annual

Review of Psychology, 55, 803-832. Biron, M., & Bamberger, P. (2010). The impact of structural empowerment on indi-

vidual well-being and performance: Taking agent preferences, self-efficacy and operational constraints into account. Human Relations, 63(2), 163-191.

Black, S. E., & Lynch, L. M. (2004). What’s driving the new economy?: The benefits of workplace innovation. The Economic Journal, 114(493), 97-116.

Blaylock, B. K., & Rees, L. P. (1984). Cognitive style and the usefulness of infor-mation. Decision Sciences, 15(1), 74-91.

Bloodgood, M. J. A., & Chilton, M. (2012). Performance implications of matching adaption and innovation cognitive style with explicit and tacit knowledge re-sources. Knowledge Management Research & Practice, 10(2), 106-117.

Bloom, N., & Van Reenen, J. (2011). Human resource management and productivity. In O. Ashenfelter & D. Card (Eds.), New developments and research on labor markets: Handbook of labor economics (Vol. 4b, pp. 1697-1767). Elsevier.

Bloom, N., Sadun, R., & Van Reenen, J. (2007). Americans do I.T. better: US multi-nationals and the productivity miracle. Working Paper 13085. National Bureau of Economic Research.

Blundell, R., & Dias, C. M. (2009). Alternative approaches to evaluation in empirical microeconomics. Journal of Human Resources, 44, 565-640.

Bommer, M., & Jalajas, D. (2002). The innovation work environment of high–tech SMEs in the USA and Canada. R&D Management, 32(5), 379-386.

Boudreau, M.-C. (2003). Learning to use ERP technology: A causal model. Proceed-ings of the 36th Annual Hawaii International Conference on System Sciences, IEEE, 1-10.

Bracht, G. H., & Glass, G. V. (1968). The external validity of experiments. American Educational Research Journal, 437-474.

Bresnahan, T. F., Brynjolfsson, E., & Hitt, L. M. (2002). Information technology, workplace organization, and the demand for skilled labor: Firm-level evidence. Quarterly Journal of Economics, 117, 339-376.

Brickley, J. A., Smith Jr, C. W., & Zimmerman, J. L. (2002). Business ethics and organizational architecture. Journal of Banking and Finance, 26(9), 1821-1835.

257

Brown, L. L. (2001). Review of the Kirton adaption-innovation inventory. In B. Plake & J. Impara (Eds.), The fourteenth mental measurements yearbook (pp. 647-649). University of Nebraska Press, Lincoln.

Brynjolfsson, E., & Hitt, L. (2003). Computing productivity: Firm-level evidence. Re-view of Economics and Statistics, 85, 793-808.

Brynjolfsson, E. (1993). The productivity paradox of information technology: Review and assignment. Communication of the ACM, 36(12), 67-77.

Brynjolfsson, E. (2003). ROI valuation: The IT productivity GAP. Optimize Maga-zine, (21).

Brynjolfsson, E., & Hitt, L. (1993). Is information systems spending productive?: New evidence and new results (pp. 47-64). MIT Sloan School of Management.

Brynjolfsson, E., & Hitt, L. (1996). Paradox lost? Firm-level Evidence on the Returns on Information Systems Spending. Management Science, 42(2), 541-558.

Brynjolfsson, E., & Milgrom, P. (2013). Complementarity in organizations. In R. Gib-bons & J. Roberts (Eds.), The handbook of organizational economics (pp. 11-55). Princeton University Press.

Brynjolfsson, E., Hitt, L. M., & Yang, S. (2002). Intangible assets: Computers and organizational capital. Brookings Papers on Economic Activity, 1, 137-198.

Buchanan, T. (2002). Online assessment: Desirable or dangerous? Professional Psy-chology: Research & Practice, 33, 148-154.

Bugamelli, M., & Pagano, P. (2004). Barriers to investment in ICT. Applied Econom-ics, 36, 2275-2286.

Bulkley, N., & Van Alstyne, M. (2007). Email, social networks and performance: An econometric case study. In L. Freeman (Ed.), Social network analysis. Sage Pub-lications, London.

Burt, R. S. (2004). Structural holes and good ideas. American Journal of Sociology, 110(2), 349-399.

Burt, R. S. (1992). Structural holes: The social structure of competition. University of Illinois at Urbana-Champaign’s Academy for Entrepreneurial Leadership His-torical Research Reference in Entrepreneurship.

Burton-Jones, A., & Straub Jr, D. W. (2006). Reconceptualizing system usage: An approach and empirical test. Information Systems Research, 17(3), 228-246.

Buser, T., & Peter, N. (2012). Multitasking. Experimental Economics, 1-15. Camerer, C. F., & Weber, R. A. (2013). Experimental Organizational Economics. In

R. Gibbons & J. Roberts (Eds.), The handbook of organizational economics (pp. 213-262). Princeton University Press.

Campbell, D. T., & Stanley, J. C. (1967). Experimental and quasi experimental de-signs for research. Chicago: Rand Mc Nally and Co.

Campion, M. A., & Thayer, P. W. (1985). Development and field evaluation of an interdisciplinary measure of job design. Journal of Applied Psychology, 70(1), 29.

Canfora, G., Cimitile, A., Garcia, F., Piattini, M., & Visaggio, C. A. (2007). Evaluat-ing performances of pair designing in industry. Journal of Systems and Software, 80(8), 1317-1327.

Cardwell, M., Clark, L., & Meldrum, C. (2001). Psychology for A2 level. HarperCol-lins.

Carland, J. C., Carland, J. W., Ensley, M. D., & Stewart, H. W. (1994). The implica-tions of cognition and learning styles for management education. Management Learning, 25(3), 413-431.

258

Caroli, E., & Van Reenen, J. (2001). Skill biased organizational change. Quarterly Journal of Economics, 116, 1448-1492.

Castells, M. (2011). The rise of the network society: The information age: Economy, society, and culture (Vol. 1). John Wiley & Sons.

Cequea, M. M., Monroy, C. R., & Bottini, M. A. N. (2011). The productivity from a human perspective: Dimensions and factors. Intangible Capital, 7(2), 549-584.

Chae, H-C., Koh, C. E., & Prybutok, R. (2014). Information technology capability and firm performance: Contradictory findings and their possible causes. MIS Quarterly, 38(1), 305-326.

Chari, M. D. R., Deavarj, S., & David, P. (2008). The impact of information technol-ogy investments and diversification strategies on firm performance. Management Science, 54(1), 224-234.

Chen, G., Kirkman, B. L., Kanfer, R., Allen, D., & Rosen, B. (2007). A multilevel study of leadership, empowerment, and performance in teams. Journal of Applied Psychology, 92(2), 331.

Chen, I. J., & Popovich, K. (2003). Understanding customer relationship management (CRM): People, process and technology. Business Process Management Journal, 9(5), 672-688.

Chesbrough, W. H. (2003). A better way to innovate. Harvard Business Review, 81(7), 12-13.

Chilton, M. A., Hardgrave, B. C., & Armstrong, D. J. (2005). Person-job cognitive style fit for software developers: The effect on strain and performance. Journal of Management Information Systems, 22(2), 193-226.

Chung, K., & Hossain, L. (2009). Measuring performance of knowledge-intensive workgroups through social networks. Project Management Journal, 40(2), 34-58.

Clements-Croome, D., & Baizhan, L. (2000). Productivity and indoor environment. Proceedings of Healthy Buildings. Helsinki, Vol. 1, 629-634.

Clements-Croome, D., & Kaluarachchi, Y. (2000). Assessment and measurement of productivity. Creating the Productive Workplace, 129-166.

Clemons, E. K., Reddi, S. P., & Row, M. C. (1993). The impact of information tech-nology on the organization of economic activity: The “move to the middle” hy-pothesis. Journal of Management Information Systems, 10(2), 9-35.

Clincy, V. A. (2003). Software development productivity and cycle time reduction. Journal of Computing Sciences in Colleges, 19(2), 278-287.

Coffield, F., Moseley, D., Hall, E., & Ecclestone, K. (2004). Learning styles and ped-agogy in post 16 learning: A systematic and critical review. London: Learning and Skills Research Centre.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159. Collins, H. (2003). Enterprise knowledge portals: next-generation portal solutions for

dynamic information access, better decision making, and maximum results. AMA-COM Div American Mgmt Assn.

Conger, J. A., & Kanungo, R. N. (1988). The empowerment process: Integrating the-ory and practice. Academy of Management Review, 13(3), 471-482.

Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design & analysis issues for field settings. Chicago: Rand McNally, Inc.

Corti, E., & Storto, C. L. (2000). The main factors to facilitate the creation of technical knowledge and competitiveness: A model to sustain the development of small firms. Innovation and Market Globalization: The Position of SME’s, 31, 101.

259

Coskun, E., & Grabowski, M. (2005). Impacts of user interface complexity on user acceptance and performance in safety-critical systems. Journal of Homeland Se-curity and Emergency Management, 2(1).

Couger, J. D., & Zawacki, R. A. (1980). Motivating and managing computer person-nel. New York, NY: Wiley.

Crano, W. D., Brewer, M. B., & Lac, A. (2014). Principles and methods of social research. Routledge.

Cross, R., Parker, A., & Sasson, L. (Eds.). (2003). Networks in the knowledge econ-omy. Oxford University Press.

Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. New York: Harper & Row.

Czerwinski, M., Horvitz, E., & Wilhite, S. (2004). A diary study of task switching and interruptions. Proceedings of the SIGCHI Conference On Human Factors in Com-puting Systems. ACM Press, 175-182.

D’Agostino, A., Serafini, R., & Ward-Warmedinger, M. (2006). Sectoral explanations of employment in Europe: The role of services. [Working paper № 625]. http://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp625.pdf. Date accessed 2012-01-05.

D’Ambra, J., & Rice R. E. (2001). Emerging factors in user evaluation of the World Wide Web. Information & Management, 38(6), 373-384.

Daft, R. L. (2000). Management. The Dryden Press, USA. Dandurand, F., Shultz, T. R., & Onishi, K. H. (2008). Comparing online and lab meth-

ods in a problem-solving experiment. Behavior Research Methods, 40(2), 428-434.

Danish, R. Q., & Usman, A. (2010). Impact of reward and recognition on job satis-faction and motivation: An empirical study from Pakistan. International Journal of Business and Management, 159-167.

Davenport, T. (2011). Powering information worker productivity with the right au-tomatization. [White paper]. http://www.idgconnect.com/download/6353/power-ing-information-worker-productivity-right-automation?source=concib-f052711c. Date accessed 2012-08-05.

Davies, H. (2005). Productivity and the knowledge worker. Proceedings of the Queensland University of Technology Research Week International Conference. Queensland University of Technology, 1-16.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-340.

Davis, S. F., & Buskist, W. (2008). 21st century psychology: A reference handbook, Vol. 1. Sage.

Day, D. V., Shleicher, D. J., Unckless, A. L., & Hiller, N. J. (2002). Self-monitoring personality at work: A meta-analytic investigation of construct validity. Journal of Applied Psychology, 87(2), 390.

de Barros Sampaio, S. C., Barros, E. A., de Aquino, G. S., & De Lemos Meira, S. R. (2010). A review of productivity factors and strategies on software development. Fifth International Conference on Software Engineering Advances (ICSEA), IEEE, 196-204.

de Koning, J., & Gelderblom, A. (2006). ICT and older workers: No unwrinkled rela-tionship. International Journal of Manpower, 27(5), 467-490.

Deci, E. L. (1975). Intrinsic motivation. New York: Plenum.

260

Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in hu-man behavior. New York: Plenum.

Deci, E. L., & Ryan, R. M. (1987). The support of autonomy and the control of be-havior. Journal of Personality and Social Psychology, 53(6), 1024.

Deci, E. L., & Vansteenkiste, M. (2004). Self-determination theory and basic need satisfaction: Understanding human development in positive psychology. Ricerche di Psichologia, 27, 17-34.

Deci, E. L., Connell, J. P., & Ryan, R. M. (1989). Self-determination in a work organ-ization. Journal of Applied Psychology, 74(4), 580.

Dedrick, D., Gurbaxani, K., & Kraemer, L. (2003). Information technology and eco-nomic performance: A critical review of the empirical evidence. ACM Computing Surveys, 35(1), 1-28.

Dedrick, J., Kraemer, K. L., & Shih, E. (2013). Information technology and produc-tivity in developed and developing countries. Journal of Management Information System, 30(1), 97-122.

Dehning, B., & Richardson, V. J. (2002). Returns on investments in information tech-nology: A research synthesis. Journal of Information Systems, 16(1), 7-30.

Delery, J. E., Gupta, N., & Shaw, J. D. (1997). Human resource management and firm performance: A systems perspective. In Southern Management Association Meet-ing, Atlanta, GA.

Delone, W. H., & McLean, E. R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research, 3(1), 60-95.

Delone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of infor-mation systems success: A ten-year update. Journal of Management Information Systems, 19(4), 9-30.

Deming, W. E. (1986). Out of the crisis. Cambridge, MA: Center for Advanced En-gineering Study.

Deng, X., Doll, W. J., & Cao, M. (2008). Exploring the absorptive capacity to inno-vation/productivity link for individual engineers engaged in IT enabled work. In-formation & Management, 45(2), 75-87.

Devaraj, S., & Kohli, R. (2003). Performance impacts of information technology: Is actual usage the missing link? Management Science, 49(3), 273-289.

Dewan, S., & Ren, F. (2011). Information technology and firm boundaries: Impact on firm risk and return performance. Information Systems Research, 22(2), 369-388.

Dewhurst, M., Hancock, B., & Ellsworth, D. (2013). Redesigning knowledge work. Harvard Business Review, 91(1), 58-64.

Dickson, G. W., Senn, J. A., & Chervany, N. L. (1977). Research in management information systems: The Minnesota experiments. Management Science, 23(9), 913-934.

Donaldson, L. (2000). Design structure to fit strategy. In E. Locke (Ed.), Handbook of principles of organizational behavior. Maiden, MA: Blackwell.

Dong, S., & Zhu, K. (2006). The business value of CRM systems: Productivity, prof-itability, and time lag. Proceedings of the Workshop on Information Systems and Economics (WISE 2006), Northwestern University, Evanston, IL.

Dong, S., Xu, S. X., & Zhu, K. H. (2009). Information technology in supply chains: The value of IT-enabled resources under competition. Information Systems Re-search, 20(1), 18-32.

Dority, G. K. (2006). Rethinking information work: A career guide for librarians and other information professionals. Libraries Unlimited.

261

Drucker, P. F. (1999). Management challenges of the 21st century. New York: Harper Business.

Duff, A. (2004). The role of cognitive learning styles in accounting education: Devel-oping learning competencies. Journal of Accounting Education, 22(1), 29-52.

Duncan, A. S. (1988). Software development productivity tools and metrics. Proceed-ings of the 10th International Conference on Software Engineering. IEEE, Com-puter Society Press, 41-48.

Dunnette, M. D. (1976). Aptitudes, abilities, and skills. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 473-520). Chicago: Rand McNally.

Dwivedi, Y. K., Wastell, D., Laumer, S., Henriksen, H. Z., Myers, M. D., Bunker, D., & Srivastava, S. C. (2015). Research on information systems failures and suc-cesses: Status update and future directions. Information Systems Frontiers, 17(1), 143-157.

Edberg, D. T., & Bowman, B. J. (1996). User-developed applications: An empirical study of application quality and developer productivity. Journal of Management Information Systems, 13(1), 167-185.

Edwards, J. R. (1991). Person-job fit: A conceptual integration, literature review and methodological critique. International Review of Industrial and Organizational Psychology, 6, 283-357.

Ee, J. (1998). Relationships among teachers’ classroom orientations, strategy-based instruction, and students’ goal orientations, self-regulated learning and achieve-ment. Unpublished doctoral dissertation, The University of Newcastle, Australia.

Ee, J., Seng, T. O., & Kwang, N. A. (2007). Styles of creativity: Adaptors and inno-vators in a Singapore context. Asia Pacific Education Review, 8(3), 364-373.

Eisenberger, R., & Byron, K. (2011). Rewards and creativity. In M. A. Runco & S. R. Pritzker (Eds.), Encyclopedia of creativity. New York, NY: Elsevier.

Emery, F. E., & Trist, E. L. (1969). Sociotechnical systems. In F. E. Emery (Ed.), Systems thinking (pp. 281-296). London: Penguin Books.

Ennen, E., & Richter, A. (2010). The whole is more than the sum of its parts – or is it? A review of the empirical literature on complementarities in organizations. Journal of Management, 36(1), 207-233.

Erasmus, B., & Schenk, H. (2008). South African human resource management: The-ory and practice. Juta and Company Ltd.

Eraut, M. (2000). Non-formal learning and tacit knowledge in professional work. Brit-ish Journal of Educational Psychology, 70, 113-136.

Faraj, S., & Azad, B. (2012). The materiality of technology: An affordance perspec-tive. Materiality and Organizing: Social Interaction in a Technological World, 237-258.

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Re-search Methods, 41, 1149-1160.

Field, A. (2011). Discovering statistics using SPSS. Thousand Oaks, California: SAGE Publications.

Freeman, L. (1979). Centrality in social networks: Conceptual clarification. Social Networks, 1, 215-239.

Frey, B. S., & Osterloh, M. (2002). Successful management by motivation: Balancing intrinsic and extrinsic incentives. Berlin, Heidelberg, New York: Springer.

262

Frick, D. E., & Drucker, P. F. (2011). Motivating the knowledge worker. Defense Acquisition Research Journal: A Publication of the Defense Acquisition Univer-sity, 18(4), 368-387.

Fried, Y., & Ferris, G. R. (1987). The validity of the job characteristics model: A review and meta-analysis. Personnel Psychology, 40, 287-322.

Fuller, L. R., & Kaplan, S. E. (2004). A note about the effect of auditor cognitive style on task performance. Behavioral Research in Accounting, 6, 131-143.

Gagne, R. M., & Medsker, K. L. (1996). The conditions of learning. Fort Worth, TX: Harcourt-Brace.

Galbraith, J. R. (1970). Environmental and technological determinants of organiza-tional design: A case study. In J. W. Lorsch & P. R. Lawrence (Eds.), Studies in organizational design (pp. 113-139). Homewood, Il.: Richard Irwin.

Gallivan, M. J. (2003). The influence of software developers’ creative style on their attitudes to and assimilation of a software process innovation. Information and Management, 40(5), 443-465.

Garg, P., & Rastogi, R. (2006). New model of job design: Motivating employees’ performance. Journal of Management Development, 25(6), 572-587.

Goodhue, D. L., & Thompson. R. L. (1995). Task technology fit and individual per-formance. MIS Quarterly, 19(2), 213-236.

Gough, H. G. (1979). A creative personality scale for the adjective check list. Journal of Personality and Social Psychology, 37, 1398-1405.

Gray, C. D., & Kinnear, P. R. (2012). IBM SPSS statistics 19 made simple. Psychology Press.

Grgecic, D., & Rosenkranz, C. (2010). Rethinking the concept of IT use. Paper pre-sented at the Diffusion Interest Group in Information Technology, St. Louis, MI.

Gupta, B. (2009). Understanding the preferences of creative and non-creative employ-ees. Indian Journal of Industrial Relations, 45(2), 289-301.

Gurbaxani, V., Melville, N., & Kraemer, K. (2000). The production of information services: A firm-level analysis of information systems budgets. Information Sys-tems Research, 11(2), 159-176.

Gyampah, A., & Salam, A. F. (2004). An extension of the technology acceptance model in an ERP implementation environment. Information & Management, 41(6), 731-745.

Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250-279.

Hackman, J. R., & Lawler, E. E. (1971). Employee reactions to job characteristics. Journal of Applied Psychology Monograph, 55(3), 259-286.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley.

Halawi, L. A., McCarthy, R. V., & Aronson, J. E. (2007). An empirical investigation of knowledge-management systems’ success. The Journal of Computer Infor-mation Systems, 48(2), 121-135.

Hammer, M., Leonard, D., & Davenport, T. (2004). Why don’t we know more about knowledge?. MIT Sloan Management Review, 45(4), 14-18.

Han, K., Chang, Y. B., & Hahn, J. (2011). Information technology spillover and productivity: The role of information technology intensity and competition. Jour-nal of Information Technology, 23(3), 185-202.

263

Haner, U. E., Kelter, J., Bauer, W., & Rief, S. (2009). Increasing information worker productivity through information work infrastructure. Proceeding of the Interna-tional Conference on Ergonomics and Health Aspects of Work with Computers. Springer Berlin Heidelberg, 39-48.

Harrison, A. W., & Rainer Jr, R. K. (1992). The influence of individual differences on skill in end-user computing. Journal of Management Information Systems, 9(1), 93-111.

Harrison, G. W. (2011). Randomization and its discontents. Journal of African Econ-omies, 20(4), 626-652.

Harter, D. E., Krishnan, M. S., & Slaughter, S. A. (2000). Effects of process maturity on quality, cycle time, and effort in software product development. Management Science, 46(4), 451-466.

Harter, S. (1978). Effectance motivation reconsidered. Toward a developmental model. Human Development, 21(1), 661-669.

Hassett, M. E., & Paavilainen-Msntymski, E. (Eds.). (2013). Handbook of longitudi-nal research methods in organisation and business studies. Edward Elgar Publish-ing.

Hayes, J., & Allinson, C. W. (1997). Learning styles and training and development in work settings: Lessons from educational research. Educational Psychology, 17(1-2), 185-193.

Hayes, J., & Allinson, C. W. (1998). Cognitive style and the theory and practice of individual and collective learning in organizations. Human Relations, 51, 847-871.

Helms-Mills, J., Dye, K., & Mills, A. J. (2008). Understanding organizational change. Routledge.

Hennessey, B. A. (1989). The effect of extrinsic constraints on children’s creativity while using a computer. Creativity Research Journal, 2(3), 151-168.

Hennessey, B. A., & Amabile, T. M. (2010). Creativity. Annual Review of Psychol-ogy, 61(1), 569-598.

Hennessey, B. A., & Zbikowski, S. M. (1993). Immunizing children against the neg-ative effects of reward: A further examination of intrinsic motivation training tech-niques. Creativity Research Journal, 6(3), 297-307.

Herzberg, F. Mausner, B., & Snyderman, B. (1959). The motivation to work. New York: John Wiley.

Hirst, G., Knippenberg, D. V., Chen, C-H., & Sacramento, C. A. (2011). How does bureaucracy impact individual creativity? A cross-level investigation of team con-textual influences on goal orientation-creativity relationships. Academy of Man-agement Journal, 54(3), 624-641.

Hitt, L. M., & Brynjolfsson, E. (1997). Information technology and internal firm or-ganization: An exploratory study. Journal of Information Systems, 14(2), 81-101.

Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institu-tions and organizations across nations. Thousand Oaks, Calif.: Sage.

Honey, P., & Mumford, A. (1982). The manual of learning styles. Maidenhead: McGraw-Hill.

Hopp, W. J., Iravani, S. M., & Liu, F. (2009). Managing white-collar work: An oper-ations-oriented survey. Production and Operations Management, 18(1), 1-32.

Horwitz, F., Heng, C. T., & Quazi, H. A. (2003). Finders, keepers? Attracting, moti-vating and retaining knowledge workers. Human Resource Management Journal, 13(4), 23-44.

264

Hou, C-H. (2012). Examining the effect of user satisfaction on system usage and in-dividual performance with business intelligence systems: An empirical study of Taiwan’s electronics industry. International Journal of Information Management, 32(6), 560-573.

Howell, W. C., & Cooke, N. J. (1989). Training the human information processor: A review of cognitive models. In I. L. Goldstein (Ed.), Training and development in organizations (pp. 121-182). San Francisco: Jossey-Bass.

Hu, Q., & Plant, R. (2001). An empirical study of the casual relationship between IT investment and firm performance. Information Resources Management Journal, 14(3), 15-26.

Hu, Q., & Quan, J. (2005). Evaluating the impact of IT investments on productivity: A causal analysis at industry level. International Journal of Information Manage-ment, 25, 39-53.

Huber, G. P. (1983). Cognitive style as a basis for MIS and DSS designs: Much ado about nothing?. Management Science, 29(5), 567-579.

Hughes, A., & Scott Morton, M. S. (2005). ICT and productivity growth – the paradox resolved? Working Paper No. 316. Centre for Business Research, University of Cambridge.

Humphrey, S. E., Nahrgang, J. D., & Morgeson, F. P. (2007). Integrating motiva-tional, social, and contextual work design features: A meta-analytic summary and theoretical extension of the work design literature. Journal of Applied Psychology, 92(5), 1332.

Hunt, R. G., Krzystofiak, F. J., Meindl, J. R., & Yousry, A. M. (1989). Cognitive style and decision making. Organizational Behavior and Human Decision Processes, 44(3), 436-453.

Hunter, G. K., & Perreault, W. D. J. (2006). Sales technology orientation, information effectiveness, and sales performance. Journal of Personal Selling and Sales Man-agement, 26(2), 95-113.

Hunter, J. E. (1986). Cognitive ability, cognitive aptitudes, job knowledge, and job performance. Journal of Vocational Behavior, 29(3), 340-362.

Huselid, M. A. (1995). The impact of human resource management practices on turn-over, productivity, and corporate financial performance, The Academy of Manage-ment Journal, 38(3), 635-672.

Ichniowski, C., & Shaw, K. (2003). Beyond incentive pay: Insiders’ estimates of the value of complementary human resource management practices. Journal of Eco-nomic Perspectives, XVII, 155-180.

Ichniowski, C., Shaw, K., & Prennushi, G. (1997). The effect of human resource man-agement practices on productivity: A study of steel finishing lines. American Eco-nomic Review, 87, 291-313.

Iivari, J. (2005). An empirical test of DeLone-McLean model of information systems success. The DATA BASE for Advances in Information Systems, 36(2), 8-27.

International Labor Organization (ILO). (2008). Labor Statistics. http://la-borsta.ilo.org/. Date accessed 2012-02-12.

Iqbal, S. T., & Horvitz, E. (2007). Conversations admits computing: A study of inter-ruptions and recovery of task activity. Proceedings of the 11th International Con-ference on User Modeling, Corfu, Greece, 2007, Berlin: Springer Heiselberg, 350-354.

Isaksen, S. G., Dorval, K. B., & Kaufmann, G. (1992). Mode of symbolic representa-tion and cognitive style. Imagination, Cognition and Personality, 11(3), 271-277.

265

Jablokow, K. W., & Booth, D. E. (2006). The impact and management of cognitive gap in high performance produce development organizations. Journal of Engi-neering and Technology Management, 23, 313-336.

Jain, V., & Kanungo, S. (2005). Beyond perceptions and usage: Impact of nature of information systems use on information system-enabled productivity. Interna-tional Journal of Human-Computer Interaction, 19(1), 113-136.

Jain, V., & Kanungo, S. (2013). Realizing IT value: Post adoptive IS usage and per-formance impacts at individual level. International Journal of Business Infor-mation Systems, 14(2), 202-222.

Jefferys, J. B. (1954). Productivity in the distributive trade in Europe. Paris: Office of European Economic Cooperation.

Jelinek, R., Ahearne, M., Mathieu, J., & Schillewaert, N. (2006). A longitudinal ex-amination of individual, organizational, and contextual factors on sales technology adoption and job performance. Journal of Marketing Theory and Practice, 14(1), 7-23.

Jez, V. (2011). Searching for the meaning of multitasking. NOKOBIT 2011, pp. 157-166.

Jorgenson, D. W. (2007). Information technology and the G7 economies. In U. Apte & U. Karmarkar (Eds.), Managing in the information economy (Vol. 1, pp. 29-57). Springer Science + Business Media, LLC.

Judd, C. M., Smith, E. R., & Kidder, L. H. (1991). Research methods in social rela-tions. New York, NY: Holt, Rinehart, and Winston.

Juran, J. M. (1974). Quality control handbook. New York: McGraw-Hill. Karasek Jr, R. A. (1979). Job demands, job decision latitude, and mental strain: Im-

plications for job redesign. Administrative Science Quarterly, 285-308. Karmarkar, U. S., & Apte, U. M. (2007). Current research on managing in the infor-

mation economy. In U. Apte & U. Karmarkar (Eds.), Managing in the information economy (Vol. 1, pp. vii-x). Springer Science+Business Media, LLC.

Kasof, J., Chen, C., Himsel, A., & Greenberger, E. (2007). Values and creativity. Cre-ativity Research Journal, 19(2-3), 105-122.

Keen, P. G., & Scott-Morton, M. (1978). Decision support systems: An organizational perspective. Reading, MA: Addison-Wesley.

Kelly, J. E. (1992). Does job re-design theory explain job re-design outcomes? Human Relations, 45, 753-774.

Kenyon, S., & Lyons, G. (2007). Introducing multitasking to the study of travel and ICT: Examining its extent and assessing its potential importance. Transportation Research, 41, 161-175.

Keppel, G. (1991). Design and analysis: A researcher’s handbook. Prentice-Hall, Inc. Kessels, J. W. (2001). Learning in organizations: A corporate curriculum for the

knowledge economy. Futures, 33(6), 497-506. Kirton, M. (1976). Adaptors and innovators: A description and measure. Journal of

Applied Psychology, 61(5), 622-629. Kirton, M. J. (1994). Adaptors and innovators. Routledge, London. Kirton, M. J. (2003). Adaption-innovation: In the context of diversity and change.

New York, NY: Routledge. Kirton, M. J., Bailey, A., & Glendinning, W. (1991). Adaptors and innovators: Pref-

erence for educational procedures. The Journal of Psychology, 125(4), 445-455.

266

Knight, M. B., & Pearson, J. M. (2005). The changing demographics: The diminishing role of age and gender in computer usage. Journal of Organizational and End User Computing (JOEUC), 17(4), 49-65.

Kock, N. F. Jr., & McQueen, R. J. (1996). Product flow, breadth and complexity of business processes: An empirical study of 15 business processes in three organi-zations. Business Process Re-engineering & Management Journal, 2(2), 8-22.

Kohli, R., & Grover, V. (2008). Business value of IT: An essay on expanding research directions to keep up with the times. Journal of the AIS, 9(1), 23-37.

Kositanurit, B., Ngwenyama, O., & Osei-Bryson, K. (2006). An exploration of factors that impact individual performance in an ERP environment: An analysis using multiple analytical techniques. European Journal of Information Systems, 15(6), 556-568.

Kossowska, M., Orehek, E., & Kruglanski, A. W. (2010). Motivation towards closure and cognitive resources: An individual differences approach. In A. Gruzka, G. Mathews, & B. Szymura (Eds.), Handbook of individual differences in cognition: Attention, money and executive control (pp. 369-382). New York: Springer Sci-ence + Business Media.

Kozlowski, S. W. J., & Bell, B. S. (2003). Work groups and teams in organizations. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Comprehensive handbook of psychology (Vol. 12, pp. 333-375). New York: Wiley.

Kraemer, K. L., & Danziger, J. N. (1990). The impacts of computer technology on the worklife of information workers. Social Science Computer Review, 8, 592-613.

Kraiger, K. (2003). Perspectives on training and development. In W. C. Borman, D. R. Ilgen & R. J. Klimoski (Eds.), Handbook of psychology (pp. 171-192). Hobo-ken, NJ: Wiley.

Kubo, I., & Saka, A. (2002). An inquiry into the motivation of knowledge workers in the Japanese financial industry. Journal of Knowledge Management, 6, 262-271.

Kvassov, V. (2004). The effects of time personality on the productivity of MIS. Jour-nal of Decision Systems, 1, 9-25.

Lal, K. (2005). In quest of the information sector: Measuring information workers for India. Malaysian Journal of Library and Information Science, 10(2), 85-104.

Laschinger, H. K. S., & Wong, C. (1999). Staff nurse empowerment and collective accountability: Effect on perceived productivity and self-rated work effectiveness. Nursing Economics, 17(6), 308.

Latham, G. P., & Locke, E. A. (1979). Goal setting – A motivational technique that works. Organizational Dynamics, 8(2), 68-80.

Lawrence, P. R., & Lorsch, J. W. (1967). Organization and environment: Managing differentiation and integration. Cambridge, MA: Harvard Graduate School of Business.

Lee, C. S. (2001). Modelling the business value of information technology. Infor-mation and Management, 39, 191-210.

Lee, S.-Y. T., Gholami, R., & Tong, T. Y. (2005). Times series analysis in the assess-ment of ICT impact at the aggregate level – Lessons and implications for the new economy. Information & Management, 42(7), 1009-1022.

Leontiev, A. N. (1978). Activity, consciousness, and personality. Upper Saddle River, NJ: Prentice Hall.

267

Lim, J. H., Richardson, V. J., & Roberts, T. L. (2004). Information technology invest-ment and firm performance: A meta-analysis. Proceedings of the 37th Hawaii In-ternational Conference on System Sciences, IEEE Computer Society, Big Island, HI, USA, 10.

Lipsey, M. W. (1990). Design sensitivity: Statistical power for experimental research, Vol. 19. Sage.

Locke, E. A., & Lathman, G. P. (2004). What should we do about motivation theory? Six recommendations for the twenty-first century. The Academy of Management Review, 29(3), 388-403.

Loher, B. T., Noe, R. A., Moeller, N. L., & Fitzpatrick, M. P. (1985). A meta-analysis of the relation of job characteristics to job satisfaction. Journal of Applied Psy-chology, 70(2), 280-289.

Lokan, C. J. (2001). Impact of subjective factors on software productivity. Proceed-ings of 7th Australian Conference on Software Metrics, Melbourne, 29-30.

Lord, R. G., & Maher, K. J. (1991). Cognitive theory in industrial/organizational psy-chology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, pp. 1-62). Palo Alto, CA: Consulting Psy-chologists Press.

Lord, R. L. (2002). Traditional motivation theories and older engineers. Engineering Management Journal, 14, 3-7.

Love, P. E. D., & Edwards, D. J. (2005). Taking the pulse of UK construction project managers’ health: Influence of job demands, job control and social support on psy-chological well-being. Engineering, Construction, and Architectural Manage-ment, 12(1), 88-101.

Loveman, G. W. (1994). An assessment of the productivity impact of information technologies. Information technology and the corporation of the 1990s: Research studies, 84-110.

Lucas, H. C., & Spitler, V. K. (1999). Technology use and performance: A field study of Broker Workstations. Decision Sciences, 30(2), 291-311.

Lund, S., Manyika, J., & Ramaswamy, S. (2012). Preparing for a new era of knowledge work. McKinsey Quarterly, 4, 103-110.

Luthans, F. (1973). The contingency theory of management: A path out of the jungle. Business Horizons, 16, 62-72.

Lynch, L. M., & Black, S. E. (1998). Beyond the incidence of training: Evidence from a national employer’s survey. Industrial and Labor Relations Review, 52(1), 64-81.

MacCormack, A., Verganti, R., & Iansiti, M. (2001). Developing products on “Inter-net time”: The autonomy of a flexible development process. Management Science, 47(1), 133-150.

Madjar, N., & Shalley, C. E. (2008). Multiple tasks’ and multiple goals’ effect on creativity: Forced incubation or just a distraction? Journal of Management, 34(4), 786-805.

Malone, T. W. (1997). Is empowerment just a fad? Control, decision making, and IT. MIT Sloan Management Review, 38(2), 23.

Mani, B. G. (2002). Performance appraisal systems, productivity, and motivation: A case study. Public Personnel Management, 31, 141-159.

Marchand, D. A., Kettinger, W. J., & Rollins, J. D. (2002). Information orientation: The link to business performance. Oxford University Press.

268

Markova, G., & Ford, C. (2011). Is money the panacea? Rewards for knowledge workers. International Journal of Productivity and Performance Management, 60(8), 813-823.

Markus, M. L., Majchrzak, A., & Gasser, L. (2002). A design theory for systems that support emergent knowledge processes. MIS Quarterly, 26(3), 179-212.

Marulanda-Carter, L., & Jackson, T. W. (2012). Effects of email addiction and inter-ruptions on employees. Journal of Systems and Information Technology, 14(1), 82-94.

Mason, R. O., & Mitroff, I. I. (1973). A program for research on management infor-mation systems. Management Science, 19(5), 475-487.

Mathieu, J., Ahearne, M., & Taylor, S. R. (2007). Longitudinal cross-level model of leader and salesperson influences on sales force technology use and performance. Journal of Applied Psychology, 2, 528-537.

Mayers, A. (2013). Introduction to statistics and SPSS in psychology. Pearson Educa-tion Limited.

McClelland, D. C. (1967). The achieving society. New York press: The Free Press. McCullough, J. S., Parente, S., & Town, R. (2013). Health information technology

and patient outcomes: The role of organizational and informational complementa-rities. Working Paper 18684. National Bureau of Economic Research.

McGill, T. J., & Klobas, J. E. (2005). The role of spreadsheet knowledge in user-developed application success. Decision Support Systems, 39(3), 355-369.

McGill, T., Hobbs, V., & Klobas, J. (2003). User-developed applications and infor-mation systems success: A test of DeLone and McLean’s model. Information Re-sources Management Journal, 16(1), 24-45.

McGrath, J. E. (1981). Dilemmatics: The study of research choices and dilemmas. American Behavioral Scientist, 25(2), 179-210.

McLeod, Jr. A. J., Clark, J. W., & Dietrich, G. (2008). The impact of information systems on end user performance: Examining the effects of cognitive style using learning curves in an electronic medical record implementation. Communications of the Association for Information Systems, 22, 165-184.

McNeese-Smith, D. (1996). Increasing employee productivity, job satisfaction, and organizational commitment. Journal of Healthcare Management, 41(2), 160.

Melville, N., Gurbaxani, V., & Kraemer, K. (2007). The productivity impact of infor-mation technology across competitive regimes: The role of industry concentration and dynamism. Decision Support Systems, 43(1), 229-242.

Melville, N., Kraemer, K., & Gurbahani, V. (2004). Information technology and or-ganizational performance: An integrative model of IT business value. MIS Quar-terly, 28(2), 283-322.

Milgrom, P., & Roberts, J. (1990). The economics of modern manufacturing: Tech-nology, strategy, and organization. American Economic Review, 80(3), 511-528.

Milgrom, P., & Roberts, J. (1995). Complementarities and fit: Strategy, structure, and organizational change in manufacturing. Journal of Accounting and Economics, 19, 179-208.

Miller, J., Daly, J., Wood, M., Roper, M., & Brooks, A. (1997). Statistical power and its subcomponents – missing and misunderstood concepts in empirical software engineering research. Information and Software Technology, 39(4), 285-295.

Miller, W. C. (1986). Employee creativity: Personal and contextual factors at work. The Academy of Management Journal, 39(3), 607-634.

269

Mitchell, V. L., & Zmud, R. W. (1999). The effects of coupling IT and work process strategies in redesign projects. Organization Science, 10(4), 424-438.

Mithas, S., Ramasubbu, N., & Sambamurthy, V. (2011). How information manage-ment capability influences firm performance. MIS Quarterly, 35(1), 237-256.

Morgeson, F. P., & Humphrey, S. E. (2006). The work design questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91, 1321-1339.

Mukhopadhyay, T., Rajiv, S., & Srinivasan, K. (1997). Information technology im-pact on the productivity and quality of mail sorting. Management Science, 43(12), 1645-1659.

Mullins, L. J. (2007). Management and organizational behavior. Pearson education. Myers, I. B., McCaulley, M. H., Quenk, N. L., & Hammer, A. L. (1998). MBTI man-

ual: A guide to the development and use of the Myers-Briggs type indicator. Con-sulting Psychologists Press, Palo Alto, CA.

North, K., & Gueldenberg, S. (2011). Effective knowledge work: Answers to the man-agement challenges of the 21st century. Emerald Group Publishing.

Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contex-tual factors at work. Academy of Management Journal, 39(3), 607-634.

Onwuegbuzie, A. J. (2003). Expanding the framework of internal and external validity in quantitative research. Research in the Schools, 10, 71-90.

Orlikowski, W. J., & Robey, D. (1991). Information technology and the structuring of organizations. Information Systems Research, 2(2), 143-169.

Oyana, T. J., & Margai, F. (2015). Spatial analysis: Statistics, visualization, and com-putational methods. CRC Press.

Ozaralli, N. (2003). Effects of transformational leadership on empowerment and team effectiveness. Leadership & Organization Development Journal, 24(6), 335-344.

Parker, S. K., & Wall, T. D. (1998). Job and work design: Organizing work to promote well-being and effectiveness. San Francisco, CA: Sage.

Parker, S. K., Wall, T. D., & Cordery, J. (2001). Future work design research and practice: Towards an elaborated model of work design. Journal of Occupational and Organizational Psychology, 74, 413-440.

Parmigiani, A., & Mitchell, W. (2009). Complementarity, capabilities, and the bound-aries of the firm: The impact of within-firm and interfirm expertise on concurrent sourcing of complementary components. Strategic Management Journal, 30(10), 1065-1091.

Pashkevich, N., & Haftor, D. (2014). IT-driven productivity at the individual level: Complementarities matter. Proceedings of the 8th European Conference on IS Management and Evaluation. Academic Conferences and Publishing Interna-tional, 380-383.

Pashkevich, N., & Haftor, D. (2016). Complementarities of effective IT-use at indi-vidual level: Preliminary results. Forthcoming in Proceedings of European Con-ference on Information Systems (ECIS 2016), Istanbul, Turkey.

Petroni, A., & Colacino, P. (2008). Motivation strategies for knowledge workers: Ev-idences and challenges. Journal of Technology, Management and Innovation, 3(3), 21-32.

Petter, S., DeLone, W., & McLean, E. (2008). Measuring information systems suc-cess: Models, dimensions, measures, and interrelationships. European Journal of Information Systems, 17, 236-263.

270

Pinder, C. C. (1984). Work motivation: Theory, issues, and applications. Scott For-esman & Co.

Pinder, W. C. C. (2011). Work motivation in organizational behavior. New York, NY: Psychology Press.

Pinsonneault, A., & Rivard, S. (1998). Information technology and the nature of man-agerial work: From the productivity paradox to the Icarus paradox?. MIS Quar-terly, 287-311.

Pisano, G. P. (1994). Knowledge, integration, and the locus of learning: An empirical analysis of process-development. Strategic Management Journal, 15, 85-100.�

Pisano, G. P. (1996). Learning-before-doing in the development of new process tech-nology. Research Policy, 25(7), 1097-1119.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and rec-ommended remedies. Journal of Applied Psychology, 88(5), 879.

Pollalis, Y. A. (2003). Patterns of co-alignment in information-intensive organiza-tions: Business performance through integration strategies. International Journal of Information Management, 23, 469-492.

Poon, S. K., Davis, J. G., & Choi, B. (2009). Augmenting productivity analysis with data mining: An application on IT business value. Expert Systems with Applica-tions, 36(2), 2213-2224.

Porat, M. U. (1977). The information economy: Definition and measurement. Office of Telecommunication (DOC), Washington, D.C.

Porter, M. E. (1990). The competitive advantage of notions. Harvard Business Re-view, 68(2), 73-93.

Porter, M., & Siggelkow, N. (2008). Contextuality within activity systems and sus-tainability of competitive advantage. The Academy of Management Perspectives, 22(2), 34-56.

Pounds, J., & Bailey, L. L. (2001). Cognitive style and learning: Performance of adap-tors and innovators in a novel dynamic task. Applied Cognitive Psychology, 15, 547-563.

Prakash, Y., & Gupta, M. (2008). Exploring the relationship between organization structure and perceived innovation in the manufacturing sector of India. Singapore Management Review, 30(1), 55-76.

Pyöria, P. (2005). The concept of knowledge work revisited. Journal of Knowledge Management, 9(3), 116-127.

Ramaprasad, A. (1987). Cognitive process as a basis for MIS and DSS design. Man-agement Science, 33(2), 139-148.

Ramirez, Y. W., & Nembhard, D. A. (2004). Measuring knowledge worker produc-tivity: A taxonomy. Journal of Intellectual Capital, 5(4), 602-628.

Randolph, K. A., & Myers, L. L. (2013). Basic statistics in multivariate analysis. Ox-ford University Press.

Reinhardt, W., Schmidt, B., Sloep, P., & Drachsler, H. (2011). Knowledge worker roles and actions – results of two empirical studies. Knowledge and Process Man-agement, 18(3), 150-174.

Reips, U.-D. (2002). Standards for Internet-based experimenting. Experimental Psy-chology, 49, 243-256.

Riding, R., & Douglas, G. (1993). The effect of cognitive style and mode of presen-tation on learning performance. British Journal of Educational Psychology, 63(2), 297-307.

271

Rigby, D. K., Sutherland, J., & Takeuchi, H. (2016). Embracing agile. Harvard Busi-ness Review, 94(5), 40-50.

Roberts, N., Galluch, P. S., Dinger, M., & Grover, V. (2012). Absorptive capacity and information systems research: Review, synthesis, and directions for future re-search. MIS Quarterly, 36(2), 625-648.

Roberts, J. (2007). The modern firm: Organizational design for performance and growth. Oxford University Press.

Robey, D. (1983). Cognitive style and DSS design: A comment on Huber’s paper. Management Science, 29(5), 580-582.

Robey, D., & Boudreau, M. C. (1999). Accounting for the contradictory organizational consequences of information technology: Theoretical directions and methodological implications. Information Systems Research, 10(2), 167-185.

Robson, C. (1993). Real world research: A resource for social scientists and practi-tioner researchers. Oxford: BlackwelL

Rose, R. C., Kumar, N., & Pak, O. G. (2011). The effect of organizational learning on organizational commitment, job satisfaction and work performance. Journal of Applied Business Research, 25(6), 55.

Rowe, A. J., & Boulgarides, J. D. (1992). Managerial decision-making: A guide to successful business decision, New York: Macmillan.

Rowe, A. J., & Mason, R. O. (1987). Managing with style. San Francisco: Jossey-Bass.

Ruch, M. A. (1994). Measuring and managing individual productivity. In D. H. Harris (Ed.), Organizational linkages understanding the productivity paradox (pp. 105-130). National Academy Press, Washington (DC).

Ryan, R., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54-67.

Ryu, C., Y. J. Kim, A. Chaudhury, H., & Rao, R. (2005). Knowledge acquisition via three learning processes in enterprise information portals: Learning-by-invest-ment, learning-by-doing, and learning-from-others. MIS Quarterly, 29(2), 245-278.

Sadler-Smith, E., & Badger, B. (1998). Cognitive style, learning and innovation. Technology Analysis & Strategic Management, 10(2), 247-266.

Sadler-Smith, E., & Smith, P. J. (2004). Strategies for accommodating individuals’ styles and preferences in flexible learning programs. British Journal of Educa-tional Technology, 35, 395-412.

Sadler, K., Robertson, T., & Kan, M. (2009). Exploring the project transitions and everyday mobile practices of freelancers: Emergent concepts from empirical stud-ies of practice. Proceedings of the 21st Annual Conference of the Australian Com-puter-Human Interaction Special Interest Group: Design: Open 24/7, ACM, 81-88.

Sagiv, L., Arieli, S., Goldenberg, J., & Goldschmidt, A. (2010). Structure and freedom in creativity: The interplay between externally imposed structure and personal cognitive style. Journal of Organizational Behavior, 31(8), 1086-1110.

Sahinidis, A. G., & Bouris, J. (2008). Employee perceived training effectiveness re-lationship to employee attitudes. Journal of European Industrial Training, 32(1), 63-76.

Sandhi, I. K. (2013). The development of productivity performance models: Based on self-efficacy, trust, system quality, and information quality. Study of information

272

systems of PT Pindo Deli Paper Products. Proceedings of the First International Conference on Law, Business and Government, UBL, Indonesia, 187-195.

Satterfield, J. M., & Hughes, E. (2007). Emotion skills training for medical students: A systematic review. Medical Education, 41(10), 935-941.

Schank, R. C., & Joseph, D. M. (1998). Intelligent schooling. In R. J. Sternberg & W. M. Williams (Eds.), Intelligence, instruction, and assessment: Theory into prac-tice (pp. 43-65). Mahwah, NJ: Erlbaum.

Schienstock, G., Bechmann, G., & Frederichs, G. (1999). Information society, work and the generation of new forms of social exclusion. SOWING). [First Interim Re-port (Literature Review)]. http://www.uta.fi/laitokset/tyoelama/sowing/re-port/SowingFinal2002.pdf. Date accessed 2012-01-05.

Schneider, B., & Smith, D. B. (2004). Personality and organizational culture. In B. Scheneider & D. B. Smith (Eds.), Personality and organization (pp. 347-369). Mahwah, New Jersey: Lawrence Erlbaum Associates.

Schonenberg, H., Mans, R., Russell, N., Mulyar, N., & van der Aalst, W. (2008). Pro-cess flexibility: A survey of contemporary approaches. In: J. L. G. Dietz & A. Albani (Eds.), Advances in enterprise engineering (pp. 16-30). Springer-Verlag Berlin Heidelberg.

Schryen, G. (2013). Revisiting IS business value research: What we already know, what we still need to know, and how we can get there. European Journal of Infor-mation System, 22, 139-169.

Sense, A. (2007). Structuring the project environment for learning. International Journal of Project Management, 25(4), 405-412.

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston: Houghton-Mifflin.

Shalley, C. E., & Perry-Smith, J. E. (2001). Effects of social-psychological factors on creative performance: The role of informational and controlling expected evalua-tion and modeling experience. Organizational Behavior and Human Decision Processes, 84(1), 1-22.

Shalley, C. E., Zhou, J., & Oldham, G. R. (2004). The effects of personal and contex-tual characteristics on creativity: Where should we go from here?. Journal of Man-agement, 30(6), 933-958.

Shao, Z., Feng, Y., & Hu, Q. (2016). Effectiveness of top management support in enterprise systems success: A contingency perspective of fit between leadership style and system life-cycle. European Journal of Information Systems, 25(2), 131-153.

Sharit, J., Czaja, S. J., Hernandez, M., Yang, Y., Perdomo, D., Lewis, J. E., & Nair, S. (2004). An evaluation of performance by older persons on a simulated telecom-muting task. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 59(6), 305-316.

Simon, H. A., & Newell, A. (1964). Information processing in computer and man. American Scientist, 52(3), 281-300.

Singh, R., & Mohanty, M. (2012). Impact of training practices on employee produc-tivity: A comparative study. Interscience Management Review (IMR), 2(2), 74.

Sisselman, M. E., & Whitt, W. (2007). Value-based routing and preference-based routing in customer contact centers. Production and Operations Management, 16(3), 277-291.

Skinner, B. F. (1976). About behaviorism. Vintage Books, N.Y.

273

Skinner, N. F., & Drake, J. M. (2003). Behavioral implications of adaption-innova-tion: Adaption-innovation, achievement motivation, and academic performance. Social Behavior and Personality, 31(1), 101-106.

Smith, L. F., Gratz, Z. S., & Bousquet, S. G. (2008). The art and practice of statistics. Cengage Learning.

Smither, R. D. (1998). The psychology of work and human performance. New York: Longman.

Snape, E., & Redman, T. (2010). HRM practices, organizational citizenship behavior, and performance: A multi-level analysis. Journal of Management Studies, 47(7), 1219-1247.

Sokoya, S. K. (2000). Personal predictors of job satisfaction for the public sector man-ager: Implications for management practice and development in a developing economy. Journal of Business in Developing Nations, 4(1), 40-53.

Solow, R. (1987). We’d better watch out. New York Times Book Review, July 12, 36. Sonnentag, S. (Ed.). (2003). Psychological management of individual performance.

John Wiley & Sons. Sonnentag, S., & Volmer, J. (2009). Individual-level predictors of task-related team-

work processes: The role of expertise and self-efficacy in team meetings. Group & Organization Management, 34(1), 37-66.

Stevens, J. P. (2012). Applied multivariate statistics for the social sciences. Routledge. Strassmann, P. A. (1990). The business value of computers: An executive’s guide. In-

formation Economics Press, New Canaan, Conn. Sun, H., & Fricke, M. (2009). Reexamining the impact of system use on job perfor-

mance from the perspective of adaptive system use. Proceedings of the AMCIS 2009, 394.

Sundaram, S., Schwarz, A., & Jones, E. (2007). Technology use on the front line: How information technology enhances individual performance. Journal of the Academy of Marketing Science, 35, 101-112.

Sutton, R. S., & Barto, A. G. (1998). Reinforcement learning: An introduction. MIT Press, Cambridge, MA.

Sykes, E. R. (2011). Interruptions in the workplace: A case study to reduce their ef-fects. International Journal of Information Management, 31(4), 385-394.

Syverson, C. (2011). What determines productivity? Journal of Economic Literature, 49(2), 326-365.

Szabo, J., & Dienes, I. (1998). Ideas and concepts on the Hungarian information econ-omy. Information Processing and Management, 24(2), 183-198.

Taherdoost, H., & Keshavarzsaleh, A. (2015). How to lead to sustainable and success-ful IT project management? Propose 5Ps guideline. International Journal of Ad-vanced Computer Science and Information Technology, 4(1), 14-37.

Takeuchi, N. (2009). How Japanese manufacturing firms align their human resource policies with business strategies: Testing a contingency performance prediction in a Japanese context. The International Journal of Human Resource Management, 20(1), 34-56.

Tallon, P. P. (2014). Do you see what I see? The search for consensus among execu-tives’ perceptions of IT business value. European Journal of Information Systems, 23(3), 306-325.

Tambe, P., Hitt, L. M., & Brynjolfsson, E. (2012). The extroverted firm: How external information practice affect innovation and productivity. Management Science, 58(5), 843-859.

274

Tangen, S. (2005). Demystifying productivity and performance. International Journal of Productivity and Performance Management, 54(1), 34-46.

Tausky, C. (1978). Work organizations: Major theoretical perspectives. F. E. Pea-cock, Itasca Illinois.

Taylor, F. W. (1947). Scientific management. New York: Harper & Row. Tett, R. P., & Burnett, D. D. (2003). A personality trait-based interactionist model of

job performance. Journal of Applied Psychology, 88(3), 500. Tett, R. P., Jackson, D. N., & Rothstein, M. (1991). Personality measures as predictors

of job performance: A meta-analytic review. Personnel Psychology, 44(4), 703-742.

Thompson, M., & Heron, P. (2005). The difference a manager can make: Organiza-tional justice and knowledge worker commitment. The International Journal of Human Resource Management, 16(3), 383-404.

Thompson, R. L., Higgins, C. A., & Howell, J. M. (1994). Influence of experience on personal computer utilization: Testing a conceptual model. Journal of Manage-ment Information Systems, 11(1), 167-187.

Tomaszewski, P., & Lundberg, L. (2005). Software development productivity on a new platform: An industrial case study. Information and Software Technology, 47(4), 257-269.

Torraco, R. J. (2005). Work design theory: A review and critique with implications for human resource development. Human Resource Development Quarterly, 16(1), 85-109.

Tybout, A., Sternthal, B., & Keppel, G. (2001). Analysis of variance. Journal of Con-sumer Psychology, 10(1-2), 5-35.

Vargo, S. L., & Lusch, R. F. (2004). The four service marketing myths remnants of a goods-based, manufacturing model. Journal of Service Research, 6(4), 324-335.

Velthouse, B. A. (1990). Creativity and empowerment: A complimentary relationship. Review of Business, 12, 13-18.

Venkatesh, V., Brown, S. A., Maruping, L. M., & Bala, H. (2008). Predicting different conceptualizations of system use: The competing roles of behavioral intention, fa-cilitating conditions, and behavioral expectation. MIS Quarterly, 32(3), 483-502.

Vroom, V. H. (1964). Work and motivation. New York: Wiley. Vygotsky, L. S. (1978). Mind and society: The development of higher psychological

processes. Cambridge, MA: Harvard University Press.� Waldman, J. D., Yourstone, S. A., & Smith, H. L. (2003). Learning curves in health

care. Health Care Management Review, 28(1), 41-54. Walsham, G. (1993). Interpreting information systems in organizations. John Wiley

& Sons, Inc. Walston, C. E., & Felix, C. P. (1977). A method of programming measurement and

estimation. IBM Systems Journal, 16(1), 54-73. Wasson, C. (2007). Multitasking during virtual meetings. HR Human Resources Plan-

ning, 27(4), 47-60. Weber, B., & Wild, W. (2005). Towards the agile management of business processes.

In K. D. Althoff, A. Dengel, R. Bergmann, M. Nick & T. Roth-Berghofer (Eds.), Professional knowledge management (pp. 409-419). Third Biennial Conference, WM. Springer Berlin Heidelberg.

Webster, M. (2012). Bridging the information worker productivity gap: New chal-lenges and opportunities for it. IDC Information and Data.

275

Wellman, B., Salaff, J., Dimitrova, D., Garton, L., Gulia, M., & Haythornthwaite, C. (1996). Computer networks as social networks: Collaborative work, telework, and virtual community. Annual Review of Sociology, 213-238.

Westfall, P., & Henning, K. S. (2013). Understanding advanced statistical methods. CRC Press.

White, H., McConnell, E., Clipp, E., Bynum, L., Teague, C., Navas, L., & Halbrecht, H. (1999). Surfing the net in later life: A review of the literature and pilot study of computer use and quality of life. Journal of Applied Gerontology, 18(3), 358-378.

Wickens, T. D., & Keppel, G. (2004). Design and analysis: A researcher’s handbook. Englewood Cliffs, NJ: Prentice-Hall.

Wilkinson, A. (1998). Empowerment: Theory and practice. Personnel Review, 27(1), 40-56.

Winship, C., & Morgan, S. L. (1999). The estimation of causal effects from observa-tional data. Annual Review of Sociology, 25, 659-706.

Wohlin, C., Runeson, P., Host, M., Ohlsson, M. C., Regnell, B., & Wesslén, A. (2000). Experimentation in software engineering: An introduction. Boston: Kluwer Academic Publishers.

Wolff, E. N. (2005). The growth of information workers in the U.S. economy. Com-munication of the ACM, 28(10), 37-42.

Womer, N. (1984). Estimating learning curves from aggregate monthly data. Man-agement Science, 30(8), 982-992.

Wu, L. (2012). Social network effects on productivity and job security: Evidence from the adoption of a social networking tool. Information System Research, 24(1), 30-51.

Wu, L., Lin, C-Y., Aral, S., & Brynjolfsson, E. (2009). Value of social network – A large-scale analysis on network structure impact to financial revenue of infor-mation technology consultants. Paper presented at the 2009 Winter Conference on Business Intelligence, February 26-28, 2009; Salt Lake City, UT.

Wu, L., Waber, B. N., Aral, S., Brynjolfsson, E., & Pentland, A. (2008). Mining face-to-face interaction networks using sociometric badges: Predicting productivity in an IT configuration task. Proceedings of the 29th Annual ICIS. Paris, France, 14-17.

Wynekoop, J. L., & Walz, D. B. (1998). Revisiting the perennial question: Are IS people different?. ACM SIGMIS Database, 29(2), 62-72.

Xue, L., Ray, G., & Sambamurthy, V. (2012). Efficiency or innovation: How do in-dustry environments moderate the effects of firms’ IT asset portfolios?. MIS Quar-terly, 36(2), 509-528.

Yaverbaum, G. J. (1988). Critical Factors in the user environment: An experimental study of users, organizations and tasks. MIS Quarterly, 12(1), 75-88.

Zabojnik, J. (2002). Centralized and decentralized decision making in organizations. Journal of Labor Economics, 20(1), 1-22.

Zmud, R. W. (1978). Note-on the validity of the analytic-heuristic instrument utilized in “The Minnesota Experiments”. Management Science, 24(10), 1088-1090.

Zmud, R. W. (1979). Individual differences and MIS success: A review of the empir-ical literature. Management Science, 25(10), 966-979.

Zuckerman, M., Porac, J., Lathin, D., Smith, R., & Deci, E. L. (1978). On the im-portance of self-determination of intrinsically motivated behavior. Personality and Social Psychology Bulletin, 4, 443-446.

.

276

277

Sammanfattning

Föreliggande avhandling adresserar följande fråga: vad är det som påverkar en individs produktivitet, när individens arbete innebär huvudsakligen inform-ationsbehandling och inkluderar användning av informationsteknologiska (IT) system? En mängd tidigare studier har vanligen inkluderat bara en typ av faktorer i taget, så som psykologiska eller organisationsmässiga. Relativt ny-ligen, har en ny föreställning formulerats om att ett antal faktorer av skilda slag måste synkroniseras på ett specifikt sätt för att uppnå höga produktivi-tetsresultat i en verksamhet – den s.k. komplementärteorin. Denna teori anger dock inte vilka specifika faktorer som bör synkroniseras för att en individ som använder sig av IT system ska prestera hög produktivitet i informationsbe-handlingsuppgifter. Denna ansats har anammats av studierna som presenteras i föreliggande avhandling.

En ny forskningsmodell har formulerats där flera faktorer inkluderas från tidigare studier. Till skillnad från tidigare försök att förstå vad det är som på-verkar produktiviteten i en individs informationsbehandling, antar förelig-gande modell en komplementär ansats där flera faktorer av skilda slag kombi-neras, grundat på en inneboende logik för informationsbehandling. Forsk-ningsmodellen innehåller teknologiska faktorer (arbetsprocesser och IT sy-stem användning), individbaserade faktorer (kognitiv profil) och organisatoriska faktorer (beslutsrätt, utbildning, incitament). Forskningsmo-dellen prövas i två distinkta och oberoende studier där informationsbehand-lande professioner undersöks. Den första studien fokuserar på produktiviteten hos säljare i ett läkemedelsföretag och är en omfattande, longitudinell och kvasi-slumpmässig fältstudie. Den andra studien undersöker produktiviteten hos programmerare av mjukvara och är en experimentell studie som genom-förs via internet.

Resultaten påvisar att införandet av ett IT system i en verksamhet inte ger upphov till individuell produktivitetsökning i arbetsuppgifter som rör inform-ationsbehandling, om inte andra komplementära faktorer adresseras på ett syn-kroniserat sätt. Studierna påvisar vikten av att flera faktorer synkroniseras samtidigt och på ett välsynkroniserat sätt för att uppnå produktivitetsökningar hos enskilda individer som använder sig av IT system i sin informationsbe-handling. Föreliggande avhandling bidrar till förståelsen av produktivitetsfak-torer på individnivå samt föreslår vidareutveckling av komplementärteorin som sådan.

222222 278

222222 279

Doctoral Theses Stockholm Business School

1

Nr Year Author Title

205 2016 Mohammad Irani Essays on Mergers and Acquisitions and Event Studies. Stockholm Business School, Stock-holm University.

204 2016 Steffi Siegert Enacting Boundaries through Social Technolo-

gies – The Dance between Work and Private Life. Stockholm Business School, Stockholm University.

203 2016 Andrea Lucarelli The Political Dimension of Place Branding.

Stockholm Business School, Stockholm Uni-versity.

202 2016 Danilo Brozovic Service Provider Flexibility – A Strategic Per-

spective.�Stockholm Business School, Stock-holm University.

201 2015 Andreas Sundström Representing Performance |Performing Repre-

sentation: Ontology in Accounting Practice. Stockholm Business School, Stockholm Uni-versity.

200 2015 Dong Zhang Essays on Market Design and Market Quality.

Stockholm Business School, Stockholm Uni-versity.

199 2015 Niklas Wällstedt Managing Multiplicity: On Control, Care and

the Individual. Stockholm Business School, Stockholm University.

198 2014 Goran Zafirov Essays on Balkan frontier stock markets. Stock-

holm Business School, Stockholm University. 197 2014 Christer Westermark Implementering av redovisning som styrmetod.

Om hållbarhetsredovisningens effekter i statligt ägda företag. Stockholm Business School, Stockholm University.

196 2014 Anna Wettermark Tales of transformation: Expatriate encounters

with local contexts. A postcolonial reading. Stockholm Business School, Stockholm Uni-versity.

195 2014 Randy Ziya Shoai Multinational Enterprises, Sociopolitical Con-

straints and Intermediaries. A Sociopolitically Informed Network Approach. Stockholm Busi-ness School, Stockholm University.

2

194 2014 Christofer Laurell Commercialising social media. A study of fash-ion (blogo)spheres. Stockholm University School of Business.

193 2014 Fredrik Jörgensen The Law Businessman - Five Essays on Legal

Self-efficacy and Business Risk. Stockholm University School of Business.

192 2013 Caihong Xu Essays on Derivatives and Liquidity. Stock-

holm University School of Business. 191 2013 Mikael Andéhn Place-of-Origin Effects on Brand Equity. Expli-

cating the evaluative pertinence of product cat-egories and association strength. Stockholm University School of Business.

190 2013 Sabina Du Rietz Accounting in the field of governance. Stock-

holm University School of Business. 189 2013 Fernholm, Johanna Uppförandekoder som etisk varumärkning? An-

svar i företag med globala värdekedjor. Stock-holm University School of Business.

188 2013 Svärdsten Nymans,

Fredrik Constituting performance: Case studies of per-formance auditing. Stockholm University School of Business.

187 2012 Kumar, Nishant Globalisation and Competitive Sustenance of

Born Global. Evidence from Indian knowledge-intensive service industry. Stockholm Univer-sity School of Business.

186 2012 Yngfalk, Carl The Constitution of Consumption. Food Label-

ing and the Politics of Consumerism. Stock-holm University School of Business.

185 2011 Fyrberg Yngfalk,

Anna Co-Creating Value. Reframing Interactions in Service Consumption. Stockholm University School of Business.

184

2011 Molander, Susanna Mat, kärlek och metapraktik. En studie i var-dagsmiddagskonsumtion bland ensamstående mödrar. Stockholm University School of Busi-ness.

183 2011 Kylsberg, Gösta Kunglig kommunikation – körkonst och tradit-

ion. En autoetnografi om autenticitet i ett kung-ligt konstföretag. Stockholm University School of Business.

3

182 2011 Lindh, Kristina Reciprocal Engagement. A grounded theory of an interactive process of actions to establish, maintain, and develop an enterprise. Stock-holm University School of Business.

181

2011 Schultz-Nybacka, Pamela

Bookonomy. The Consumption Practice and Value of Book Reading. Stockholm University School of Business.

180 2011 Lund, Ragnar Leveraging cooperative strategy – cases of

sports and arts sponsorship. Stockholm Uni-versity School of Business.

179 2010 Svendsen, Jens Martin Gör som jag säger! igen och igen – om lojalitet

och lek i marknadsföringen: en beskrivning av legitimeringssystematik. Stockholm University School of Business.

178 2010 Hansson, Jörgen Köp av tjänster för ledningskompetens – en po-lyfonisk process. Stockholm University School of Business.

177 2010 Ljungbo, Kjell Language as a Leading Light to Business Cul-tural Insight. A Study on Expatriates’ Intercul-tural Communication in Central and Eastern Europe. Stockholm University School of Busi-ness.

176 2010 Demir, Robert Strategy as Sociomaterial Practices: Planning, Decision-Making, and Responsiveness in Cor-porate Lending. Stockholm University School of Business.

175 2010 Radón, Anita The Rise of Luxury Brands Online: A study of how a sense of luxury brand is created in an online environment. Stockholm University School of Business.

174 2010 Martinsson, Irene Standardized Knowledge Transfer: A study of Project-Based Organizations in the Construc-tion and IT Sectors. Stockholm University School of Business.

173 2009 Digerfeldt-Månsson,

Theresa Formernas liv i designföretaget - om design och design management som konst. Stockholm University School of Business.

172 2009 Larsson Segerlind, Tommy

Team Entrepreneurship – A Process Analysis of the Venture Team and the Venture Team Roles in relation to the Innovation Process. Stock-holm University School of Business.

4

171

2009 Svensson, Jenny The Regulation of Rule - Following. Imitation and Soft Regulation in the European Union. Stockholm University School of Business.

170 2009 Wittbom, Eva Att spränga normer - om målstyrningsprocesser för jämställdhetsintegrering. Stockholm Uni-versity School of Business.

169 2009 Wiesel, Fredrika Kundorientering och ekonomistyrning i offent-lig sektor. Stockholm University School of Business.

168 2008 Essén, Anna Technology as an Extension of the Human

Body: Exploring the potential role of technol-ogy in an elderly home care setting. Stockholm University School of Business.

167 2008 Forslund, Dick Hit med pengarna! Sparandets genealogi och den finansiella övertalningens vetandekonst. Stockholm University School of Business.

166 2008 Gustafsson, Clara Brand Trust: Corporate communications and consumer-brand relationships. Stockholm Uni-versity School of Business.

165 2008 Jansson, Elisabeth Paradoxen (s)om entreprenörskap: En roman-tisk ironisk historia om ett avvikande entrepre-nörskapande. Stockholm University School of Business.

164 2008 Jüriado, Rein Learning within and between public-private partnerships. Stockholm University School of Business.

163 2008 Söderholm Werkö, Sophie

Patient Patients? Achieving Patient Empower-ment through active participation, increased knowledge and organisation. Stockholm Uni-versity School of Business.

162 2008 Tomson, Klara Amnesty in Translation. Ideas, Interests and Organizational Change. Stockholm University School of Business.

161 2007 Carrington, Thomas Framing Audit Failure - Four studies on qual-

ity discomforts. Stockholm University School of Business.

160 2007 Dahl, Matilda States under scrutiny. International organiza-tions, transformation and the construction of progress. Stockholm University School of Business.

5

159 2007 Gawell, Malin Activist Entrepreneurship - Attac´ing Norms and Articulating Disclosive Stories. Stockholm University School of Business.

158 2007 Ihrfors, Robert Spelfrossa - Spelets makt och maktens spel. Stockholm University School of Business.

157 2007 Karlsson, Anders Investment Decisions and Risk Preferences among Non-Professional Investors. Stockholm University School of Business.

156 2007 Vigerland, Lars Homo Domesticus. En marknadsanalys av bo-stadskonsumenters strategier och preferenser. Stockholm University School of Business.

155 2007 Värlander, Sara Framing and Overflowing. How the Infusion of Information Technology Alters Proximal Ser-vice Production. Stockholm University School of Business.

154 2006 Ahlström Söderling,

Ragnar Regionala företags förutsättningar för internat-ionell konkurrenskraft. Stockholm University School of Business.

153 2006 Beckius, Göran Företagsetik. En studie av etiskt organiserande i några svenska företag. Stockholm University School of Business.

152 2006 Ferdfelt, Henrik Pop. Stockholm University School of Business.

151 2006 Sjödin, Ulrika Insiders´ Outside/Outsiders´Inside - rethinking the insider regulation. Stockholm University School of Business.

150 2006 Skoglund, Wilhelm Lokala samhällsutvecklingsprocesser och ent-reprenörskap. Stockholm University School of Business.

149 2005 Bengtsson, Elias Shareholder activism of Swedish institutional

investors. Stockholm University School of Business.

148 2005 Holmgren, Mikael A passage to organization. Stockholm Univer-sity School of Business.

147 2005 Thornquist, Clemens The Savage and the Designed: Robert Wilson and Vivienne Westwood as Artistic Managers. Stockholm University School of Business.

146 2004 Sjöstrand, Fredrik Nätverkskoordineringens dualiteter. Stockholm

University School of Business.

6

145 2004 Khan, Jahangir Hoss-ain

Determinants of Small Enterprise Development of Bangladesh. Stockholm University School of Business.

144 2004 Almqvist, Roland Icons of New Public Management. Four studies on competition, contract and control. Stock-holm University School of Business.

143 2004 Yazdanfar, Darush Futures som ett mångsidigt instrument. En em-pirisk studie av oljebolag som använder futu-reskontrakt. Stockholm University School of Business.

142 2003 Skoog, Matti Intangibles and the transformation of manage-ment control systems - Five studies on the changing character of management control sys-tems in Swedish organizations. Stockholm Uni-versity School of Business.

141 2003 Elmersjö, Carl-Åke Moralisk ekonomi i sjukvården? - Om etik och ekonomi i sjukhusets vardagsorganisering. Stockholm University School of Business.

140 2003 Koponen, Anja Företagens väg mot konkurs. Stockholm Uni-versity School of Business.

139 2003 Frostling-Hennings-son, Maria

Internet Grocery Shopping - A Necessity, A Pleasurable Adventure, or an Act of Love. Stockholm University School of Business.

138 2003 Köping, Ann-Sofie Den Bundna friheten. Om kreativitet och relat-ioner i ett konserthus. Stockholm University School of Business.

137 2003 Bagelius, Nils Svenska företag åter i österled: Hur svenska fö-retag positionerade sig i Öst och minskade sin exponering för risk och osäkerhet. Stockholm University School of Business.

136 2003 Lindqvist, Katja Exhibition enterprising - six cases of realisa-tion from idea to institution. Stockholm Univer-sity School of Business.

135 2003 Soila-Wadman, Marja Kapitulationens estetik. Organisering och le-darskap i filmprojekt. Stockholm University School of Business.

134 2003 Lundkvist, Anders Conversational Realities - Five Studies of User Interactions as Sources of Innovation. Stock-holm University School of Business.

7

133 2003 Willstrand-Holmer, Sofia

Att konstruera kunskap om kunder - en studie om förändring och berättelser i ICA-samman-slutningen. Stockholm University School of Business.

132 2003 Roy, Sofie Navigating in the Knowledge Era. Metaphors and Stories in the Construction of Skandia’s Navigator. Stockholm University School of Business.

131 2003 Tollhagen, Renate Skräddare utan tråd - en illustration av fyra fö-retag i klädbranschen. Stockholm University School of Business.

130 2002 Hansson, Johan Omtänkbara organisationer – Sagor och utsa-

gor om Astrid Lindgrens Barnsjukhus. Stock-holm University School of Business.

129 2002 Pramborg, Bengt Empirical Essays on Foreign Exchange Risk Management. Stockholm University School of Business.

128 2002 Axén-Ruzicka, Jea-nette

Införande av ny teknik. En studie av problem vid införande av elektroniska marknadsplatser. Stockholm University School of Business.

127 2002 Torpman, Jan Rättssystemets Lärande. Stockholm University School of Business.

126 2002 Dahlström, Karin Värdeskapande produktutveckling i tjänstein-

tensiva företag. Stockholm University School of Business.

125 2002 Gravesen, Inger Fitnessövningar och husförhör: Om förbätt-ringsprocesser i företag. Stockholm University School of Business.

124 2001 Gottfridsson, Patrik Småföretags tjänsteutveckling - en studie av

hur småföretag utvecklar individuellt anpas-sade tjänster. Stockholm University School of Business.

123 2001 Engström, Malin Essays on Equity Options. Stockholm Univer-sity School of Business.

122 2001 Gatarski, Richard Artificial Market Actors: Explorations of Auto-mated Business Interactions. Stockholm Uni-versity School of Business.

121 2001 Hansson, Bo Essays on Human Capital Investments. Stock-holm University School of Business.

8

120 2001 Wåhlstedt, Håkan Resultatredovisning för hållbar utveckling. Na-turekonomiska principer för kommunal tillämp-ning. Stockholm University School of Business.

119 2001 Golubeva, Olga Foreign Investment Decision-Making in Transi-tion Economies. Stockholm University School of Business.

118 2001 Catasús, Bino Borders of Management. Five Studies of Ac-counting, Organizing and the Environment. Stockholm University School of Business.

117 2001 Eklöv, Gunilla Auditability as Interface - Negotiation and Sig-nification of Intangibles. Stockholm University School of Business.

116 2001 Lennstrand, Bo HYPE IT - IT as Vision and Reality - on Diffu-sion, Personalization and Broadband. Stock-holm University School of Business.

115 2001 von Friedrich-Gräng-sjö, Yvonne

Destinationsmarknadsföring. En studie av tur-ism ur ett producentperspektiv. Stockholm Uni-versity School of Business.

114 2001 Wetterström, Jeanette Stor opera - små pengar. Stockholm University School of Business.

113 2001 Friman, Henrik Strategic Time Awareness - Implications of Strategic Thinking. Stockholm University School of Business.

112 2001 Apéria, Tony Brand Relationship Management: den varu-märkesbyggande processen. Stockholm Univer-sity School of Business.

111 2001 Johansson, Stig G Individens roll i strategiska informationssy-stem. Stockholm University School of Busi-ness.

110 2001 Carlell, Camilla Technology in Everyday Life - A study of Con-sumers and Technology in a Banking Context. Stockholm University School of Business.

109 2001 Maravelias, Christian Managing Network Organisations. Stockholm University School of Business.

108 2000 Holmqvist, Mikael The Dynamics of Experiential Learning. Bal-

ancing Exploitation and Exploration Within and Between Organizations. Stockholm Uni-versity School of Business.

107 2000 Hamde, Kiflemariam Shifting Identities: Teamwork and Supervisors in Swedish Change Programmes for the Last Three Decades. Stockholm University School of Business.

9

106 2000 Uggla, Henrik Managing the Brand-Association Base: Explor-ing Facets of Strategic Brand Management from the Imaginary Organization perspective. Stockholm University School of Business.

105 2000 Rämö, Hans The Nexus of Time and Place in Economical Operations. Stockholm University School of Business.

104 2000 Aronsson, Britt Prisdifferentieringars janusansikte. Prisdiffe-rentieringar mot mellanled som ett konkurrens-politiskt problem. Stockholm University School of Business.

103 2000 Porsander, Lena TITT-SKÅP FÖR ALLA - en berättelse om hur Stockholm blev kulturhuvudstad. Stockholm University School of Business.

102 2000 Hagelin, Niclas Empirical Essays on Financial Markets, Firms, and Derivatives. Stockholm University School of Business.

101 2000 Berglund, Åke Blomquist, Anders

Från affärskompetens till affärsutveckling i småföretag. Stockholm University School of Business.

100 2000 Näsman, Birgitta Pappas flickor. Entreprenöriella processer i kvinnoföretagandets tillkomst. Stockholm Uni-versity School of Business.

99 1999 Lundgren, Maths Bankens natur - miljöfrågans genomslag i

svenska banker. Stockholms universitet, Före-tagsekonomiska institutionen.

98 1998 Björkman, Ivar Sven Duchamp - Expert på auraproduktion:

Om entreprenörskap, visioner, konst och före-tag. Stockholms universitet, Företagsekono-miska institutionen.�

97 1998 Millak, Jurek Organisatorisk kompetens. Stockholms univer-sitet, Företagsekonomiska institutionen.

96 1998 Wiklander, Levi Intertextuella strövtåg i Akademia. Stockholms universitet, Företagsekonomiska institutionen.

95 1998 Bay, Thomas ...AND...AND...AND - Reiterating Financial Derivation. Stockholm University, School of Business.

10

94 1998 Malver, Henrik Service in the Airlines - Customer or Competi-tion Oriented? Stockholm University, School of Business.

93 1998 Granberg, Georg Vägar mot ökad konkurrens och marknadsstyr-ning av offentlig sektor. Stockholms universitet, Företagsekonomiska institutionen.

92 1998 Bjurklo, Margareta Kardemark, Gunnel

Nyckelord - en nyckel vid kompetensutveckling. Stockholms universitet, Företagsekonomiska institutionen.

91 1997 Wallin Andreassen,

Tor Dissatisfaction with Services - The Impact of Satisfaction with Service Recovery on Corpo-rate Image and Future Repurchase Intention. Stockholm University, School of Business.

90 1997 Alkebäck, Per Do Dividend Changes Really Signal? – Evi-dence from Sweden. Stockholm University, School of Business.

89 1997 Lagrosen, Stefan Kvalitetsstyrning i skolan? - En analys av TQM:s tillämpbarhet inom den svenska grund-skolan sett från en företagsekonomisk utgångs-punkt. Stockholms universitet, Företagsekono-miska institutionen.

88 1997 Andersson, Göran Framgång i kommersiella tjänsteverksamheter. Stockholms universitet, Företagsekonomiska institutionen.

87 1996 Le Duc, Michaël Constructivist Systemics - Theoretical Elements

and Applications in Environmental Informatics. Stockholm University, School of Business.

86 1996 Preiholt, Håkan The Organization of Manufacturing Know-How. Stockholm University, School of Busi-ness.

85 1996 Green, Bo Analys av komplexa samhällssystem - Aktions-inriktade fallstudier och metodologiska kon-klusioner. Stockholms universitet, Företagseko-nomiska institutionen.

84 1996 Edenius, Mats Ett modernt dilemma - organiserandet kring elektronisk post. Stockholms universitet, Före-tagsekonomiska institutionen.�

83 1996 Hedlin, Pontus Accounting Investigations. Stockholm Univer-sity, School of Business.

82 1996 Yakhlef, Ali Organizing as Discursive Practices: The Exam-ple of Information Technology Outsourcing. Stockholm University, School of Business.

11

81 1996 Wahlgren, Ingela Vem tröstar Ruth? Stockholms universitet, Fö-retagsekonomiska institutionen.

80 1996 Rutihinda, Cranmer Resource-based internationalization: Entry Strategies of Swedish Firms into the Emerging Markets of Eastern Europe Stockholm Univer-sity, School of Business.

79 1996 Liljefors, Ole Efterfrågan och utbud av kompetensutveck-lande ledningsarbete. Stockholms universitet, Företagsekonomiska institutionen.

78 1995 Asproth, Viveca Visualization of Dynamic Information. Stock-

holm University, School of Business.

77 1995 Håkansson, Anita Models and Methods for the Management of Dynamic Information in GEOinformatic Sys-tems. Stockholm University, School of Busi-ness.

76 1995 Khodabandehloo, Ak-bar

Marknadsföring som utbyte: en idéhistoria. en pluralistisk ansats. Stockholms universitet, Fö-retagsekonomiska institutionen.

75 1995 Rylander, Leif Tillväxtföretag i startfas. Från dimma och mör-ker till relationslyft. Stockholms universitet, Företagsekonomiska institutionen.

74 1995 Malmström, Li Lärande organisationer? Krisen på den svenska fastighetsmarknaden. Stockholms uni-versitet, Företagsekonomiska institutionen.

73 1995 Brunson, Karin Dubbla budskap. Hur riksdag och regering pre-senterar sitt budgetarbete. Stockholms univer-sitet, Företagsekonomiska institutionen.

72 1994 Sveiby, Karl-Erik Towards a knowledge perspective on organisa-

tion. Stockholm University, School of Busi-ness.

71 1994 Bergqvist, Erik Belöningar och prestationer i offentlig verk-samhet - En utvärdering av fyra fall inom Stockholms läns landsting. Stockholms univer-sitet, Företagsekonomiska institutionen.

70 1994 Paul, Ann-Sofi Organisationsutveckling genom personalenkä-ter – en personalekonomisk utvärdering.Stock-holms universitet, Företagsekonomiska institut-ionen.

69 1994 Bergström, Cecilia A Female Cooperative Perspective on Power Influence and Ownership. Stockholm Univer-sity, School of Business.

12

68 1994 Borg, Erik European Markets and Management Action: Making Sense of a Europe Without Frontiers. Stockholm University, School of Business.

67 1994 Olsson, Birgitta Kortare arbetsdag - en väg till ett mer ekolo-giskt arbetsliv? Stockholms universitet, Före-tagsekonomiska institutionen.

66 1993 Thomasson, Bertil Tjänstekvalitet - Kundorienterad och kompe-

tensbaserad kvalitetsutveckling. Stockholms universitet, Företagsekonomiska institutionen.

65 1993 Tesfaye, Besrat Determinants or Entrepreneurial Processes. A Case Study of Technology-Based Spin-off Com-pany Formations. Stockholm University, School of Business.

64 1993 Norling, Per Tjänstekonstruktion - Service Design. Stock-holms universitet, Företagsekonomiska institut-ionen och Högskolan i Karlstad.

63 1993 Ramfelt, Lena Näringspolitiska samverkansprojekt ur ett or-ganisationsperspektiv – Substantiella och sym-boliska aspekter på organisatoriskt handlande. Stockholms universitet, Företagsekonomiska institutionen.

62 1993 Sigfridsson, Jan Strategisk ekonomistyrning i tidningsföre-tag - Aktionsforskning i ekonomisk ledningsin-formation. Stockholms universitet, Företagse-konomiska institutionen.

61 1992 Olsen, Morten Kvalitet i banktjänster. Privatkunders upplevda

problem med banktjänster – En studie med kri-tisk-händelse-metoden. Stockholms universitet, Företagsekonomiska institutionen.

60 1992 Gustavsson, Bengt The Transcendent Organization. Stockholm: Stockholm University, School of Business.

59 1992 Borgert, Leif Organisation som mode. Kontrasterande bilder av svensk hälso- och sjukvård. Stockholms uni-versitet, Företagsekonomiska institutionen.

58 1992 Osarenkhoe, Aihie Improving Food Product Distribution in Devel-oping Countries: A Case Study of Nigeria. Stockholm University, School of Business.

57 1992 Westerberg, Lillemor Föreställningar på arenan. Ett utvecklingsar-bete kring eget budgetansvar på kommunala barnstugor. Stockholms universitet, Företagse-konomiska institutionen.

13

56 1992 Johanson, Ulf Nilson, Marianne

Personalekonomiska beräkningars användbar-het. Stockholms universitet, Företagsekono-miska institutionen.

55 1991 Feurst, Ola Kost och hälsa i marknadsföringen. En analys

av system och processer i vilka våra matvanor formas. Med betoning på livsmedelsannonsers spegling av kostideal och konsumtion 1950-85. Stockholms universitet, Företagsekonomiska institutionen.

54 1991 Pihliamäki, Klara Media Technology and Communication Pat-terns in the Organizational Interface. Stock-holm University, School of Business.

53 1990 Ekvall, Arne Affärsidéer - En empirisk studie av hur företags verksamhetsinriktning kan analyseras och besk-rivas utifrån ett affärsidébegrepp. Stockholms universitet, Företagsekonomiska institutionen.

52 1990 Sotto, Richard Man without Knowledge - Actors and Specta-tors in Organizations. Stockholm University, School of Business.

51 1990 Zineldin, Mosad The Economics of Money and Banking - a The-oretical and Empirical Study of Islamic Inter-est-Free Banking. Stockholm University, School of Business.

50 1990 Tollin, Karin Konsumentbilder i marknadsföringen av livs-

medel - en studie om marknadsföringens kon-text inom svensk lantbrukskooperativ livsme-delsindustri. Stockholms universitet, Företags-ekonomiska institutionen.�

49 1990 Wagué, Cheick Entrepreneurship and industrial policy in de-veloping countries. A case study of principal policy constraints which limit the development and expansion of private sector industrial en-terprises in Mali. Stockholm University, School of Business.

48 1989 Eriksson, Gunilla Framtidsinriktade aktörsperspektiv på

branscher - metodsynpunkter med utgångs-punkt från en studie i svensk dagligvaruindu-stri. Stockholms universitet, Företagsekono-miska institutionen.�

47 1989 Winai, Peter Gränsorganisationer. Egenskaper, problem och utvecklingsmöjligheter hos organisationer i gränslandet mellan privat och offentlig sektor. Stockholms universitet, Företagsekonomiska institutionen. �

14

46 1989 Åredal, Åke Den osynliga styrningen. En hermeneutisk stu-die av styrningen inom svensk tandvård. Stock-holms universitet, Företagsekonomiska institut-ionen.�

45 1989 Kaiser, Bo Produktlivscykler i dagligvaruhandeln. För-ändringar i utbudet av några livsmedel och ke-misk-tekniska produkter efter 1970. Stockholms universitet, Företagsekonomiska institutionen.

44 1988 Scheutz, Curt Företagsfissioner. Avknoppningar till Stock-

holms Fondbörs och OTC-marknaden - en em-pirisk undersökning av motiv och konsekvenser. Stockholms universitet, Företagsekonomiska institutionen.

43 1988 Eriksson, Lars Torsten Myndigheters marknadsorientering. Om mark-nadsföringsfrågor i avgiftsfinansierade statliga myndigheter. Stockholms universitet, Före-tagsekonomiska institutionen.

42 1987 Barius, Bengt Investeringar och marknadskonsekvenser. En empirisk undersökning av investeringsärenden och särskilt av möjligheter att bedöma investe-ringars framtida marknadskonsekvenser. Stockholms universitet, Företagsekonomiska institutionen.

41 1987 Liukkonen, Paula Det lokala arbetsmiljöarbetets effektivitet. En fallstudie från kvarteret Garnisonen. Stock-holms universitet, Företagsekonomiska institut-ionen.

40 1987 Öhrming, Jan Förvaltning av flerbostadshus. Om arbetsorga-nisation och föreställningar som villkor för samspel och boendemedverkan. Stockholms universitet, Företagsekonomiska institutionen.

39 1987 Kostopoulos, Trifon The Decline of the Market: the ruin of capital-ism and anti-capitalism. Stockholm University, School of Business.

38 1987 de Ridder, Adri Access to the Stock Market. An empirical study of the efficiency of the British and the Swedish primary markets. Stockholm University, School of Business.

37 1986 Ehrengren, Lars Riskhantering vid u-landsinvestering. En teore-

tisk studie och en empirisk undersökning av ett antal svenska industriföretags produktionsinve-steringar. Stockholms universitet, Företags-ekonomiska institutionen.

15

36 1985 Senning, Eva-Marie Kostnadssamband och kostnadsstyrning inom fastighetsförvaltning. Med tillämpningar på Riksbyggens drift- och underhållskostnader. Stockholms universitet, Företagsekonomiska institutionen.

35 1985 Onwuchekwa, Chris-tian Ike

Agricultural Cooperatives and Problems of Transition. A study of organizational develop-ment problems in rural development. Univer-sity of Stockholm, Department of Business Ad-ministration.

34 1985 Hilding, Madeleine Arbetstrivsel och psykisk påfrestning. En studie av arbetsmiljö i samband med omlokalisering av statlig verksamhet. Stockholms universitet, Företagsekonomiska institutionen.

33 1985 Valdemarsson, Bengt Förväntningar inför arbetslivet. En longitudi-nell studie hos några ungdomar av förväntning-ars uppkomst och deras betydelse för inställ-ningen till arbetslivet i industriföretag. Stock-holms universitet, Företagsekonomiska institut-ionen.

32 1985 Badran, Mohga Coordination In Multiactor Programs: An Em-pirical Investigation of Factors Affecting Coor-dination among Organizations at the Local Level in the Egyptian Family Planning Pro-gram. University of Stockholm, Department of Business Administration.

31 1984 Myrsten, Karl Lönsam samverkan. En studie av utvecklings-

processer inom området fastighetsreparationer. Stockholms universitet, Företagsekonomiska institutionen.

30 1981 Abdel-Khalik, Ali The Production and Distribution of Milk and

Dairy Products in Egypt: towards a Co-opera-tive System. University of Stockholm, Depart-ment of Business Administration.

29 1981 Hedvall, Maria Participation i företag. En jämförelse mellan ett jugoslaviskt och ett svenskt tobaksföretag. Stockholms universitet, Företagsekonomiska institutionen.

28 1980 Vavrin, Jeanette The Airline Insurance Industry. A future study.

University of Stockholm, Department of Busi-ness Administration.

27 1980 Håkansson, Stefan Kostnadsvariationer inom sjukvården - jämfö-rande studier på landstings- och kliniknivå. Stockholms universitet, Företagsekonomiska institutionen.

16

26 1980 Bergström, Erik Projektorienterad marknadsföring. En studie av fem försäljningar av komplexa anläggningar. Stockholms universitet, Företagsekonomiska institutionen.

25 1980 Edsbäcker, Göran Marginal Cost Pricing of Electricity. Univer-sity of Stockholm, Department of Business Ad-ministration.

24 1980 Högberg, Olle Föreställningar och spelregler i kommunal pla-nering. Stockholms universitet, Företagsekono-miska institutionen.

23 1980 Klingberg, Tage Byggforskningen T10:1980, En studie av bygg-nadsnämndens tillsyn. Stockholms universitet, Företagsekonomiska institutionen.

22 1978 Lindgren, Christer Broms och inlärning. Tjänstemän i Västerås planerar ett bostadsområde. Stockholms uni-versitet, Företagsekonomiska institutionen.

21 1978 Granqvist, Roland Studier i sjukvårdsekonomi. Stockholms univer-sitet, Företagsekonomiska institutionen.

20 1978 Gröjer, Jan-Erik Stark, Agneta

Social redovisning. Stockholms universitet, Fö-retagsekonomiska institutionen.

19 1978 Khan, Sikander A Study of Success and Failure in Exports. An empirical investigation of the export perfor-mance of 165 market ventures of 83 firms in the chemical and electronics manufacturing indus-tries. University of Stockholm, Department of Business Administration.

18 1977 Mills, Peter New Perspectives on Post-Industrial Organiza-

tions. An empirical investigation into the theo-ries and practices of service firms. University of Stockholm, Department of Business Admin-istration.

17 1977 Bergström, Sören Konsumentperspektiv på dagligvaruföretag. En analys av hur företagens arbetssätt och arbets-förutsättningar inverkar på konsumentproblem. Stockholms universitet, Företagsekonomiska institutionen.

16 1977 Gummesson, Evert Marknadsföring och inköp av konsulttjäns-ter. En studie av egenskaper och beteenden i producenttjänstmarknader. Stockholms univer-sitet, Företagsekonomiska institutionen.

17

15 1977 Hansson, Roland Friställd. En studie av konsekvenserna för de anställda vid två företagsnedläggningar. Stock-holms universitet, Företagsekonomiska institut-ionen.

14 1976 Widman, Leif Alternativa distributionssystem. En samhällse-konomisk modellstudie av dagligvarudistribut-ionen. Stockholms universitet, Företagsekono-miska institutionen.

13 1975 Lilja, Johan Läkares läkemedelsval ur samhällets synvinkel. En stuide av möjligheterna att med hjälp av of-fentliga åtgärder påverka läkarnas preparatval utanför sjukhus. Stockholms universitet, Företagsekonomiska institutionen.

12 1975 Söderman, Sten Industrial Location Planning. An empirical in-vestigation of company approaches to the prob-lem of locating new plants. University of Stock-holm, Department of Business Administration.

11 1975 Ljung, Birger Selmer, Jan

Samordnad planering i decentraliserade före-tag. En studie av Danzig & Wolfe's dekomposit-ionsalgoritm. Stockholms universitet, Före-tagsekonomiska institutionen.

10 1974 Rapp, Birger Models for Optimal Investment and Mainte-

nance Decisions. University of Stockholm, De-partment of Business Administration.

9 1973 Lindberg, Jens Externa effekter av dryckesförpackningar. En

studie av några åtgärder. Stockholms universi-tet, Företagsekonomiska institutionen.

8 1973 Rundfelt, Rolf Reklamens kostnader och bestämningsfaktorer. Stockholms universitet, Företagsekonomiska institutionen.

7 1973 Leonardz, Björn To Stop or Not to Stop, Some Elementary Opti-mal Stopping Problems with Economic Inter-pretations. University of Stockholm, Depart-ment of Business Administration.

6 1972 Sellstedt, Bo Selection of Product Development Projects Un-

der Uncertainty. University of Stockholm, De-partment of Business Administration.

18

5 1972 Åhrell, Lars Nedskräpning som ett ekonomiskt problem. Stockholms universitet, Företagsekonomiska institutionen.

4 1971 Lönnstedt, Lars Operationsanalys i börsnoterade företag.

Stockholms universitet, Företagsekonomiska institutionen.

3 1970 Gullander, Staffan En studie i produktionsplanering. Stockholms

universitet, Företagsekonomiska institutionen.

2 1970 Engwall, Lars Size Distributions of Firms. Stockholms univer-sitet, Företagsekonomiska institutionen.

1 1969 Bergendahl, Göran Models for investments in a road network. Uni-

versity of Stockholm, Department of Business Administration.

19

The Swedish Research School of Management and Information Technology (MIT) is one of 16 national research schools supported by the Swedish Government. MIT is jointly operated by the following institutions: Blekinge Institute of Technology, IT University of Göteborg, Jönköping International Business School, Karlstad Univer-sity, Linköping University, Linnaeus University Växjö, Lund University, Mälardalen University College, Stockholm University, Umeå University, Örebro University, and Uppsala University, host to the research school. At the Swedish Research School of Management and Information Technology (MIT), research is conducted, and doctoral education provided, in three fields: management information systems, business ad-ministration, and informatics.

DISSERTATIONS FROM THE SWEDISH RESEARCH SCHOOL OF MANAGEMENT AND INFORMATION

TECHNOLOGY Doctoral theses (2003- ) 1. Baraldi, Enrico (2003), When Information Technology Faces Resource Interaction: Us-

ing IT Tools to Handle Products at IKEA and Edsbyn. Department of Business Studies, Uppsala University, Doctoral Thesis No. 105.

2. Wang, Zhiping (2004), Capacity-Constrained Production-Inventory Systems: Modelling and Analysis in both a Traditional and an E-Business Context. IDA-EIS, Linköpings uni-versitet och Tekniska Högskolan i Linköping, Dissertation No. 889

3. Ekman, Peter (2006), Enterprise Systems & Business Relationships: The Utilization of IT in the Business with Customers and Suppliers. School of Business, Mälardalen University, Doctoral Dissertation No 29.

4. Lindh, Cecilia (2006), Business Relationships and Integration of Information Technol-ogy. School of Business, Mälardalen University, Doctoral Dissertation No 28.

5. Frimanson, Lars (2006), Management Accounting and Business Relationships from a Supplier Perspective. Department of Business Studies, Uppsala University, Doctoral The-sis No. 119.

The Swedish Research School

of Management and Information Technology MIT

20

6. Johansson, Niklas (2007), Self-Service Recovery. Information Systems, Faculty of Eco-nomic Sciences, Communication and IT, Karlstad University, Dissertation KUS 2006:68.

7. Sonesson, Olle (2007), Tjänsteutveckling med personal medverkan: En studie av bank-tjänster. Företagsekonomi, Fakulteten för ekonomi, kommunikation och IT, Karlstads uni-versitet, Doktorsavhandling, Karlstad University Studies 2007:9.

8. Maaninen-Olsson, Eva (2007), Projekt i tid och rum: Kunskapsintegrering mellan pro-jektet och dess historiska och organisatoriska kontext. Företagsekonomiska institutionen, Uppsala universitet, Doctoral Thesis No. 126.

9. Keller, Christina (2007), Virtual learning environments in higher education: A study of user acceptance. Linköping Studies in Science and Technology, Dissertation No. 1114.

10. Abelli, Björn (2007), On Stage! Playwriting, Directing and Enacting the Informing Pro-cesses. School of Business, Mälardalen University, Doctoral Dissertation No. 46.

11. Cöster, Mathias (2007), The Digital Transformation of the Swedish Graphic Industry. Linköping Studies in Science and Technology, Linköping University, Dissertation No. 1126.

12. Dahlin, Peter (2007), Turbulence in Business Networks: A Longitudinal Study of Mergers, Acquisitions and Bankruptcies Involving Swedish IT-companies. School of Business, Mä-lardalen University, Doctoral Thesis No. 53.

13. Myreteg, Gunilla (2007), Förändringens vindar: En studie om aktörsgrupper och konsten att välja och införa ett affärssystem. Företagsekonomiska institutionen, Uppsala universi-tet, Doctoral Thesis No. 131.

14. Hrastinski, Stefan (2007), Participating in Synchronous Online Education. School of Economics and Management, Lund University, Lund Studies in Informatics No. 6.

15. Granebring, Annika (2007), Service-Oriented Architecture: An Innovation Process Per-spective. School of Business, Mälardalen University, Doctoral Thesis No. 51.

16. Lövstål, Eva (2008), Management Control Systems in Entrepreneurial Organizations: A Balancing Challenge. Jönköping International Business School, Jönköping University, JIBS Dissertation Series No. 045.

17. Hansson, Magnus (2008), On Closedowns: Towards a Pattern of Explanation to the Closedown Effect. Swedish Business School, Örebro University, Doctoral Thesis No. 1.

18. Fridriksson, Helgi-Valur (2008), Learning processes in an inter-organizational context: A study of krAft project. Jönköping International Business School, Jönköping University, JIBS Dissertation Series No. 046.

19. Selander, Lisen (2008), Call Me Call Me for some Overtime: On Organizational Conse-quences of System Changes. Institute of Economic Research, Lund Studies in Economics and Management No. 99.

21

20. Henningsson, Stefan (2008), Managing Information Systems Integration in Corporate Mergers & Acquisitions. Institute of Economic Research, Lund Studies in Economics and Management No. 101.

21. Ahlström, Petter (2008), Strategier och styrsystem för seniorboende-marknaden. IEI-EIS, Linköping universitetet och Tekniska Högskolan i Linköping, Doktorsavhandling, Nr. 1188.

22. Sörhammar, David (2008), Consumer-firm business relationship and network: The case of ”Store” versus Internet. Department of Business Studies, Uppsala University, Doctoral Thesis No. 137.

23. Caesarius, Leon Michael (2008), In Search of Known Unknowns: An Empirical Investi-gation of the Peripety of a Knowledge Management System. Department of Business Stud-ies, Uppsala University, Doctoral Thesis No. 139.

24. Cederström, Carl (2009), The Other Side of Technology: Lacan and the Desire for the Purity of Non-Being. Institute of Economic Research, Lund University, Doctoral Thesis, ISBN: 91-85113-37-9.

25. Fryk, Pontus, (2009), Modern Perspectives on the Digital Economy: With Insights from the Health Care Sector. Department of Business Studies, Uppsala University, Doctoral Thesis No. 145.

26. Wingkvist, Anna (2009), Understanding Scalability and Sustainability in Mobile Learn-ing: A Systems Development Framework. School of Mathematics and Systems Engineer-ing, Växjö University, Acta Wexionesia, No. 192, ISBN: 978-91-7636-687-5.

27. Sällberg, Henrik (2010), Customer Rewards Programs: Designing Incentives for Re-peated Purchase. Blekinge Institute of Technology, School of Management, Doctoral Dis-sertation Series No. 2010:01.

28. Verma, Sanjay (2010), New Product Newness and Benefits: A Study of Software Products from the Firms’ Perspective, Mälardalen University Press, Doctoral Thesis.

29. Iveroth, Einar (2010), Leading IT-Enabled Change Inside Ericsson: A Transformation Into a Global Network of Shared Service Centres. Department of Business Studies, Upp-sala University, Doctoral Thesis No. 146.

30. Nilsson, Erik (2010), Strategi, styrning och konkurrenskraft: En longitudinell studie av Saab AB, IEI-EIS, Linköpings universitet och Tekniska Högskolan i Linköping, Doktors-avhandling, Nr. 1318.

31. Sjöström, Jonas (2010), Designing Information Systems: A pragmatic account, Depart-ment of Informatics and Media, Uppsala University, Doctoral Thesis.

32. Numminen, Emil (2010), On the Economic Return of a Software Investment: Managing Cost, Benefit and Uncertainty, Blekinge Institute of Technology, School of Management, Doctoral Thesis.

22

33. Frisk, Elisabeth (2011), Evaluating as Designing: Towards a Balanced IT Investment Approach, IT University, Göteborg, Doctoral Thesis.

34. Karlsudd, Peter (2011), Support for Learning: Possibilities and Obstacles in Learning Applications, Mälardalen University, Doctoral Thesis.

35. Wicander, Gudrun (2011), Mobile Supported e-Government Systems: Analysis of the Education Management Information System (EMIS) in Tanzania, Karlstad University, Doctoral Thesis. Karlstad University Studies 2011:49.

36. Åkesson, Maria (2011), Role Constellations in Value Co-Creation: A Study of Resource Integration in an e-Government Context, Karlstad University, Doctoral Thesis. Karlstad University Studies 2011:36.

37. Nfuka, Edephonce N. (2012), IT Governance in Tanzanian Public Sector Organisations, Department of Computer and Systems Sciences, Stockholm University, Doctoral Thesis.

38. Larsson, Anders Olof (2012), Doing Things in Relation to Machines: Studies on Online Interactivity, Department of Informatics and Media, Uppsala University, Doctoral Thesis.

39. Andersson, Bo (2012), Harnessing Handheld Computing: Framework, Toolkit and De-sign Propositions, Lund University, Doctoral Thesis.

40. Erixon, Cecilia (2012), Information System Providers and Business Relationships: A Study on the Impact of Connections, Mälardalen University, Doctoral Thesis.

41. Svensson, Martin (2012), Routes, Routines and Emotions in Decision Making of Emer-gency Call Takers, Blekinge Institute of Technology, Doctoral Dissertation Series No. 2012:04.

42. Svensson, Ann (2012), Kunskapsintegrering med informationssystem I professionsorien-terade praktiker, Institutionen för tillämpad IT, Göteborgs universitet, Doktorsavhandling.

43. Pareigis, Jörg (2012), Customer Experiences of Resource Integration: Reframing Ser-vicescapes Using Scripts and Practices, Karlstad University, Doctoral Thesis. Karlstad University Studies 2012:38.

44. Röndell, Jimmie (2012), From Marketing to, to Marketing with Consumers, Department of Business Studies, Uppsala University, Doctoral Thesis No. 155.

45. Lippert, Marcus (2013), Communities in the Digital Age: Towards a Theoretical Model of Communities of Practice and Information Technology, Department of Business Studies, Uppsala University, Doctoral Thesis No. 156.

46. Netz, Joakim (2013), Diffusa spänningar eller spännande tillväxt? Företagsledning i ti-der av snabb förändring, Mälardalens högskola, Doktorsavhandling nr 135.

47. Thorén, Claes (2013), Print or Perish? A Study of Inertia in a Regional Newspaper In-dustry, Karlstad University, Doctoral Thesis. Karlstad University Studies 2014:10 (Ny uppl.).

23

Stockhult, Helén (2013), Medarbetare i dialog: en studie om viljan att göra mer än det formellt förväntade, Örebro universitet, Örebro Studies in Business Dissertations, 4.

48. Mihailescu, Daniela (2013), Explaining the Use of Implementation Methodology in En-terprise Systems Implementation Context: A Critical Realist Perspective, Lund University, Doctoral Thesis.

49. Ghazawneh, Ahmad (2012), Towards a Boundary Resources Theory of Software Plat-forms, Jönköping International Business School, Doctoral Thesis.

50. Shams, Poja (2013), What Does it Take to Get your Attention? The Influence of In-Store and Out-of-Store Factors on Visual Attention and Decision Making for Fast-Moving Con-sumer Goods, Karlstad University, Doctoral Thesis. Karlstad University Studies 2013:5.

51. Osowski, Dariusz (2013), From Illusiveness to Genuineness: Routines, Trading Zones, Tools and Emotions in Sales Work, Department of Business Studies, Uppsala University, Doctoral Thesis No. 160.

52. Höglund, Linda (2013), Discursive Practises in Strategic Entrepreneurship: Discourses and Repertoires in Two Firms, Örebro University, Doctoral Thesis.

53. Persson Ridell, Oscar (2013), Who is the Active Consumer? Insight into Contemporary Innovation and Marketing Practices, Department of Business Studies, Uppsala University, Doctoral Thesis.

54. Kask, Johan (2013), On business relationships as Darwinian systems: An exploration into how Darwinian systems thinking can support business relationship research, Örebro University, Doctoral Thesis.

55. Paulsson, Wipawee Victoria (2013), The Complementary Use of IS Technologies to Sup-port Flexibility and Integration Needs in Budgeting, Lund University, Doctoral Thesis.

56. Kajtazi, Miranda (2013), Assessing Escalation of Commitment as an Antecedent of Non-compliance with Information Security Policy, Linnaeus University, Doctoral Thesis.

57. Hasche, Nina (2013), Value Co-Creating Processes in International Business Relation-ships: Three empirical stories of co-operation between Chinese customers and Swedish suppliers, Örebro University, Doctoral Thesis.

58. Pierce, Paul (2013), Using Alliances to Increase ICT Capabilities, Lund University, Doc-toral Thesis.

59. Mansour, Osama (2013), The Bureaucracy of Social Media: An Empirical Account in Organizations, Linnaeus University, Doctoral Thesis.

60. Osmonalieva, Zarina (2013), Factors Determining Exploitation of Innovative Venture Ideas: A study of nascent entrepreneurs in an advisory system, Mälardalen University, Doctoral Thesis.

24

61. Holmberg, Nicklas (2014), The Purity of Separation of Concerns: The Service Oriented Business Process - a Design Approach for Business Agility, Lund University, Doctoral Thesis.

62. Poth, Susanna (2014), Competitive Advantage in the Service Industry. The Importance of Strategic Congruence, Integrated Control and Coherent Organisational Structure: A Lon-gitudinal Case Study of an Insurance Company, Department of Business Studies, Uppsala University, Doctoral Thesis.

63. Safari, Aswo (2014), Consumer Foreign Online Purchasing: Uncertainty in the Con-sumer-Retailer Relationship, Department of Business Studies, Uppsala University, Doc-toral Thesis.

64. Sandberg, Johan (2014), Digital Capability: Investigating Coevolution of IT and Busi-ness Strategies, Umeå University, Doctoral Thesis.

65. Eklinder Frick, Jens (2014), Sowing Seeds for Innovation: The Impact of Social Capital in Regional Strategic Networks, Mälardalen University, Doctoral Thesis.

66. Löfberg, Nina (2014), Service Orientation in Manufacturing Firms: Understanding Chal-lenges with Service Business Logic, Karlstad University, Doctoral Thesis. Karlstad Uni-versity Studies 2014:30.

67. Gullberg, Cecilia (2014), Roles of Accounting Information in Managerial Work, Depart-ment of Business Studies, Uppsala University, Doctoral Thesis No. 171.

68. Bergkvist, Linda (2014), Towards a Framework for Relational-Oriented Management of Information Systems Outsourcing: Key Conditions Connected to Actors, Relationships and Process, Karlstad University, Doctoral Thesis. Karlstad University Studies 2014:31.

69. Tavassoli, Sam (2014), Determinants and Effects of Innovation: Context Matters, Ble-kinge Institute of Technology, Doctoral Thesis No. 2014:10.

70. Högström, Claes (2014), Fit In to Stand Out: An Experience Perspective on Value Crea-tion, Karlstad University, Doctoral Thesis. Karlstad University Studies 2014:44.

71. Jansson, Tomas (2015), Agila projektledningsmetoder och motivation, Karlstads univer-sitet, Doctoral Thesis. Karlstad University Studies 2015:9.

72. Ryzhkova, Natalia (2015), Web-Enabled Customer Involvement: A Firms’ Perspective, Blekinge Institute of Technology, Doctoral Thesis.

73. Sundberg, Klas (2015), Strategisk utveckling och ekonomistyrning: Ett livscykelperspek-tiv. Företagsekonomiska institutionen, Uppsala universitet, Doctoral Thesis No. 173.

74. Nylén, Daniel (2015), Digital Innovation and Changing Identities: Investigating Organi-zational Implications of Digitalization, Umeå University, Doctoral Thesis.

75. Chowdhury, Soumitra (2015), Service Logic in Digitalized Product Platforms: A Study of Digital Service Innovation in the Vehicle Industry, Gothenburg University, Doctoral Thesis.

25

76. Jogmark, Marina (2015), Den regionala transformationsprocessens sociala dimension. Karlskrona 1989-2002, Blekinge Tekniska Högskola, Doctoral Thesis.

77. Sundström, Angelina (2015), Old Swedish Business in New International Clothes: Case Studies on the Management of Strategic Resources in Foreign-Acquired Swedish R&D Firms, Mälardalen University, Doctoral Thesis.

78. Öbrand, Lars (2015), Information Infrastructure Risk: Perspectives, Practices & Tech-nologies, Umeå University, Doctoral Thesis.

79. Brozović, Danilo (2016), Service Provider Flexibility: A Strategic Perspective, Stock-holm University, Doctoral Thesis.

80. Siegert, Steffi (2016), Enacting Boundaries through Social Technologies: A Dance be-tween Work and Private Life, Stockholm University, Doctoral Thesis.

81. Linton, Gabriel (2016), Entrepreneurial Orientation: Reflections from a Contingency Perspective, Örebro University, Doctoral Thesis.

82. Akram, Asif (2016), Value Network Transformation: Digital Service Innovation in the Vehicle Industry, Department of Applied Information Technology, Chalmers University of Technology and University of Gothenburg, Doctoral Thesis.

83. Hadjikhani, Annoch (2016), Executive Expectation in the Internationalization Process of Banks: The Study of Two Swedish Banks Foreign Activities, Department of Business Studies, Uppsala University, Doctoral Thesis No. 177.

84. El-Mekawy, Mohamed (2016), From Theory to Practice of Business-IT Alignment: Bar-riers, an Evaluation Framework and Relationships with Organizational Culture, DSV, Stockholm University, Doctoral Thesis.

85. Salavati, Sadaf (2016), Use of Digital Technologies in Education: The Complexity of Teachers’ Everyday Practice, Linnaeus University, Doctoral Thesis.

86. Pashkevich, Natallia (2016), Information Worker Productivity Enabled by IT System Us-age: A Complementary-Based Approach, Stockholm Business School, Stockholm Univer-sity, Doctoral Thesis.