An empirical examination of therelationship between scientists’work environment and researchperformance

James C. Ryan1 and John Hurley2

1Department of Human Resource Management, Massey University, Private Bag 11222,Palmerston North, New Zealand. j.c.ryan@massey.ac.nz2Dublin City University Business School, DCU, Glasnevin, Dublin 9, Ireland

This research is an empirical examination of the relationship between organizational

characteristics and scientific research effectiveness. A sample of research-active scientists

(N¼ 295) from 25 biological and chemical science university research departments took part

in this study. Data were collected using the Organizational Culture Survey (Glaser et al.,

1987), which measures six organizational characteristics of the research environment.

Organizational characteristics are analysed across a measure of departmental research

performance. Results support the hypothesis that specific characteristics of the organizational

environment are related to research performance. The implications of these findings for the

existing literature and the future management and organization of scientific research

departments are discussed.

1. Introduction

Scientific research makes a fundamental con-tribution to the medical, technological, envir-

onmental, and social advancement of the humanrace. That the scientific enterprise is important isreflected in the extent to which governmentsallocate human and financial resources to it.Despite the obvious influence and importance ofscience to national economies and society, ourunderstanding of the human conditions best sui-ted to its continuing success within the context ofthe organization is limited (Hurley, 1997).

The span of activities that take place under theconceptual umbrella of ‘science’ is vast. Conse-quently, it is necessary to identify more clearly the‘type’ of science that is under examination in thecurrent study. There are obvious and notable

differences between the nature and characteristicsof the resources, personnel, methods, manage-ment, and outcomes of the two extremes of theresearch spectrum from basic or knowledge gen-eration scientific research to more applied routinescientific work. For this reason it is inappropriate,from a research perspective, to view them under asingle heading. The current study focuses morespecifically on scientific research of a basic nature.That is, research directed towards the discoveryand creation of new knowledge is distinct fromapplied research that is directed towards thedevelopment and application or testing of existingknowledge. The logic and reasoning behind thisdecision is simple. In a chronological analysis ofscientific advancement, the discovery must, bynecessity, come before the application and devel-opment of new knowledge. Therefore, discovery,

R&D Management 37, 4, 2007. r 2007 The Authors. Journal compilation r 2007 Blackwell Publishing Ltd, 3459600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 02148, USA

or new knowledge creation, is of primary impor-tance, for without it, development and refinementcannot occur. The primary importance of basicscientific research in the development of theinternational competitiveness of nations in tech-nology and science fields is also highlighted byLim and Song (1996). Furthermore, contempor-ary investigations into the nature and organiza-tion of scientific effectiveness have tended to focuson applied scientific research. Consequently, em-pirical research on the effectiveness of basicscientific research has been neglected. Further-more, research is badly needed, which is groundedin actual scientific research, using measurableoutcomes, rather than relying on simulation,student samples, or opinion-based measures.

Using a categorization developed by Wilts(2000), research institutes can be classified underthree headings Service Providers, Research Con-tractors, and Knowledge Seekers.

In this study, scientific research refers primarilyto research conducted in organizations classifiedby Wilts (2000) as knowledge seekers. That is,research concerned with the production and crea-tion of new knowledge. Specifically, the researchorganizations under examination are university-based research departments that are relativelyautonomous in their decision-making processes,and allow scientists working within them reason-able latitude in the selection and approach to theirfields of research.

2. Literature review

A review of the literature on studies conducted onthe organizational environment of research, iden-tifies several attempts to model the organization ofscientific research as well as a number of identifi-able organizational variables believed to influencethe performance of scientific research.

One of the earliest attempts to model andexplain the administration of modern science wasconducted by Glueck and Thorp (1971). Duringthe course of compiling an annotated bibliographyof research on the organization and administra-tion of scientific research, they developed a generalmodel of the management of research.

Glueck and Thorp (1971) identified the ‘Man-agement Research Process’ as a collection oforganizational characteristics that include factorssuch as preparation and planning, organization/coordination, control/conflict, organization cli-mate, reward structure, and evaluation.

UNESCO (1979) conducted one of the largeststudies on the organizational environment ofscience. The first round of this ‘InternationalComparative Study on the Organization andPerformance of Research Units’ (ICSOPRU) ex-amined the performance-effectiveness of scientificresearch units in six European countries (Austria,Belgium, Finland, Hungary, Poland, and Swe-den). In all, 1222 research units were examined.The structural equation models (SEM) of re-search and development effectiveness constructedfrom the UNESCO (1979) data provide usefuldescriptions of the interactions that take placewithin the research unit. The relative strength ofrelationships between the variables (such assupervision, group climate, and R&D effective-ness) examined in the UNESCO (1979) studyvaried from statistically weak to moderate. Whilethe models derived from this research are ofconsiderable interest, the lack of a theoreticalframework through which to view identified re-lationships limits the usefulness of results.

Thamhain and Wilemon’s (1987) developmentof a systems model of research team performanceillustrates how resources and objectives are trans-ferred into results and team characteristicsthrough the influence of various drivers andbarriers. Drivers include factors such as stimulat-ing work, freedom, communication, good inter-personal relations, and proper planning, whilebarriers include factors such as unclear objectives,conflict, lack of commitment, poor communica-tion, and differing interests (Thamhain and Wi-lemon, 1987). Such concepts have been morerecently described as energy-generating or en-ergy-draining activities of the laboratory environ-ment and found to be related to job satisfactionand organizational culture (Dhawan et al., 2002).

An alternative approach to understanding theinfluence of the immediate environment on scien-tific research comes from the field of creativityresearch. Amabile suggests that ‘the ultimatedriving force behind all of scientific progress’ ishuman creativity (Amabile, 1994). As one of theforemost writers and researchers in the area ofcreativity, Amabile proposes a componentialmodel of scientific creativity. By focusing on theeffect of social factors on scientists’ motivationand on the effect of motivation on creativity, thismodel recognizes the influential nature of theorganizational environment. Experimental sup-port for Amabile’s work is generally based onstudent samples in contrived research settings(Conti et al., 1996). This methodological limita-tion poses concerns over the usefulness and

James C. Ryan and John Hurley

346 R&D Management 37, 4, 2007 r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

validity of the componential model of creativity inexplaining scientific performance in the ‘realworld’ of the research organizations, but doessucceed in, once again, highlighting the importantlink between the research environment and re-search outcomes.

The theoretical model of scientific discoverypresented by Hurley (1997) highlights the impor-tance of the suitable combination or ‘fit’ of orga-nizational variables with individual variables. Thistheoretical model suggests that the suitabilityof the scientist’s ability and personality, coupledwith the suitability of the organizations resourcesand dynamics, determines the likelihood of adiscovery taking place. Hurley’s (1997) modelhas a strong theoretical base, but is lacking inany empirical support. By his own admission, themodel is ‘essentially a speculation on the nature ofreality’ (Hurley, 1997, p. 148). Yet despite thisweakness, the model offers an interesting theore-tical perspective on the discovery process, andprovides suggestions for future research in thisarea.

In addition to attempts to model the nature of theorganization of scientific research, researchers inthis field have also identified specific characteristicsof the research environment that relate to researchperformance or proxies of research performancesuch as the motivation to perform research.

Baumgartel’s (1956) research on the attitudesand motivations of scientists identifies leadershipstyle as a characteristic that is related to scientists’job satisfaction and their motivation to engage inresearch. Argyris (1968) identifies nine basic vari-ables that influence the effectiveness of the researchorganization. These include organizational struc-ture, technology, administrative controls, humancontrols, leadership styles, interpersonal relationsand communication, group effectiveness, inter-group relations, and norms of the living system.

Individual organizational factors such as leader-ship/supervision, cooperation, and participation(Keller et al., 1996) have been demonstrated toaccount for variation in publication quantity, andquality in some samples. Influencing styles ofproject managers have also been related to a varietyof indicators of project success (Thamhain andGemmill, 1974), while the social context in whichscience takes place and the characteristics andstates of communication processes have beenshown to relate to eminence and research produc-tivity (Simonton, 1992; Pruthi and Nagpaul, 1994).

In a review of the literature on the characteristicsof the productive research environment, Bland andRuffin (1992) identify 12 common and recurring

characteristics. These include the existence of cleargoals that serve a coordinating function, a researchemphasis, a distinctive organizational culture, apositive group climate, assertive participative gov-ernance, a decentralized organizational structure,the size, age and diversity of the group, the rewardstructure of the organization, frequent and opencommunication, a well-developed recruitment andselection process, adequate resources, and finallyeffective leadership.

Factors identified by Hurley (1997) includemany of those identified by Bland and Ruffin(1992) with the addition of others such as teamautonomy, group morale, group membership,and supervision of the research team. Utilizingan ethnographic approach, Mouly and Sankaran(1998) identified several factors, such as excessivebureaucracy, lenient selection and recruitment,lack of team identity, lack of autonomy, poorinterpersonal relations, and authoritarian andapathetic supervision, which characterize the dys-functional research environment.

Utilizing a data set from a further round of theUNESCO (1979) study, Chawla and Singh (1998)identified several organizational characteristicsrelated to research productivity. These organiza-tional characteristics were chosen for their abilityto describe the effectiveness of research organiza-tions in terms of management practices andresources, and were used to identify five distincttypes of research organizations with varying levelsof effectiveness. The list of characteristics itselfincludes leadership, work environment, policies,planning, communication, and resources.

Table 1 presents a summary of key researchthat has examined a variety of organizationalcharacteristics, believed to influence the perfor-mance-effectiveness of scientific research.

The summary of organizational characteristicspresented in Table 1 highlights the degree of con-ceptual similarity among many of the factorsidentified across the various studies summarizedhere. It also highlights the complexity of ap-proaches and various methodological perspectivesthat have been taken into consideration whileexamining the aspects of the scientific work envir-onment and poses questions as to how to empiri-cally operationalize and measure these concepts.

3. Measuring the organizationalenvironment and research performance

The most intuitive and logical approach to themeasurement of those organizational characteristics

Relationship between scientists work environment and research performance

r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

R&D Management 37, 4, 2007 347

Table1.Summary

ofcharacteristics

oftheresearchenvironmentidentified

asinfluencingscientificresearcheffectiveness

Argyris

(1968)

Baumgartel

(1956)

Glueckand

Thorp

(1971)

Thamhain

and

Wilem

on(1987)

Blandand

Ruffin(1992)

Hurley

(1997)

Mouly

and

Sankaran

(1998)

Chawla

and

Singh(1998)

Organizational

structure

Leadership

Preparation

andplanning,

Leadership

Cleargoals

Goodselection

processes

Lenience

inrecruitment

processes,

professional

mediocrity

Leadership

Technology

Freedom

indecision

making

Organization/

coordination

Jobcontent

Research

emphasis

Highmorale

Poorself-image

Work

environment

Administrative

controls

Control/

conflict

Personalgoals

Distinctive

organizational

culture

Positivegroup

mem

ber-ship

andsupervision

Lack

ofteam

identity

androle

clarity

Policies

Human

controls

Organization

clim

ate

Work

environment

Adequate

recruitment

andselectionprocess

Autonomyof

work

teams

Lack

ofinstitutional

autonomyin

project

selection

Communication

Leadership

styles

Evaluation

Assertive

participative

governance

Supportive

organizational

culture

Excessive

bureaucracy,

apathetic

attitudes

Resources

Interpersonal

relationsand

communication

Decentralized

organizational

structure

Effective

communication

Strained

interpersonal

relationships

Planning

Group

effectiveness

Sizeageand

diversity

of

thegroup

Effective

leadership

styles

Apathetic

supervision,

authoritarian-ism

within

theinstitutes

hierarchy

Inter-group

relations

Appropriate

reward

structure

Norm

softhe

livingsystem

Frequentand

open

communication

Positivegroup

clim

ate

Adequate

resources

Effective

leadership

James C. Ryan and John Hurley

348 R&D Management 37, 4, 2007 r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

believed to influence the performance of scientists isto integrate existing measures that have shownrelationships between the environment of scienceand scientific research effectiveness. However, asevident from the examination of research on theorganization of scientific research, presented in theprevious paragraphs and in Table 1, the methodo-logical disparity and lack of easily identifiable quan-titative instruments to operationalize many of theorganizational characteristics identified in the litera-ture do not facilitate this approach.

An alternative perspective is to view the orga-nization more holistically, and examine the pos-sibility of applying existing measures oforganizational characteristics to the study of thespecific organization of science. Such approacheshave been taken in the study of organizations; inparticular, such an approach is evident in thestudy of the overriding concept of organizational‘culture’.

The concept and term organizational culture inthe academic literature can be traced back toPettigrew (1979). The concept of culture began as,what has been described as, a management ‘fad’but its value was quickly recognized and it hasbecome a primary area of organizational research(Hofstede et al., 1990) spawning many definitions,interpretations, and measurement instruments. Ar-gument over the actual definition and conceptuali-zation of culture still exists, but several usefuldefinitions have been put forward in the literature.Deal and Kennedy (1982) describe culture as ‘theway things are done around here’. Such a definitionincludes wide aspects of organizational memberbehaviour and organizational processes. A morespecific definition by Schein (1983) refers to the‘pattern of basic assumptions that a given grouphas invented, discovered, or developed in learningto cope with its problems of external adaptationand internal integration’. An alternative definitionprovided by Koberg and Chusmir (1987) explainsthe concept of organizational culture as ‘a systemof shared values and beliefs that produce norms ofbehaviour and establish an organizational way oflife’. Recognition of the variety of and difference inthe definition of organizational culture is impor-tant, because subtleties within these definitionsdrive how various researchers in the area of orga-nizational culture operationalize the concept ofculture.

Two broad components of organizational cul-ture that exist in the literature, and that can beidentified in definitions of the concept, include theexamination of culture as (a) a set of cognitions(i.e.) values and beliefs, and (b) the outcome of

these values and beliefs in the form of observablebehavioural components. It is the latter aspectand perspective of culture that is of possibleinterest to the current study. The examination ofbehavioural norms across various dimensions ofthe organization relate directly to what can alsobe viewed as the human characteristics of theresearch organization. Essentially such measuresoperationalize culture by examining behaviouralnorms across various dimensions of organiza-tional functioning. Such dimensions of organiza-tional functioning can include factors such asleadership, communication, teamwork, or othercharacteristics of the research organization iden-tified in this literature review.

In this instance the Organizational CultureSurvey (OCS) (Glaser et al., 1987) was chosento measure characteristics of the organizationalenvironment. The OCS measures six dimensionsof organizational functioning including Team-work, Morale, Information Flow, Involvement,Supervision, and Meetings. Immediate links canbe drawn between many of these concepts and thecharacteristics of the research organization iden-tified in a review of the literature,. for example,the characteristic of morale referred to by Hurley(1997), or the characteristic of supervision re-ferred to by Mouly and Sankaran (1998). Inaddition to this, a detailed examination of theitems that comprise each of the factors of the OCSshow that the labelling of some of these factorsbelies their obvious conceptual relationship tomany of the organizational characteristics identi-fied in the literature summarized in Table 1. Anintellectual synthesis of these concepts and thefactor items of the OCS highlight these concep-tual similarities. For example, the OCS factor ofTeamwork refers to the degree to which membersof a team are able to communicate effectively,resolve disagreement, and constructively confrontproblems. An examination of the concepts con-tained in Table 1 show the concept of teamworkas measured by the items of the OCS to be relatedto several of the concepts contained within thistable. For example good interpersonal relationsand communication are identified by Argyris(1968) as organizational factors that influencethe effectiveness of research organizations. Theseconcepts relate directly to the concept of team-work as measured by the OCS. There is alsoconceptual overlap between the OCS concept ofteamwork and Argyris’ concept of intergrouprelations. Similarly, the positive group climatereferred to by Bland and Ruffin (1992) is synon-ymous with the OCS concept of teamwork.

Relationship between scientists work environment and research performance

r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

R&D Management 37, 4, 2007 349

Hurley’s (1997) identification of positive groupmembership also shows considerable overlap withthe OCS concept of teamwork, as does the con-cept of strained interpersonal relationships identi-fied by Mouly and Sankaran (1998). The conceptof the work environment as identified in theUNESCO (1979) study and by Thamhain andWilemon (1987) similarly shows considerable con-ceptual overlap with the OCS concept of team-work. Teamwork, as operationalized by the OCSis also related to the concept of control/conflict, asidentified by Glueck and Thorp (1971). A detailedexamination of the items that comprise the OCSfound that other factors were similarly related tomany of the concepts listed in Table 1.

In the current study, the measurement of re-search performance across participating researchdepartments was facilitated by an existing andreadily accessible measure known as the ResearchAssessment Exercise (RAE). The RAE is a UKgovernment-initiated research performance mea-sure that is designed to enable higher educationauthorities to distribute public research fundsselectively to research departments in tertiaryinstitutions on the bases of their research quality.The assessment, which takes place every 4–5years, provides research-quality ratings for aca-demic departments across all academic disci-plines, and on the bases of these ratingsdistributes approximately UKd1 billion a yearHigher Education and Research Opportunitiesin the United Kingdom (HERO, 2002a).Thequality ratings themselves, which range from 1to 5*, are based on detailed submissions from theparticipating departments.

While the RAE is certainly not a perfect orflawless evaluation procedure, it does provide arelatively thorough and detailed evaluation of thequality of research across a large number ofresearch departments in UK universities, basedon the information contained within each sub-mission. This information is similar in nature tothat employed in the evaluation of research per-formance in the UNESCO (1979) study in that itincludes a variety of dimensions in the analysis ofperformance. As such the RAE offers the mostreadily available and comprehensive evaluation ofresearch effectiveness suitable for use in the cur-rent study. It also offers the additional benefit ofallowing for the stratified sampling of scientistsfrom research departments of varying levels ofresearch effectiveness.

Following a review of the literature and theidentification of a suitable measure, a hypothesiswas formulated.

Hypothesis 1 : Research departments that exhibitmore-favourable organizational char-acteristics will exhibit correspond-ingly higher levels of researchperformance.

4. Methodology

The first stage of sample selection was the identi-fication of a population of scientists from whichto draw a participant sample. In the currentstudy, the population of scientists was limited toresearch-active scientists working in the UK uni-versity departments who had taken part in theUK’s RAE, and were working in research depart-ments in the fields of biological and/or chemicalresearch. A list of such departments and theirRAE scores was collected from data availablefrom the website of the HERO (2002b).

Departments were then separated by RAEgrade from 5* to 1. The lists of departments ineach grade of the RAE from 5* through to 5, 4,3a, 3b, 2, and 1 were then randomized. Contactwas then made with the heads of department ineach list starting at the top of each list. The headsof departments were contacted via email andphone and the nature of the study was explainedto them, while access to their staff was requested.This step was repeated down the lists of depart-ments until sufficient sample size was procured.At an early stage in the research it became clearthat insufficient numbers of responses in the 1, 2,and 3b categories would be attained to allow for auseful examination of departments with thesegrades. Consequently; they were excluded fromfurther investigation. The reasons for this areelaborated on further in the discussion. In total,727 research scientists were requested to partici-pate in the study, out of which 295 responded,giving a response rate of 30.6%. Participation inthe study at both the departmental and individuallevels was voluntary, and confidentiality wasassured. Of the total of 295 respondents, 243(82.4%) were male and 52 (17.6%) were female.

A paper and pencil copy of the OrganizationalCulture Survey (Glaser et al., 1987) was preparedfor each participant. A questionnaire was alsoincluded to collect basic biographical informa-tion. Data were collected from individual scien-tists whose departments agreed to participate inthe study, in one of two ways. In the first instance,where possible the researcher visited the depart-ment at an agreed date and time. During this visit,

James C. Ryan and John Hurley

350 R&D Management 37, 4, 2007 r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

the booklet of questionnaires was distributed inperson to potential participants, either directly attheir office/laboratory or left for them in the internalmail system. Participants were requested to completethe booklet and return it to the researcher by the endof the day or mail the completed booklet to theresearcher at a time suitable to them. In the secondinstance, where an onsite visit to the department wasnot favoured by the participating department a copyof the booklet of questionnaires was mailed to thework addresses of all research-active staff in theparticipating department. Included with the ques-tionnaire booklet was a self-addressed envelope toassist participants in returning the completed ques-tionnaire to the researcher. Data from the completedquestionnaires were then inputted into an SPSS filefor analysis. The data from a haphazard selection of30 completed questionnaires were then rechecked toensure accuracy in data entry.

The analysis of data from the current study wasconducted in two stages. The first stage examinedthe scale reliabilities of the OCS (1987). Thesecond stage examined the primary hypothesispresented previously, examining the relationshipbetween characteristics of the organizational en-vironment and research performance.

5. Results

Before analysis of the primary relationship pro-posed in the hypothesis, an examination of thescale reliabilities of the OCS was conducted. Theinstrument is reported to be well constructed withprincipal-component factor analysis yielding asix-factor solution with eigenvalues in excess of1.0. In addition, no items in the scales have anitem loading less than 0.56 (Rubin et al., 1994).Glaser et al. (1987) reported scale reliabilitiesranging from 0.63 to 0.91 using Cronbach’s a.Independent examinations of the Cronbach a for

each subscale were conducted on the results fromthe current study. Table 2 lists the Cronbach areliabilities for each subscale of the OCS for thecurrent sample of research scientists.

As we can see from Table 2, the OCS Cronbacha found in the current study mirror quite closelythose reported by Glaser et al. (1987). In generalterms an a coefficient above 0.7 is deemed accep-table and reflects a suitable level of internal con-sistency among scale items. The higher a’s ofTeamwork, Morale and Supervision are not sur-prising as these scales contain the most items (eight,seven and eight, respectively), which increases thelikelihood of higher a’s for internally consistentscales when scale items have a relatively highdegree of shared variance. A visual examinationof the scale items of the OCS also suggests solidface validity and offers support for the dimension-ality of each scale. In short, the OCS seems to offera useful measure of the organisational environmentthat is reliably consistent across time and in differ-ing cultural and occupational contexts.

Before further analysis of the data, an exam-ination of the number of responses for eachcategory of the RAE was conducted. For easeof data entry and analysis, a slight readjustmentof the labelling of RAE scores was made. Forexample, 5*-ranked departments were recordedas ‘6’. The RAE rankings of 5 and 4 remained thesame while 3a rankings were relabelled ‘3’. Asmentioned previously, both the small number ofdepartments in categories 3b, 2, and 1 and theirsmall size resulted in their exclusion from analy-sis. The number of respondents in each depart-mental category and the number of departmentsfrom which they were drawn are presented inTable 3.

The primary hypothesis present below predictsa linear relationship between characteristics of theresearch environment and research performance,with performance improving as organizationalcharacteristics become more favourable.

Table 2. OCS subscale Cronbach a reliabilities forsample of participants in current study (N¼ 295)

Subscale Cronbach a

Teamwork 0.90Morale 0.91Information flow 0.78Involvement 0.78Supervision 0.88Meetings 0.75

OCS, organizational culture survey.

Table 3. Number of participating departments andscientists in each RAE category

RAE category Number ofparticipatingdepartmentsin each RAEcategory

Total number ofparticipatingscientists ineach category N

3 8 774 8 705 6 696 3 79

RAE, research assessment exercise.

Relationship between scientists work environment and research performance

r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

R&D Management 37, 4, 2007 351

Hypothesis 1 : Research departments that exhibitmore-favourable organizational char-acteristics will exhibit correspond-ingly higher levels of researchperformance.

The testing of this hypothesis requires theexamination of the relationship between depart-mental scores for organizational characteristicsand departmental score for research performance.Departmental scores for each of the six factors ofthe OCS were compiled by aggregating responsesfrom participating scientists within each depart-ment. As the measure of research performanceused in the current study (the RAE) is essentiallyan ordinal ranking, a one-tailed Spearman’ rwas conducted for each of the OCS factorsand RAE rankings. Results are presented inTable 4.

Results identify significant positive correlationsbetween all measured characteristics of organiza-tional environment and the performance ratingsfor each department, and consequently confirmthe primary hypothesis that ‘research departmentsthat exhibit more-favourable organizational char-acteristics will exhibit correspondingly higher le-vels of research performance’.

6. Discussion

Before discussing the implications of the currentfindings, it is important to note its limitations andview its findings cautiously. The identification ofsignificant correlations between organizationalfactors and a measure of research performanceis of significant importance to the literature on themanagement of science, but does not prove orimply a causal relationship between the variablesunder examination. The structure of the currentstudy can neither show temporal order in thevariables nor can it eliminate all alternative ex-planations. However, the logical association be-tween these factors is well presented in theliterature reviewed previously.

Another weakness of the current study is thelack of a complete range of research performanceacross departments. As mentioned previously, thedepartmental measure of research performanceused in the current study were the departmentalRAE scores for UK university departments. Thesescores range from 5* to 1. However, in practice thenumbers of 1-and-2 ranked departments are few innumber. This is likely due to the onerous nature ofthe RAE process and the probability that depart-ments that surmised their RAE score to be a 1 or 2would deem the funding associated with such arank to be incommensurate to the effort exerted onthe RAE submission. An examination of the size ofsuch departments also reveals the general trendthat low-scoring departments generally comprisefewer staff. This has implications for the degree towhich a limited number of respondents can beviewed as providing an accurate representation ofthe characteristics of the organizational environ-ment. These factors contributed to the studies’inability to collect sufficient data to provide acomplete range of research departments from thevery highest ranking to the lowest ranking, whichdoes limit the ability to generalize results.

The voluntary nature of participation must alsobe taken into account when examining the currentresults. It is possible that the responses gatheredmay represent the experiences of a subpopulationof research scientist within each department andnot accurately reflect the general experience of allscientists within each department. Of further noteis the use of the RAE rankings as a measure ofresearch performance at the departmental level.Attempts at quantifying the measurement ofscientific research vary considerably and includebasic publication counts, citation counts, mea-surement of impact factors, complex bibliometricanalysis of scientific fields, patent counts, funding

Table 4. Summary of Spearman rho correlations forOCS factors and RAE scores for participating researchdepartments

OCS revisedfactor

Correlation with researchassessment exercise(RAE score)

Teamwork Pearson correlation 115*Significance (one-tailed) 0.024N 294

Morale Pearson correlation 0.299**Significance (one-tailed) 0.01N 294

Information Flow Pearson correlation 0.211**Significance (one-tailed) 0.000N 295

Involvement Pearson correlation 0.239**Significance (one-tailed) 0.000N 295

Supervision Pearson correlation 236**Significance (one-tailed) 0.000N 282

Meetings Pearson correlation 0.143**Significance (one-tailed) 0.007N 294

*Correlation is significant at 0.05 level (one-tailed).**Correlation is significant at 0.01 level (one-tailed). OCS,organizational culture survey.

James C. Ryan and John Hurley

352 R&D Management 37, 4, 2007 r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

awards, science medals, etc. What is clear fromcommunications with scientists in various fields isthat there is no one universally acceptable methodof measuring research performance/effectiveness.With this in mind, the RAE can be viewed as aflawed but nonetheless acceptable and indicativemeasure of research performance.

The identification of correlations betweenvariables examined in the current study doessupport the primary hypothesis. Table 4 showscorrelations ranging between 0.155 and 0.299, andthese are statistically significant though notstrong. More-favourable organizational charac-teristics are shown to exist in tandem with higherlevels of research performance. However, whilethe hypothesis was supported, the relative weak-ness of the correlations reported suggests theexistence of other important variables in ourunderstanding of research effectiveness. Some ofthese alternative variables have been empiricallyexamined. The role and importance of develop-mental characteristics, educational environment,individual differences, and other social conditionsare reported by several researchers (Zuckerman,1977; Simonton, 1991, 1992; Hurley, 1997; Feistand Gorman, 1998; Ryan, 2003). Results of thecurrent study add to this literature by providingempirical evidence of the relationship betweenorganizational characteristics and performance,and highlighting the need for a greater under-standing of the organizational and managerialenvironment in which scientists work.

Where empirical evidence has been gathered inthis field of research, it has tended to be of aqualitative nature. This leads to difficulties inconfirming relationships and in many cases stu-dies are irreplicable. However, current findingsprovide empirical evidence of the significant linkbetween the organization and scientific researcheffectiveness (Argyris, 1968; Bland and Ruffin,1992; Chawla and Singh, 1998; Mouly and San-karan, 1998). Concepts contained within Tham-hain and Wilemon’s (1987) systems model ofresearch effectiveness suggest that single organi-zational factors can act as either drivers orbarriers depending on their relative position ona positive to negative continuum. Essentially asingle organizational characteristic, say informa-tion flow, can act as a driver or a barrier depend-ing on whether it is manifested in a positive,efficient, and constructive manner (driver) or ismanifested in a dysfunctional, ineffective, anddestructive manner (barrier). By viewing resultsfrom the current study from the perspective ofThamhain and Wilemon’s (1987) systems theory,

we might suggest that the increasingly positivecharacteristics of Morale, Information Flow,Supervision, Meetings, and Involvement contri-buting to a positive increase in the performance ofresearch conducted with that organizational en-vironment culture do not hinder or act as barriersto research performance.

The findings of this study also offer support forHurley’s (1997) hypothetical model of scientificeffectiveness, confirming the influence of the or-ganization while also recognizing and allowingfor the influence of other non organizational/individual conditions. Hurley suggests that it isthe suitable combination of individual character-istics, such as creativity, persistence, and technicalknowledge, along with organizational character-istics such as those identified in the current studythat ultimately determine the likelihood of successin the scientific endeavour. These results go someway to supporting that hypothesis.

While the current research findings inform ex-isting theory on the organization and managementof science, they also offer cautious practical direc-tion to science managers and practitioners. Thecurrent research suggests that elements of a re-search culture can manifest in tangible, observableorganizational processes and human interactions.

Such interactions include the supervisor–subor-dinate relationship. The current study suggestsa need for research supervisors to be participativein their decision-making processes, to be opento critique, and to be generally facilitative andsupportive of a collaborative environment. Addi-tionally, practical steps can be taken to ensuremeetings occurring within the research environ-ment are focused, purposeful, and facilitate thecontribution of all team members. The currentstudy also shows that involvement in decision-making is related to higher levels of researchperformance. Encouraging the meaningful involve-ment of research personnel can also contributepositively to feelings of meaningfulness and moralein the research setting, promoting a positive teamenvironment and generally improving the researchculture that could in turn promote positive researchoutcomes.

Future research in this area should seek toaccount for both the characteristics of the orga-nizational environment and individual character-istics of the scientists working within the researchenvironment. Such studies could explain morecomprehensively the relative importance andinfluence of the individual and organizationalvariables in determining research performance-effectiveness.

Relationship between scientists work environment and research performance

r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd

R&D Management 37, 4, 2007 353

In conclusion, the findings presented here re-present advancement in our understanding of theway in which the component characteristics of anorganizations’ culture may be related to theperformance of scientists within that organiza-tion. Our findings suggest that the culture of theorganizational environment can be seen to berelated to research performance, and can distin-guish between the lower and higher performers inthe scientific arena. However, the low to moderatecorrelations reported here also suggest the influ-ence of alternative variables, including furtherorganizational and non-organizational variables,in explaining the nature of scientific effectiveness.

References

Amabile, T.M. (1994) The atmosphere of pure work:

creativity in research and development. In Shadish,

W.R. and Fuller, S. (eds), The Social Psychology of

Science. London: The Guilford Press.

Argyris, C. (1968) On the effectiveness of research and

development organisations. American Scientist, 56, 4,

344–355.

Baumgartel, H. (1956) Leadership, motivation and

attitudes in research laboratories. Journal of Social

Issues, 12, 2, 24–31.

Bland, C.J. and Ruffin, M.T. (1992) Characteristics of a

productive research environment: literature review.

Academic Medicine, 67, 385–397.

Chawla, A. and Singh, J.P. (1998) Organizational

environment and performance of research groups: a

typological analysis. Scientometrics, 43, 3, 373–391.

Conti, R., Coon, H. and Amabile, T.M. (1996) Evi-

dence to support the componential model of creativ-

ity: secondary analyses of three studies. Creativity

Research Journal, 9, 4, 385–389.

Deal, T.E. and Kennedy, A.A. (1982) Corporate Cul-

tures: The Rites and Rituals of Corporate Life. Read-

ing, MA: Addison-Wesley.

Dhawan, S.K., Roy, S. and Kumar, S. (2002) Organi-

zational energy: an empirical study of Indian R&D

laboratories. R&D Management, 32, 5, 397–408.

Feist, G.J. and Gorman, M.E. (1998) The psychology

of science: review and integration of a nascent dis-

cipline. Review of General Psychology, 2, 1, 3–47.

Glaser, S.R., Zamanou, S. and Hacker, K. (1987)

Measuring and interpreting organizational culture.

Management Communication Quarterly, 1, 173–198.

Glueck,W.F. and Thorp, C.D. (1971) TheManagement of

Scientific Research. Missouri: University of Missouri.

HERO (2002a) A Guide to the 2001 Research Assess-

ment Exercise. Available at: http://www.hero.ac.uk/

rae/Pubs/other/raeguide.pdf (accessed August 2002).

HERO (2002b) Higher Education and Research Oppor-

tunities in the United Kingdom. Available at: http://

www.hero.ac.uk/research/research_assessment_exer

cise_2485.cfm (accessed August 2002).

Hofstede, G., Neuijen, B., Ohayv, D.D. and Sanders,

G. (1990) Measuring organizational cultures: a qua-

litative and quantitative study across twenty cases.

Administrative Science Quarterly, 35, 286–316.

Hurley, J. (1997) Organisation and Scientific Discovery.

Chichester: Wiley.

Keller, R.T., Julian, S.D. and Kedia, B.L. (1996) A

multinational study of work climate, job satisfaction,

and the productivity of R&D teams. IEEE Transac-

tions on Engineering Management, 42, 1, 48–55.

Koberg, C.S. and Chusmir, L.H. (1987) Organizational

culture relationships with creativity and other job-related

variables. Journal of Business Research, 15, 5, 397–409.

Lim, Y.T. and Song, C.H. (1996) An international

comparative study of basic scientific reserach capa-

city: OECD countries, Taiwan and Korea. Techno-

logical Forecasting and Social Change, 52, 75–94.

Mouly, V.S. and Sankaran, J.K. (1998) The behaviour

of Indian R&D project groups: an ethnographic

study. Advances in Qualitative Research, 1, 137–160.

Pettigrew, A.M. (1979) On studying organizational cul-

tures. Administrative Science Quarterly, 24, 570–581.

Pruthi, S. and Nagpaul, P.S. (1994) Communication

and productivity in scientific research. Journal of

Scientific and Industrial Research, 53, 840–849.

Rubin, R.B., Palmgreen, P. and Sypher, H.E. (1994)

Communication Research Measures: A Sourcebook.

New York: Guilford Press.

Ryan, J.C. (2003) The contribution of cognitive and

organisational psychology. In Hurley, J. (ed.), Scien-

tific Research Effectiveness: The Organisational Dimen-

sion. Dordecht, the Netherlands: Kluwer Academic

Publishers, pp. 197–213.

Schein, E.H. (1983) The role of the founder in creating

organizational culture. Organizational Dynamics, 12,

1, 13–28.

Simonton, D.K. (1991) Career landmarks in science:

individual differences and interdisciplinary contrasts.

Developmental Psychology, 27, 1, 119–130.

Simonton, D.K. (1992) The social context of career

success and course for 2,026 scientists and inventors.

Personality and Social Psychology Bulletin, 18, 452–463.

Thamhain, H.J. and Gemmill, G.R. (1974) Influence

styles of project managers: some project performance

correlates. Academy of Management Journal, 17, 2,

216–224.

Thamhain, H.J. and Wilemon, D.L. (1987) Building high

performance engineering project teams. IEEE Transac-

tions on Engineering Management, 34, 3, 130–137.

UNESCO. (1979) Scientific Productivity: The Effective-

ness of Research Groups in Six Countries. Cambridge:

Cambridge University Press.

Wilts, A. (2000) Forms of research organisations and

their responsiveness to external goal setting. Re-

search Policy, 29, 767–781.

Zuckerman, H. (1977) Scientific Elite. New York: Free

Press.

James C. Ryan and John Hurley

354 R&D Management 37, 4, 2007 r 2007 The Authors

Journal compilation r 2007 Blackwell Publishing Ltd