Image Congruence and theAdoption of Service Innovations

Mirella KleijnenVrije Universiteit Amsterdam

Ko de RuyterUniversity of Maastricht

Tor W. AndreassenNorwegian School of Management BI

In this article, the authors investigate whether the fit be-tween consumers’ image and the image of an innovation,commonly referred to as image congruence, has an impacton customers’attitudes toward innovative services as wellas the intention to use these new services. Moreover, situa-tional influences (i.e., private vs. public, and friends vs.colleagues) are examined. The results of the two experi-ments show that not only does image congruence have asignificant impact on consumer attitudes and the adoptiondecision, but there is also an interaction between imagecongruence and the consumption situation. Contrary tothe authors’expectations, consumers with low image con-gruence are influenced more by their surroundings thanconsumers with high image congruence. It seems that aslong as the fit between the consumer’s self-image and theservice image is high, signals from the consumer’senvironment are overruled.

Keywords: image congruence; consumer adoption; ser-vices innovation; mobile commerce

As a rule of thumb, innovative services typically do notmake it beyond the early market of technology enthusiasts(e.g., Moore 2002). Ample research in the field of market-ing has focused on what makes visionaries and pragma-tists adopt an innovative service, that is, cross the prover-bial “innovation chasm” (Moore 2002). Although theimportance of investigating technology characteristics iswidely acknowledged in the literature (e.g., Davis,Bagozzi, and Warshaw 1989), several authors point to therelevance of studying consumer-related factors as well(e.g., Meuter et al. 2000). Recent studies in both the tech-nology acceptance literature (Venkatesh and Davis 2000)and innovation research (Plouffe, Vandenbosch, andHulland 2001) have emphasized the need to expand tradi-tional adoption theory by including image as a relevantfactor in the adoption process. Whereas these previous

We would like to thank the editor and three anonymous reviewers of the Journal of Service Research for their constructive feedbackallowing us to develop a better paper. We would also like to thank Philipp Graf von Bassewitz for his help in the data collection. All au-thors have contributed equally to the article. Correspondence concerning this article should be addressed to Mirella Kleijnen, VrijeUniversiteit Amsterdam, Faculty of Economics and Business Administration, Department of Marketing, De Boelelaan 1105, 1081 HVAmsterdam, the Netherlands; phone: +31 (0)20 598-9861; fax: +31 (0)20 598-6005; e-mail: mkleijnen@feweb.vu.nl.

Journal of Service Research, Volume 7, No. 4, May 2005 343-359DOI: 10.1177/1094670504273978© 2005 Sage Publications

studies focus solely on the image of the product or serviceinnovation, Schlosser (2002) suggested that the use ofmobile technology is dependent on the consumer’s self-identity, as the use of such technology allows the con-sumer to express who he or she is. This seems especiallyrelevant for the study at hand, which examines a new ser-vice category called mobile commerce. Mobile commercerefers to services that are delivered via wireless tech-nology, by means of mobile devices, such as a mobilephone, and can be consumed anywhere and at any time(Balasubramanian, Peterson, and Jarvenpaa 2002). Theimportance of self-image in the adoption of mobile ser-vices is also illustrated in practice. For example, whenNokia replaced the old 9110 Communicator with theCommunicator 9210i, the advertising focused on the ben-efits to image and status that users would enjoy from asso-ciation with the new model. Rather than communicatingtechnology features, Nokia focused on communicatingperceived benefits, such as the match between the servicesthe Nokia communicator delivers and the consumer’s life-style or image. Why this change in communication strat-egy? Research has found that consumers often preferbrands and stores with images that are consistent with theirself-image (e.g., Sirgy 1982, 1985). In the literature, thisphenomenon is referred to as the image congruencehypothesis (e.g., Grubb and Grathwohl 1967). Becauseconsumers may buy brands that can create and enhanceself-image (i.e., how a consumer actually sees himself orherself), we suggest that image congruence plays a crucialrole in the consumer adoption process. The serious needfor new knowledge in this area is reflected in the Market-ing Science Institute’s research priorities,1 which definethe interplay between users and image as a subtopic of sig-nificant interest.

In this article, we examine image and innovation adop-tion in the context of consumer usage situations. Severalstudies in the field of consumer behavior illustrate theimpact of the consumer’s surroundings on decision mak-ing. For example, Graeff (1997) suggested that certainproducts or brands are consumed in a particular situationbecause the consumption of these items bestows on theuser a certain image. Therefore, we suggest that the con-sumer’s usage situation has an effect on the decision ofwhether to adopt mobile services. Previous researchregarding image has investigated private versus publicbrands and/or products (e.g., Schmitt and Shultz 1995),where private relates to situations in which consumers areisolated and public to situations in which consumers are inproximity to others. This seems particularly relevant whenstudying the adoption of mobile services. Due to the ubiq-uity of mobile services (Balasubramanian, Peterson, and

Jarvenpaa 2002), they can be consumed in a number ofdifferent consumption situations. The customer is in com-plete control of the decision as to when, what, and particu-larly where services are consumed, implying that theseservices can be used in private as well as public situations.

The objectives of this article, therefore, are (a) to testthe image congruency theory in the adoption of an innova-tive service and (b) to investigate the effects of consump-tion situations on attitude and intention to adopt an innova-tive service. The article is organized as follows. First, wepresent an overview of the literature regarding image. Sec-ond, we discuss our first experiment that focuses on theeffects of image congruence and the effects of private ver-sus public situations. A second experiment is then con-ducted to confirm and extend the findings of the firstexperiment. Finally, the implications of the findings arediscussed.

LITERATURE REVIEW

As image is central to this study, and because it has beendiscussed and studied in the services marketing and in theadoption literature, the following section provides a briefsummary of relevant literature.

Image, Decision Making, and Adoption

According to Olson (1972), an object’s image can be anextrinsic information cue and may influence customerchoice and adoption. In this respect, image becomes anissue of attitudes and beliefs affecting awareness and rec-ognition (Aaker 1991), customer satisfaction, and con-sumer behavior (Fornell 1992). As customers receive newinformation or have new experiences pertaining to anobject, this information will act as an input to either main-tain or update (be it positively or negatively) their percep-tions and attitudes toward the object.

In adoption research, image has been implicitly includedin the factor of relative advantage. For example, Rogers(1995) indicated that one aspect of relative advantage ismost certainly image. However, Tornatzky and Klein(1982), contrary to previous studies, found support forseparating the constructs of image and relative advantage.Similarly, Chau (1996) has recognized that the desire togain social status might be one of the most important moti-vations for a consumer to adopt an innovation. Follow-ing this line of reasoning, we argue that image shouldbe recognized as a construct in itself, distinctly differ-ent from traditional adoption drivers such as relativeadvantage.

We take this thought one step further by incorporat-ing image congruence as a relevant adoption criteria.

344 JOURNAL OF SERVICE RESEARCH / May 2005

1. “2002-2004 Research Priorities. A Guide to MSI Research Pro-grams and Procedures,” Marketing Science Institute.

Martineau (1958) was one of the first to propose that a con-sumer exhibits a preference for a store that has a person-ality consistent with the consumer’s image of himself orherself. Specifically, the attitudes or purchase intentionstoward a product or service seem to be highly dependenton the extent to which these products or services displaysimilarities to the consumer’s image of himself or herself,also referred to as the self-image (Graeff 1996). Onkvisitand Shaw (1987) suggested that this self-image is signifi-cant and relevant to the study of consumer behavior, asmany purchases made by consumers are directly influ-enced by the image individuals have of themselves. The fitbetween the self-image and the service image has beendefined as the image congruence hypothesis (Grubb andGrathwohl 1967). Despite its intuitive logic, numerousstudies report inconclusive findings concerning this rela-tionship (cf. Hogg, Cox, and Keeling 2000). In addition,research on image congruence within a service context isvirtually nonexistent. In the context of innovative servicesassociated with lifestyle products (i.e., mobile phones),the investigation of image congruence seems to be particu-larly relevant. Services by definition are intangible, whichmakes them difficult to evaluate. Signals relating to, forexample, added value or trustworthiness become crucial,especially in the case of innovative mobile services, whereinformation is limited in comparison to other services.Consequently, the current study aims to clarify the impactof image congruence on consumer adoption by focusingon the creation of consumer attitudes toward mobile ser-vices as well as the intention to use these services.

Both attitudes and intentions have been recognized ascentral constructs in the adoption process. Ajzen andFishbein (1980) indicated that attitude is one of the focalconstructs to be included when studying intentionalbehavior. According to Graeff (1996), a consumer’s self-concept can be used to influence attitudes and purchasedecisions. Creating an image of a product or service thatcompares favorably with the image of the consumer willpositively influence his or her attitude toward the productor service (e.g., Sirgy and Ericksen 1992). Other research-ers have suggested direct effects of image perceptionson consumer choices (e.g., Plouffe, Vandenbosch, andHulland 2001). Therefore, we consider both attitudes andintentions as relevant dependent variables in our research.

Finally, usage situations are taken into account. Previ-ous studies have indicated that usage situations have a con-siderable effect on consumer decision making (e.g., Belk,Balm, and Mayer 1982). As the role of image congruencemight be different for consumers depending on their usagesituation (e.g., Graeff 1997; Hogg, Cox, and Keeling2000), we believe that there is a distinction between pri-vate and public situations. This assertion is consistent withliterature pertaining to categorization (i.e., members fallwithin fuzzy boundaries) and stereotyping (i.e., the

dynamic process through which social groups make senseof and pursue their identity-related goals within inter-group contexts). Therefore, we investigate the role of theusage situation and its direct impact on consumer attitudesand intentions, as well as the joint effect of image congru-ence and usage situation. In the following section, weexplain the first study we conducted, investigating the roleof image congruence and the impact of private versus pub-lic usage situations.

HYPOTHESES DEVELOPMENT

A substantial amount of literature has been devoted tothe role of image congruence in consumer decision mak-ing (e.g., Hogg, Cox, and Keeling 2000). The main mes-sage that can be derived from this research is that imagecongruence is a relevant factor to consider. Another streamof relevant literature we draw on is self-brand connections.According to studies in this field, consumers tend to makedecisions that are consistent with their existing self-conceptions. Self-brand connections can be explained asthe extent to which individuals have incorporated a brandinto their self-concept (Escalas and Bettman 2003). Whenbrand associations are used to form the consumer’s self-image or to communicate this self-image to others, astrong connection is developed between the brand and theconsumer’s self-identity. Subsequently, this has a positiveeffect on consumers’ evaluations of the brand (Homer andBatra 1997). Consumers are more likely to adopt innova-tions that display a good fit with their own values, beliefs(Rogers 1995), and lifestyle (Kleijnen, de Ruyter, andWetzels 2004). This idea is similar to that of Grubb andGrathwohl (1967), who linked the individual’s self-concept with the symbolic value of consumer purchases.Therefore, consumers may rely on image congruence as asignal and decide whether to adopt an innovation contin-gent on the level of congruence. This is particularly rele-vant for mobile services. Due to the unique advantages ofmobile commerce (Kleijnen, de Ruyter, and Wetzels2004), mobile services are considered more personal thanany other form of remote service delivery. This also sug-gests a need for a fit with the self-image, as m-commerce isindissolubly connected to its users. This notion is con-firmed by Schlosser (2002), who stated that the use ofwireless technology is highly related to the consumer’sself-identity and the extent to which wireless innovationsfit to this self-identity. Therefore, a match between theimage of mobile services and the consumer’s self-imageseems crucial for attitude formation and the adoption deci-sion. Consequently, we suggest the following hypothesis:

Hypothesis 1: Consumers who perceive high image con-gruence will experience innovative, wireless ser-

Kleijnen et al. / IMAGE CONGRUENCE 345

vices more positively than consumers who experi-ence low image congruence in terms of (a) attitudeand (b) intention to adopt.

Past studies of image have identified image as a multi-dimensional construct (for an overview, see Sirgy 1982).Several researchers have recognized self-image as a two-fold concept: the actual self-image—how an individualsees himself or herself—and the ideal self-image—howhe or she would ideally wish to be seen. Onkvisit and Shaw(1987) argued that researchers should make an explicitspecification and justifiable reasoning for selecting a par-ticular self-component. The merit of differentiatingbetween types of self-concepts has been validated by sev-eral studies (e.g., Dolich 1969; Landon 1974). On the basisof these studies, it has been concluded that some prefer-ences are associated with the actual self-image, whereasothers display stronger correlation with the ideal self-image (Onkvisit and Shaw 1987). The underlying reasonfor this difference originates from the consumption situa-tion, where some services are typically consumed in a pri-vate situation and other products most often in a public sit-uation. In other words, “The self is a collection of maskseach tied to a particular set of social circumstances”(Sirgy, Grewal, and Mangleburg 2000). Previous researchhas investigated these effects by focusing on, for instance,private and public brands. However, in these studies, it waspossible to differentiate brands to represent either the pri-vate or the public brand. With mobile services, this type ofdistinction cannot be made. The elimination of geographicand time-related boundaries is regarded as the key advan-tage of mobile services (Balasubramanian, Peterson, andJarvenpaa 2002) by companies and consumers. Thisimplies that mobile services can and will be consumed indifferent types of situations, that is, private and public.Moreover, earlier research (Graeff 1997) has also indi-cated that consumers reveal strong similarities betweenactual and ideal self-image, and different effects of theactual versus ideal self-image congruence are hardlydistinguishable. Therefore, we hypothesize the following:

Hypothesis 2: In the adoption of innovative, wirelessservices, actual self-image congruence and idealself-image congruence will have an equally strong(positive) effect on (a) consumer attitude and (b)intention to adopt.

Besides image congruence, we also investigated theeffects of usage situations on consumer adoption of inno-vative services. Research has demonstrated how consum-ers exhibit different preferences in different consumptionsituations (e.g., Hogg, Cox, and Keeling 2000). Researchin both psychology and marketing has confirmed the needfor understanding situational contexts, which allows us to

considerably enhance the ability to explain, predict, andcontrol behavior (Schmitt and Shultz 1995). Social iden-tity theory (Tajfel and Turner 1986) suggests that consum-ers are inclined to let their choice be determined by thesocial groups surrounding them in order to gain socialacceptance of peer groups considered relevant (Chau1996; Graeff 1997). In addition, the self-monitoring litera-ture suggests that self-enhancement is dependent on theproduct or brand being a publicly recognized symbol(Belk, Balm, and Mayer 1982). Snyder and Gangestad(1986) suggested that some people are more sensitive toadapting to social contexts (self-monitoring). Whereaspeople described as low self-monitors are said to rely ondispositional information and direct their behavior inter-nally, high monitors rely on situational information todirect their behavior externally (Snyder 1979). The theoryof self-monitoring therefore suggests that consumersmake different decisions in different situations and thatthese decisions are particularly dependent on who is sur-rounding them.

In earlier research, the distinction between private andpublic situations has been made repeatedly (Dolich 1969;Graeff 1996, 1997). In a private situation, consumers arenot surrounded by others and, therefore, do not experienceany social pressure. Conversely, consumers will be influ-enced by their surroundings in a public setting to a largeextent. Previous research has recognized social pressure asa usage motivation (e.g., Karahanna and Limayem 2000).Generally, mobile phones and user extensions, such asmobile services, are described as innovative (Kleijnen, deRuyter, and Wetzels 2004), trendy (BBC 2000), and social(Jarvenpaa et al. 2003), implying a positive public opiniontoward mobile services. Consequently, consumers mightevaluate mobile services more positively in a public set-ting due to group influences. Finally, in Fishbein andAjzen’s (1975) theory of reasoned action and Bagozzi andWarshaw’s (1990) theory of trying, the importance of sub-jective norms (or social beliefs) in consumers’ decision-making processes is discussed. Numerous studies haveconfirmed the idea that consumers are heavily influencedby their surroundings when evaluating products or ser-vices (e.g., Karahanna and Limayem 2000; Venkatesh andDavis 2000). On the basis of the previous discussion, weposit the following:

Hypothesis 3: Consumers in a public situation will eval-uate innovative, wireless services more positivelythan consumers in a private situation in terms of (a)attitude and (b) intention to adopt.

Although studies on the image congruence hypothesishave been generally supportive, there is no extensive bodyof research that has examined variables interacting withimage congruence (Graeff 1996). According to Schmitt

346 JOURNAL OF SERVICE RESEARCH / May 2005

and Shultz (1995), the impact of situational variables onconsumer preferences seems particularly relevant forproducts or services that are typically recognized as imageenhancers. In fact, consumers might find it more importantto bestow a particular image on themselves in certain situa-tions than in others. However, the various situations inwhich this phenomenon might occur are only developed toa limited extent.

Consumers in a private situation (e.g., at home) have alesser need to portray a strong image. Consequently,image congruence will not have such an influential rolewhen compared to public situations. In contrast to the pri-vate situation, we argue that in a public situation, consum-ers often seek to profile themselves in a certain environ-ment to create an identity of social presence (Grewal,Mehta, and Kardes 2000). Consequently, in cases of highcongruence, consumers’ attitudes and intentions to useinnovative services in public situations will increase, dueto the opportunity to express their image to others. Con-sequently, the following hypothesis is developed:

Hypothesis 4: Consumers in a public situation experi-encing high image congruence will evaluate innova-tive, wireless services more positively than consum-ers in a private situation experiencing high imagecongruence in terms of (a) attitude and (b) intentionto adopt.

To test our hypotheses, an experiment was conducted.The following section discusses the details of the researchdesign and presents our findings.

EXPERIMENT 1

Research Context

According to Barwise and Strong (2002), the mobilechannel is an extremely personal medium that consumerscarry with them at all times, and optimistic forecasts claimthat mobile Internet will develop into a pervasive elementof everyday business and personal life. Rapid advances arebeing made, especially with regard to wireless financialservices (Barnard et al. 2001; Viquez 2001). Several bankshave launched initiatives to create wireless opportunitiesfor their customers. For example, a large European bank(www.postbank.nl) gave out 500.000 Wireless Applica-tion Protocol (WAP)-enabled mobile phones to their cus-tomers, enabling them to use mobile banking services.However, at this time, the market seems to be more supplythan demand driven. Although some consumers haveexpressed interest in interacting with service providers viaremote channels (Adriaenssens 2001), the wireless deliv-ery of banking services is not yet accepted by the majority

of consumers (Adrian 2001). The reason for this might bethat companies have stressed the functional aspects ofmobile services as the key advantage, thereby neglectingthe social value that can be derived from mobile services.Consequently, the image of mobile services could affect aconsumer’s decision as to whether to adopt such services.This has also been acknowledged by Becker, Mutschler,and Christi (2002), who emphasized the need for financialservice providers to create favorable images for newbanking channels in order to be successful.

Experimental Design

We conducted a between-subjects experiment to exam-ine whether there is a relationship between image congru-ence and attitudes toward the use of mobile services aswell as the intention to adopt these services, manipulatingonly the situation that participants were assigned to.Before the situation was presented, participants wereasked to fill out a questionnaire containing questionsrelated to the image of mobile banking, their actual self-image, and their ideal self-image. Following this, a partic-ular consumption situation was presented to them (privateor public). Participants were randomly assigned to a sce-nario (see Figure 1). The respondents were then asked toimagine themselves in that situation and answer questionsconcerning their attitudes and intentions to adopt. Toassess whether these scenarios reflected the intended stateof mind, manipulation checks were added (Perdue andSummers 1986).

The scenarios were subjected to a qualitative pretest inwhich 20 respondents were asked to critically evaluate thescenarios in terms of clarity and realism. The interviewsresulted in only minor adaptations of the scenarios, relatedto fine-tuning sentence formulations rather than radicalchanges in content.

Questionnaire Development

We employed a series of semantic differential scales tomeasure the various types of image, as this has been recog-nized as the most popular technique for this type of study(Graeff 1997). The dimensions were adapted partly fromGraeff (1997) and expanded with relevant dimensionsbased on a review of domain-specific (i.e., mobile ser-vices) literature. These dimensions were evaluated anddiscussed extensively by means of individual interviewswith 20 respondents, evaluating both the clarity and therelevance of the image dimensions. The resulting dimen-sions are reported in Table 1. These dimensions were usedto measure three types of image: image of wirelessfinance, the actual self-image of the respondent, and theideal self-image of the respondent. Attitudes toward using

Kleijnen et al. / IMAGE CONGRUENCE 347

mobile services as well as the intention to adopt mobileservices were included as the relevant dependent variablesin our study (see Table 1).

Sample

Respondents were approached on the streets of midsizecities in a small European country, during a fourteen-dayperiod that included weekdays, weekends, and differenthours of the day. In our sample selection, we employed thefollowing selection criteria: possession of mobile phone,acquaintance with mobile technologies such as WAP, andfamiliarity with brokerage (meaning that they are knowl-edgeable about stock trading and have personal experi-ence with trading stocks). These qualifications were nec-essary in order for the respondents to understand thescenarios presented to them. Questionnaire completionwas obtained by means of personal interviews. In total,298 people were approached, of which 29% were not will-ing to participate or did not fit the selection criteria. Afterscreening the questionnaires for incompleteness andabnormalities (Tabachnik and Fidell 1996), a usable sam-

ple of 209 respondents was obtained, which were ran-domly assigned to a scenario (104 private respondents vs.105 public respondents). The sample contained 49%women and 51% men and a broad range of age categories:younger than 25 years: 26%, 25 to 40 years: 26%, 41 to 55years: 26%, and older than 55 years: 22%.2

RESULTS

Manipulation Check

Manipulation checks were performed to assess whetherthe intended manipulations were understood by therespondents. Bipolar scales ranging from 1 to 7, where 4represents the neutral point (“this situation can be describedas a public situation” vs. “this situation can be described asa private situation”) were used to measure the respondents’evaluation of the situation. The results of the manipulation

348 JOURNAL OF SERVICE RESEARCH / May 2005

Base Scenario

You have put aside some money every month for a number of years and saved a considerable amount. Because the return on your savings account is very low, your financial adviser has suggested that you invest the money in stocks. You have followed his advice.

Situation Manipulation

Private: You are sitting at home in front of the television with nothing special on TV. In this situation, you receive an SMS from your financial adviser. He is advising you to buy certain stocks now, because of the real time developments that are occurring on the stock exchange market. You can execute the transaction he has proposed via your mobile phone by using the mobile transaction services. If you decide to execute the transaction, you will later receive an SMS with the details of the transaction.

Public: You are sitting in a lounge, surrounded by strangers. You do not know any of the people around you. In this situation, you receive an SMS from your financial adviser. He is advising you to buy certain stocks now, because of the real time developments that are occurring on the stock exchange market. You can execute the transaction he has proposed via your mobile phone by using the mobile transaction services. If you decide to execute the transaction, you will later receive an SMS with the details of the transaction. If you decided to use your mobile phone to execute the transaction, the people around you would notice that you use these mobile services.

FIGURE 1Scenario Description: Private Versus Public Situation

NOTE: SMS = Short Message Service.

2. Similar patterns were found within each subsample, leading to nosignificant differences between subsamples.

checks strongly suggested that there were differencesbetween the private and public situations, F(1, 207) =1,269.126, p < .001.

Measuring Congruence

In line with previous research (Graeff 1996, 1997), theoverall difference score reflecting degree of congruencebetween mobile services and self-image was calculatedwith a general Euclidean distance model:

D P Si ii

n

= −=∑ ( ) 2

1

,

where

D = overall difference score reflecting degree of congru-ence between service and self-image

i = a particular personality image dimensionPi = perception of the mobile service on image dimension iSi = perception of self on image dimension i

A small (large) difference indicates a high (low) degree ofsimilarity between the mobile service and self-image.

First, we investigated the effect of different types ofimage congruence (actual and ideal) on the dependentvariables in different situations. These relationships were

examined using correlation analysis. Table 2 presents thecorrelations between image congruence (self and ideal)and attitudes and intentions.

The significant correlations indicate that image con-gruence had a significant positive relationship with thevariables defined (remember that the congruence score issimilar to a reversed scale, therefore the negative correla-tion actually implies that when the congruence increases,the intention to adopt increases). Based on these results,actual as well as ideal self-image have a significant influ-ence on attitudes (Hypothesis 1a) and intentions to use(Hypothesis 1b).

Although there seem to be some (marginal) differencesbetween the image congruence correlations in the varioussituations, these differences are rather small, as wasexpected (Hypothesis 2). To test the differences betweencorrelations, several authors have suggested Hotelling’sT1 as an appropriate tool (e.g., Ferguson 1976). However,according to Steiger (1980), this test is not to be used, asthe underlying assumptions regarding distributions areoften not met. Therefore, Steiger suggests other methods,of which the T2 statistic developed by Williams (1959) isconsidered the most optimal solution. Comparing the cor-relations based on the T2 statistic reveals that there areactually no statistically significant differences betweenthe correlations for actual and ideal congruence (all p val-ues > .05). Therefore, we can conclude that both typesof image congruence have a significant effect on the

Kleijnen et al. / IMAGE CONGRUENCE 349

TABLE 1Questionnaire Items

Variable Items Alpha Study 1 Alpha Study 2 Source

Image Not very career focused—career focused Bank: .87 Bank: .75 Derived from Graeff (1997)Old-fashioned—modern Actual: .86 Actual: .72Unsuccessful—successful Ideal: .85 Ideal: .75Local—globalEasygoing—economicalInefficient—efficientSlow—FastConventional/solid—young urban

professional/trendyUnimportant—influentialTechnological follower—technological pioneer

Attitude Good—bad .96 .95 Dabholkar and Bagozzi (2002)Pleasant—unpleasantHarmful—beneficialFavorable—unfavorable

Intention Unlikely—likely .97 .97 Bruner II and Hensel (1998)/Improbable—probable Dabholkar and Bagozzi (2002)Impossible—possibleUncertain—certainDefinitely would not use—definitely would use

dependent variables, regardless of the situation presented(Hypothesis 2).

Considering these results and the high correlationbetween actual and ideal congruence (r = .92, p < .01), wewill examine the joint effects of (actual and ideal) imagecongruence and consumption situation on attitudes andintentions to adopt in the following analysis. This is in linewith the suggestions of Graeff (1997). We calculated theaverage of actual and ideal image congruence and usedthis variable of “overall” image congruence as a covariatein our further analysis.

Testing Assumptions

Considering the large sample size and the robustness ofMANOVA to departures from multivariate normality(Tabachnik and Fidell 1996), violations of multivariatenormality are not expected to be severe. Inspection of thehistograms, normal-probability plots, skewness, and kur-tosis for each dependent measure for each cell showedsome departures from normality. However, skewness andkurtosis are well within the expected values (Hair et al.1995). Moreover, in larger samples of 100 or more cases,the impact of departures from normality disappear (e.g.,Tabachnik and Fidell 1996). Consequently, MANCOVAhas generally been shown to be robust against violations ofthe multivariate normal distribution, and therefore we con-tinued the analysis maintaining the variables in their origi-nal form. Moreover, MANCOVA assumes linear relation-ships among all pairs of dependent variables in each cell,

as a deviation of this assumption would lead to a reductionin power of the statistical tests. Plotting the dependentvariables indicates a clear presence of linear relationships.The same procedure was conducted to inspect whether thesame holds true for correlations of the covariates with thedependent measure(s). As this is the case, the inclusion ofimage congruence as a covariate is appropriate. Finally,applying MANCOVA would be superfluous in case thedependent variables are uncorrelated, therefore the corre-lation between attitude and intention to adopt was evalu-ated as significant (r = .72, p < .01).

MANCOVA: Private Versus Public

To test our hypotheses, we conducted a MANCOVA.The results are reported in Table 3.

Research concerning the robustness of the statisticsavailable for MANCOVA has proposed the Pillai-Bartletttrace criterion (V) as the most robust statistic for gen-eral protection against departures from multivariate nor-mality and homogeneity of variance-covariance matrices(Tabachnik and Fidell 1996). The results of this test indi-cate significant results for the interaction effect, V = .237,F(2, 204) = 31.725, p ≤ .001, as well as significant effectsfor situation, V = .059, F(2, 204) = 6.366, p = .002, andimage congruence, V = .515, F(2, 204) = 108.229, p <.001.

In addition to our hypothesized model, we investigatedan alternative model. In this model, we used the image ofthe wireless services as an independent variable instead of

350 JOURNAL OF SERVICE RESEARCH / May 2005

TABLE 2Correlations: Private Versus Public

Private Consumption Situation Public Consumption Situation(n = 104) (n = 105)

Actual Congruence Ideal Congruence Actual Congruence Ideal Congruence

Attitude –.83** –.80** –.36** –.39**Intention –.75** –.74** –.36** –.40**

**Significant at the .01 level (two-tailed).

TABLE 3Results of MANCOVA Tests: Private Versus Public

Private Versus Public

Effect Pillai-Bartlett Trace (V) F(2, 204) p Value Partial η2

Main effectsSituation .059 6.366 .002 .059Image congruence .515 108.229 < .001 .515

Two-way interactionSituation × Congruence .237 31.725 < .001 .237

NOTE: Adjusted R2 attitude = .568, adjusted R2 intention = .432.

image congruence, as past literature has indicated that thisimage in itself might explain attitude formation as well asconsumer intentions (see our discussion: “Image, Deci-sion Making, and Adoption”). This alternative model,however, yielded a very low R2 adjusted for both attitude(adjusted R2 = .065) and intention (adjusted R2 = .021).This leads to the conclusion that our original model pro-vides a better explanation for the role of image with regardto attitudes and intentions. To further investigate ourhypotheses we used univariate analyses. The results of theomnibus ANCOVA tests are presented in Table 4. More-over, the cell means for the interaction between imagecongruence and situation are depicted in Figure 2.

It should be pointed out that in case of significant inter-action effects, the main effects in themselves do not fullyexplain the results of our experiment and should be inter-preted in light of the interaction effect (Keppel 1991).Therefore, we start by interpreting the interaction effectsfirst. With regard to attitude toward wireless finance (Table4), the test indicates significant results for the interactioneffect, F(1, 208) = 64.054, p < .001, η2 = .230. However,looking at the means presented in Figure 2, the results indi-cate that the type of consumption situation had a signifi-cant effect on attitudes only for the participants exhibitinglow overall congruence, whereas the attitudes for partici-pants with high overall image congruence did not differ

between situations. Therefore, although the interactioneffect is present, Hypothesis 4a is not supported.

The main effect of the situation, although significant(Hypothesis 3a), is relatively weak, F(1, 208) = 12.284,p = .001, η2 = .057, which is not surprising considering theinteraction effect we found. The results for image congru-

Kleijnen et al. / IMAGE CONGRUENCE 351

TABLE 4Results of Omnibus ANCOVA Tests:

Private Versus Public

Private Versus Public

PartialDependent Variables and Effect F(1, 208) p Value η2

Attitude (adjusted R2 = .568)Main effects

Situation 12.284 .001 .057Image congruence 186.849 < .001 .477

Two-way interactionSituation × Congruence 64.054 < .001 .230

Intention to adopt (adjusted R2 = .432)Main effects

Situation 5.044 .026 .024Image congruence 125.536 < .001 .380

Two-way interactionSituation × Congruence 24.727 < .001 .108

3.3

4.9 5.25.5

1

2

3

4

5

6

7

Private Public

Consumption Situation

Att

itud

e

4.1

2.8

5.25.4

1

2

3

4

5

6

7

Private Public

Consumption SituationIn

tent

ion

FIGURE 2Effects of Consumption Situation and Overall Congruence (Private vs. Public Situation)

NOTE: The solid line indicates high congruence, and the dotted line indicates low congruence. Averages are based on a median split of total image congru-ence, resulting in a high group and a low group.

ence, F(1, 208) = 186.849, p < .001, seem to suggest thatthe main effect of image congruence should not be over-looked (Hypothesis 1a) as the partial effect size is highestfor this effect (η2 = .477). In light of the significant interac-tion effect, however, the evidence of the ANCOVA alone isnot conclusive. Therefore, we conducted additional teststo inspect the main effects without the inclusion of theinteraction effect (Jaccard, Turrisi, and Wan 1990). One-way ANOVAs were conducted with attitude as a depen-dent variable. These analyses also provided support for themain effect of image congruence, F(37, 171) = 2.08, p =.005, suggesting in line with Hypothesis 1a that partici-pants with a high overall congruence exhibited a morefavorable intention to adopt than participants with a lowoverall congruence. In addition, we found a main effect ofthe situation, F(1, 207) = 12.284, p < .001, as attitudeswere significantly higher in the public situation comparedwith the private situation, supporting Hypothesis 3a.

The general pattern of the means for intention to adoptdisplays great similarity to that of the means for the atti-tudes. Again, the interaction effect, F(1, 208) = 24.727, p <.001, was significant (Table 4). Also, in this case, theresults reveal that the type of consumption situation onlyinfluenced the intention to adopt wireless finance for theparticipants exhibiting low overall congruence, whereasthe intention for participants exhibiting high overall con-gruence did not differ between situations. This result doesnot support the expected interaction effect, therefore wecannot accept Hypothesis 4b. Again, the ANCOVA sug-gests significant main effects for image congruence, F(1,208) = 125.536, p < .001, and situation, F(1, 208) = 5.044,p = .026. Participants with a high overall congruenceexhibited a more favorable intention to adopt than partici-pants with a low overall congruence. Moreover, intentionto adopt was significantly higher in the public situationcompared with the private situation. To further investigatethese effects, the one-way ANOVAs are inspected forimage congruence, F(37, 171) = 5.928, p < .016, and situa-tion, F(1, 207) = 3.077, p < .001). Based on these results,both Hypothesis 1b and Hypothesis 4b are supported. Toensure correct interpretation of our results, we also posedan open question to our respondents as to what motivatedtheir choices in the current scenario. After careful screen-ing, we were able to determine five broad answer cate-gories: novelty, fit with personality and/or style, socialsurroundings, control, and convenience. In the private sit-uation, “fit with personality and/or style” was mentionedmost frequently (73% mentioned this as one of the rea-sons), whereas in the public situation, social surroundingwas the most dominating motivator (85%). This suggeststhat our interpretations of the results are correct.

The results we found are only partially supported by theliterature discussed. Moreover, this study does not define

any particular group in the public setting. To further inves-tigate the effects of image congruence and usage situation,we conducted a second experiment where a more refinedinsight into the effects of the public setting is given.

SIGNIFICANT OTHERS:FRIENDS VERSUS COLLEAGUES

Whereas the first study merely investigates the effect ofthe presence of others, this study investigates the type ofothers that are of influence. Members of a social group aredefined on the basis of their similarities or differences witha particular group (Graeff 1997; Tajfel and Turner 1986).To validate our findings and to further refine the effects ofthe public surroundings, we investigated the effects of par-ticular groups of significant others (Shah 2003), namely,friends and colleagues. Previous adoption research hasfocused on either investigating friends (e.g., Grewal,Mehta, and Kardes 2000) or colleagues (e.g., Karahannaand Limayem 2000). It seems particularly relevant to in-clude both groups when studying mobile service adoption.Due to the ubiquity of mobile services (Balasubramanian,Peterson, and Jarvenpaa 2002), they can be consumed indifferent consumption situations, dissolving tradition-ally predefined borders between leisure time and work(Kleijnen, de Ruyter, and Wetzels 2004).

Friends are recognized as people who are known inti-mately, whereas coworkers are people who are known at amore superficial level. We therefore expect the socialinfluence of friends to be stronger than the influence ofcolleagues. Moreover, many of the mobile services offeredare based on interactions and social networks with friendsrather than with colleagues. This leads to the followinghypothesis:

Hypothesis 5: Consumers in a “friends” situation willevaluate innovative, wireless services more posi-tively than consumers in a “colleagues” situation interms of (a) attitude and (b) intention to adopt.

Also, in this study, we expect an interaction effectbetween image congruence and the usage situation. How-ever, the results of our previous experiment indicate a dif-ferent interaction than we hypothesized. It seems that con-sumers experiencing low congruence are influenced bytheir surroundings instead of consumers with high congru-ence. The reason for this might be that consumers feel aneed to express an image that is not their own (by means ofmobile service use) in order to gain social approval of theirenvironment. This is in line with Herr, Kardes, and Kim(1991), who stated that signals from the consumer’s sur-roundings overrule other negative information. Becausefriends are known intimately, there is no reason to portray

352 JOURNAL OF SERVICE RESEARCH / May 2005

a certain image. Colleagues, however, might pass judg-ment, and therefore clear profiling might be necessary inorder to gain a certain position in this particular groupsetting. Therefore, the following hypothesis is formulated:

Hypothesis 6: Consumers in a “colleagues” situationexperiencing low image congruence will evaluatewireless services more positively than consumers ina “friends” situation experiencing low image con-gruence in terms of (a) attitude and (b) intention toadopt.

Again, we test our hypotheses by means of an experiment.

EXPERIMENT 2

The design and procedure for Experiment 2 are iden-tical to those of Experiment 1. The objective of this ex-periment is to (a) replicate our findings concerning therelationship between image congruence and attitudetoward mobile services and the intention to adopt and (b)

investigate if a particular group of people (friends vs. col-leagues) surrounding the consumer have an influenceon these relationships. The situations are presented inFigure 3.

The colleagues’ scenario mainly differs on the descrip-tion of the people surrounding the respondent: “Imagineyou are in your office and colleagues are present.” Fur-thermore, these colleagues were described as “casualacquaintances.” Again, these scenarios were presented to20 respondents, who critically assessed clarity and realismof both scenarios. In addition, in this case, the scenarioswere adjusted in terms of clarity rather than actual content.

The questionnaire contained the same items as the pre-vious experiment. The coefficient alphas are againreported in Table 1. The sample, which was retrieved in asimilar way to the first experiment (number of people con-tacted was 305, actual participation rate 67%), consistedof 102 respondents for the friends scenario, and 102respondents for the colleagues scenario. The sample con-sisted of 49% women and 51% men and can be character-ized as follows in terms of age: younger than 25 years:

Kleijnen et al. / IMAGE CONGRUENCE 353

Base Scenario

You have put aside some money every month for a number of years and saved a considerable amount. Because the return on your savings account is very low, your financial adviser has suggested that you invest the money in stocks. You have followed his advice.

Situation Manipulation

Friends: You are in a lounge with your friends. You have known each other for years and do not have any big secrets from one another. In this situation, you receive an SMS from your financial adviser. He is advising you to buy certain stocks now, because of the real time developments that are occurring on the stock exchange market. You can execute the transaction he has proposed via your mobile phone by using the mobile transaction services. If you decide to execute the transaction, you will later receive an SMS with the details of the transaction. If you decided to use your mobile phone to execute the transaction, your friends would notice that you use these mobile services.

Public: You are in a lounge with your colleagues. You recently got a new job, so your colleagues do not know you very well. In this situation, you receive an SMS from your financial adviser. He is advising you to buy certain stocks now, because of the real time developments that are occurring on the stock exchange market. You can execute the transaction he has proposed via your mobile phone by using the mobile transaction services. If you decide to execute the transaction, you will later receive an SMS with the details of the transaction. If you decided to use your mobile phone to execute the transaction, your colleagues would notice that you use these mobile services.

FIGURE 3Scenario Description: Friends Versus Colleagues Situation

NOTE: SMS = Short Message Service.

25%, 25 to 40 years: 27%, 41 to 55 years: 25%, and olderthan 55 years: 23%.3

RESULTS

Manipulation Check

Bipolar scales (“situation where I am surrounded byfriends who know me well” vs. “situation where I am sur-rounded by colleagues who do not know me very well”)were used to measure the respondents’ evaluation of thesituation. The results of the manipulation checks stronglysuggested that there were differences between the friendsand colleagues situation, F(1, 202) = 489.394, p < .001.

Measuring Congruence

Again, the overall difference score reflecting degree ofcongruence between service and self-image was calcu-lated with a general Euclidean distance model. To validateour findings from the previous study, we again investi-gated the effect of different types of image congruence(actual and ideal) on the dependent variables in the two sit-uations. Table 5 presents the correlations between imagecongruence (actual and ideal) and the dependent variables.

Based on these results, actual as well as ideal self-image has a significant influence on attitudes and intentionto use (Hypothesis 1). We also find significant correlationsin this study indicating a significant positive relationshipbetween image and the variables defined (Hypothesis 2).Again, based on the T2 statistic, we could not find any sig-nificant differences between the correlations for actualand ideal congruence with regard to attitudes and inten-tions. The correlation between actual and ideal con-gruence for this sample was .90 (p < .01). Similar to theprevious study, we used the means of actual and ideal

image congruence to further investigate the joint effects ofcongruence.

MANCOVA Friends Versus Colleagues

After testing the assumptions using the same proce-dures as in Study 1 (correlation between attitude and inten-tion was r = .72, p < .01), a MANCOVA was conducted tofurther explore the effect of public surroundings, that is,we investigated the effect of different peer groups: friendsversus colleagues. The results of the MANCOVA are sum-marized in Table 6. For these scenarios we also found asignificant interaction effect, V = .143, F(2, 199) = 16.607,p < .001, as well as significant main effects for both situa-tion, V = .049, F(2, 199) = 5.151, p = .007, and image con-gruence, V = .667, F(2, 199) = 198.864, p < .001. Again,the alternative model with image of mobile bankinginstead of image congruence yielded poor results(adjusted R2 attitude = .054, adjusted R2 intention = .073).ANCOVAs were conducted to further explore our hypoth-eses. The results are presented in Table 7. The cell meansfor the interaction between image congruence andsituation are depicted in Figure 4.

In line with Study 1, the significant interaction effect,F(1, 200) = 33.214, p < .001, indicates that the type of con-sumption situation had a significant effect on attitude onlyfor the respondents exhibiting low overall congruence,whereas the attitudes for respondents exhibiting highoverall congruence did not differ between situations.Although we expected this interaction effect based on ourfirst experiment, attitudes are evaluated more positivelywhen the consumer is surrounded by friends instead ofcolleagues. Therefore, we cannot accept Hypothesis 6a. Inaddition, the results imply that high overall congruenceleads to a more favorable attitude than a low overall con-gruence, F(1, 200) = 373.004, p < .001. Furthermore,respondents had a significantly more favorable attitude inthe friends situation compared with the colleagues situa-tion, F(1, 200) = 10.245, p = .002. The main effects areagain separately investigated by means of one-wayANOVAs. These again illustrate significant results forimage congruence, F(27, 176) = 3.183, p = .001, providingadditional support for Hypothesis 1a. Moreover, signifi-cant results are found for the situation, F(1, 202) = 10.570,p < .001, confirming Hypotheses 5a.

The same results were found regarding intention toadopt wireless services. A significant interaction effectwas found, F(1, 200) = 23.296, p < .001, which again illus-trated that the type of consumption situation had an effecton respondents displaying a low overall congruence.However, in this case, the effect was found to be strongerin the friends situation, implying that we cannot acceptHypothesis 6b. Moreover, the main effects seem to suggest

354 JOURNAL OF SERVICE RESEARCH / May 2005

TABLE 5Correlations: Friends Versus Colleagues

Friends Consumption Situation Colleagues Consumption Situation(n = 102) (n = 102)

Actual Ideal Actual IdealCongruence Congruence Congruence Congruence

–.77** –.77** –.78** –.79**–.76** –.74** –.74** –.75**

**Significant at the .01 level (two-tailed).

3. Similar patterns were found within each subsample, leading to nosignificant differences between subsamples.

Kleijnen et al. / IMAGE CONGRUENCE 355

TABLE 6Results of MANCOVA Tests: Friends Versus Colleagues

Friends Versus Colleagues

Effect Pillai-Bartlett Trace (V) F(2, 199) p Value Partial η2

Main effectsSituation .049 5.151 .007 .049Image congruence .667 198.864 < .001 .667

Two-way interactionSituation × Congruence .143 16.607 < .001 .143

NOTE: Adjusted R2 attitude = .420, adjusted R2 intention = .482.

TABLE 7Results of Omnibus ANCOVA Tests: Friends Versus Colleagues

Friends Versus Colleagues

Dependent Variables and Effect F(1, 200) p Value Partial η2

AttitudeMain effects

Situation 10.245 .002 .049Image congruence 373.004 < .001 .651

Two-way interactionSituation × Congruence 33.214 < .001 .162

Intention to adoptMain effects

Situation 5.243 .023 .026Image congruence 290.981 < .001 .593

Two-way interactionSituation × Congruence 23.296 < .001 .104

2.4

4.0

5.65.6

1

2

3

4

5

6

7

Friends Colleagues

Consumption Situation

Inte

ntio

n

4.3

3.0

5.55.4

1

2

3

4

5

6

7

Friends Colleagues

Consumption Situation

Att

itud

e

FIGURE 4Effects of Consumption Situation and Overall Congruence (Friends vs. Colleagues Situation)

NOTE: The solid line indicates high congruence, and the dotted line indicates low congruence. Averages are based on a median split of total image congru-ence, resulting in a high group and a low group.

that respondents with an overall high congruence display asignificantly higher intention to use wireless servicesthan those with low overall congruence, F(1, 200) =290.981, p < .001, whereas consumers in the friends situa-tion demonstrate a higher intention to use wireless ser-vices than respondents in the colleagues situation, F(1,200) = 5.243, p = .023. Further inspection of the maineffects leads to further support for our hypotheses. Forboth image congruence, F(27, 176) = 3.978, p < .001, andsituation, F(1, 202) = 12.730, p < .001, significant resultswere found, validating our initial results regardingHypothesis 1b and providing support for Hypothesis 5b.Again, we used an open-ended question as a check for ourresults. In the friends situation, the social surrounding wasmentioned most often by the respondents as the reason fortheir behavior (94%). In the colleagues scenario, however,fit with personality and/or style was cited most frequently(75%). The implications of these findings are nowdiscussed.

DISCUSSION

The general purpose of this study was to identify theeffect of image congruence on attitudes and intentions toadopt wireless services. Moreover, we aimed to investi-gate how these effects were moderated by the consump-tion situation. Consistent with previous findings (e.g.,Graeff 1997; Hogg, Cox, and Keeling 2000), image con-gruence seems to have a substantial effect on consumers’evaluations. However, concerning the effects of differenttypes of self-image, a number of interesting results werefound. On the basis of our findings, we can conclude thatboth actual and ideal self-image have an influence on theadoption of mobile services, regardless of the situation.Both effects are present and have a significant impact onattitudes and intentions to adopt. Our results also enable usto nuance this general finding with regard to the differenttypes of consumption situations that have an effect on theconsumer’s attitude and intention to adopt. Specifically,consumers seem to be influenced by social situations. Thisis consistent with our expectations based on research con-cerning social identity (Grewal, Mehta, and Kardes 2000;Tajfel and Turner 1986). When investigating the type ofsocial situation, results indicate that consumers are influ-enced more strongly by their friends than by their col-leagues. An explanation for this might be that the respon-dents experience a stronger social tie, or even socialpressure, when surrounded by their friends as opposed totheir colleagues. In addition, the working environmentmight be recognized as a less personal environment wheresocial influences are experienced less strongly by consum-ers. However, these types of strategies will only be effec-tive for consumers whose self-image is inconsistent with

the service image, whereas the results demonstrate thatconsumers exhibiting a high overall congruence seem tobe fairly stable in their evaluations. Although surprising,this effect may be interpreted as follows. When the imagecongruence is low, consumers may look for other signals.Herr, Kardes, and Kim (1991) in their research have foundthat influences from the consumer’s environment lead tofavorable attitudes, even when extensive, diagnosticattribute information is available that would lead to lesspositive evaluations. Similarly, in the current setting, thesituational influence overrules the effect of low congru-ence, leading to a more positive attitude and intention toadopt when consumers are surrounded by others, and spe-cifically by friends.

Overall, we can conclude that image congruence hasa strong impact on consumer evaluations. Although pre-vious research has already demonstrated the relevanceof image in the adoption process as such (Plouffe,Vandenbosch, and Hulland 2001), these results indicatethe importance of the fit between a consumer’s image andthe image of innovative, wireless services.

SUGGESTIONS FORFURTHER RESEARCH

During the process of our research, several limitationswere encountered. These limitations suggest directions forfuture research and point to a number of theoretical impli-cations. First of all, the experimental design pertained to alaboratory experiment. As a consequence, the general-izability of the findings would be limited with regard toreal-life settings. Furthermore, an experimental designmight suffer from limitations pertaining to a possible lackof realism. For instance, our study involved a single sam-ple of each subject group judging one hypothetical sce-nario on the basis of limited information. Even though theresults of the manipulation check demonstrate successfulmanipulation of various consumer situations, there mightbe a difference between simulation and real experience,affecting the way in which respondents react to the sit-uation. Several researchers (e.g., de Ruyter and Wetzels1998) have suggested the use of audiovisual scenarios thatare more realistic than verbal stimuli. However, simulatingconsumption situations might be complicated in the cur-rent research setting.

Although our study adds to the understanding of therole image congruence plays in determining consumerintentions, it is merely a first step in fully understandingthe impact of the various dimensions of the self-concept.Sirgy (1982) provided an overview of these dimensions.Although the two dimensions actual and ideal self aremost predominantly present in the literature, furtherinsights could be gained by incorporating additional

356 JOURNAL OF SERVICE RESEARCH / May 2005

dimensions of the self-image, such as the relational self orsocial self.

In addition, other situational influences could be incor-porated. Although the focus in the current research hasbeen on significant others surrounding the consumer, fur-ther research may investigate consumer predispositionsthat are known to influence consumer behavior (e.g.,mood). Similarly, personality traits have not been includedin the current research. A useful segmentation variable inthis area might be Parasuraman’s (2000) TechnologyReadiness Index. For example, consumers who are less“technology ready” might be more willing to adopt inno-vations if they perceive a fit with their own image, that is,the signal is more important for them than for consumerswho are more technology ready. Furthermore, thisresearch does not take into account changes over time.Karahanna, Straub, and Chervany (1999) have made asubstantial contribution in this field by investigating pre-and postadoption beliefs. However, as Karahanna, Straub,and Chervany stated, our knowledge about how these be-liefs and attitudes change over time is still limited, indi-cating a need for further research in this area.

Finally, future research should focus on a broader rangeof services. However, it can be argued that other categoriesare even more embedded in social settings (e.g., mobileentertainment and communication services), which leadsus to believe that investigation of these types of servicesshould confirm our findings. Moreover, past research hasmainly focused on functional aspects when studying adop-tion of new technology or services (e.g., Davis 1989),whereas socioemotional aspects have been included im-plicitly (e.g., Rogers 1995). The relevance of image con-gruence in the current study signals the need for researchregarding adoption theory that goes beyond functional rel-evance. In particular, more attention should be devoted tothe role of social networks in the adoption and diffusion ofinnovative networks.

MANAGERIAL IMPLICATIONS

There are several managerial implications to be drawnfrom this study. First, the current research suggests thatemploying various consumption situations as a promo-tional tool might be very effective with respect to consum-ers whose self-image is not consistent with the wirelessservice image. Instead of focusing on the overwhelming(functional) possibilities mobile services offer, managersshould exploit the use of socioemotional aspects whenmarketing wireless services. While most companies havefocused on the individual, marketers should realize theimportance of the social surroundings in the adoptionprocess.

Moreover, techniques related to viral marketing mightbe useful in promoting wireless services. Using consum-ers to introduce other consumers to new services might bethe most effective way to create consumer acceptance.Companies need to create opportunities for “wirelessworth of mouth,” for example, by allowing users of a par-ticular service to invite other potential customers for a freetrial. Focusing on social networks to spread these innova-tions might be particularly helpful, as other interactive ser-vices such as Short Message Service (SMS) have beenspread effectively through these networks.

Finally, given the impact of image congruence on theadoption of new services, marketers should consider seg-menting their markets in terms of consumer image. Ana-lyzing the firm’s customer base may disclose differentclusters of images. Developing and positioning new prod-ucts or services close to these images may speed up theadoption process. For example, through Nokia 5100,Nokia is targeting consumers with a sporty self-image.The phone is designed to resemble a “six-pack” (i.e., well-defined abdominal muscles), and the promotional cam-paigns portray athletic people. The promotional slogan“shaped by performance” emphasizes the match betweenthe athletic consumer and the phone design to an evengreater extent. In general, Nokia claims that their phonesare “designed to make an impression” (www.nokia.com).This strategy can certainly be extended to numerous inno-vative (wireless) services.

REFERENCES

Aaker, David A. (1991), Managing Brand Equity: Capitalizing on theValue of a Brand Name. New York: Free Press.

Adriaenssens, Charles (2001), “The Potential for M-Commerce in RetailFinancial Services,” in Mobile Commerce—Connecting with theFuture, Mike Elysée, Charlie Henderson, Nicholas A. Hopwood,John Machin, and Malcolm Marshall, eds. London: TechnologyPublishing.

Adrian, Bradford (2001), “Wireless Financial Services Delivery: ACloud of Hype,” Gartner Research Report (COM-13-5433), 1-4.

Ajzen, I. and M. Fishbein (1980), Understanding Attitudes and Predict-ing Social Behavior. Englewood Cliffs, NJ: Prentice Hall.

Bagozzi, Richard P. and Paul R. Warshaw (1990), “Trying to Consume,”Journal of Consumer Research, 17 (2), 127-40.

Balasubramanian, Sridhar, Robert A. Peterson, and Sirkka L. Jarvenpaa(2002), “Exploring the Implications of M-Commerce for Marketsand Marketing,” Journal of the Academy of Marketing Science, 30(4), 348-61.

Barnard, Chris, Elena Carreri, Pim Bilderbeek, Barbara Blesio, DanieleBonfanti, and Tim Sheedy (2001), “Banking on Mobile: Prospects forWestern Europe,” IDC Research Report, 1-58.

Barwise, Patrick and Colin Strong (2002), “Permission-Based MobileAdvertising,” Journal of Interactive Marketing, 16 (1), 14-24.

BBC (2000), “A Simple Guide to M-Commerce,” BBC News BusinessSection, retrieved February 18, 2004, from www.thefeature.com.http://news.bbc.co.uk/1/hi/business/992292.stm

Becker, Ursula, Robert Mutschler, and Bettina Christi (2002), “Effi-cient Multichannel Banking,” Forrester Research TechStrategyReport, 1-28.

Kleijnen et al. / IMAGE CONGRUENCE 357

Belk, R. W., K. D. Balm, and R. N. Mayer (1982), “Developmental Rec-ognition of Consumption Symbolism,” Journal of ConsumerResearch, 9 (June), 4-17.

Bruner, Gordon C., II and Paul J. Hensel (1998), Marketing Scales Book:A Compilation of Multi-Item Measures, 2nd ed. Chicago: AmericanMarketing Association.

Chau, Patrick Y. K. (1996), “An Empirical Assessment of a ModifiedTechnology Acceptance Model,” Journal of Management Informa-tion Systems, 13 (2), 185-204.

Dabholkar, Pratibha A. and Richard P. Bagozzi (2002), “An AttitudinalModel of Technology-Based Self-Service: Moderating Effects ofConsumer Traits and Situational Factors,” Journal of the Academy ofMarketing Science, 30 (3), 184-201.

Davis, F. D. (1989), “Perceived Usefulness, Perceived Ease of Use, andUsage of Information Technology: A Replication,” MIS Quarterly, 16(2), 319-39.

, R. P. Bagozzi, and P. R. Warshaw (1989), “User Acceptance ofComputer Technology: A Comparison of Two Theoretical Models,”Management Science, 35, 982-1003.

de Ruyter, Ko and Martin G. M. Wetzels (1998), “On the Complex Natureof Patient Evaluations of General Practice Service,” Journal of Eco-nomic Psychology, 19, 565-90.

Dolich, Ira J. (1969), “Congruence Relationships between Self Imagesand Product Brands,” Journal of Marketing Research, 6 (February),80-84.

Escalas, Jennifer Edson and James R. Bettman (2003), “You Are WhatThey Eat: The Influence of Reference Groups on Consumers’Connections to Brands,” Journal of Consumer Psychology, 13 (3),339-48.

Ferguson, George A. (1976), Statistical Analysis in Psychology & Educa-tion, 4th ed. New York: McGraw-Hill.

Fishbein, M. and I. Ajzen (1975), Belief, Attitude, Intention and Behavior.Reading, MA: Addison-Wesley.

Fornell, Claes (1992), “A National Customer Satisfaction Barometer:The Swedish Experience,” Journal of Marketing, 56 (1), 6-21.

Graeff, Timothy R. (1996), “Using Promotional Messages to Manage theEffects of Brand and Self-Image on Brand Evaluations,” Journal ofConsumer Marketing, 13 (3), 4-18.

(1997), “Consumption Situations and the Effect of Brand Imageon Consumers’Brand Evaluations,” Psychology & Marketing, 14 (1),49-70.

Grewal, R., R. Mehta, and F. R. Kardes (2000), “The Role of the Social-Identity Function of Attitudes in Consumer Innovativeness and Opin-ion Leadership,” Journal of Economic Psychology, 21, 233-52.

Grubb, Edward L. and Harrison L. Grathwohl (1967), “Consumer Self-Concept, Symbolism and Market Behavior: A Theoretical Ap-proach,” Journal of Marketing, 31 (October), 22-27.

Hair, Joseph F., Jr., Rolph E. Anderson, Ronald L. Tatham, and WilliamC. Black (1995), Multivariate Data Analysis with Readings, 4th ed.Englewood Cliffs, NJ: Prentice Hall.

Herr, Paul M., Frank R. Kardes, and John Kim (1991), “Effect of Word-of-Mouth and Product-Attribute Information on Persuasion: AnAccessibility-Diagnosticity Perspective,” Journal of ConsumerResearch, 17, 454-62.

Hogg, Margaret K., Alastair J. Cox, and Kathy Keeling (2000), “TheImpact of Self-Monitoring on Image Congruence and Product/BrandEvaluation,” European Journal of Marketing, 34 (5/6), 641-66.

Homer, Pamela M. and Rajeev Batra (1997), “The Impact of Susceptibil-ity to Normative Influence on the Relationship between ConsumerPersonality and Brand Personality,” in Proceedings of the WinterConference of the Society of Consumer Psychology, C. Pechman andS. Ratneshwar, eds. Washington, DC: American PsychologicalAssociation, 132-40.

Jaccard, J., R. Turrisi, and C. K. Wan (1990), Interaction Effects in Multi-ple Regression. Beverly Hills, CA: Sage.

Jarvenpaa, Sirkka L., Karl Reiner Lang, Yoko Takeda, and Virpi KristiinaTuunainen (2003), “Mobile Commerce at Cross Roads,” Communi-cations of the ACM, 46 (12), 41-45.

Karahanna, E. and M. Limayem (2000), “E-Mail and V-Mail Usage:Generalizing across Technologies,” Journal of Organizational Com-puting and Electronic Commerce, 10 (1), 49-66.

Karahanna, Elena, Detmar W. Straub, and Norman L. Chervany (1999),“Information Technology Adoption across Time: A Cross-SectionalComparison of Pre-Adoption and Post-Adoption Beliefs,” MIS Quar-terly, 23 (2), 183-213.

Keppel, G. (1991), Design and Analysis: A Researcher’s Handbook.Englewood Cliffs, NJ: Prentice Hall.

Kleijnen, Mirella H. P., J. C. de Ruyter, and Martin G. M. Wetzels (2004),“Consumer Adoption of Wireless Services: Discovering the Rules,While Playing the Game,” Journal of Interactive Marketing, 18 (2),51-61.

Landon, E. Laird, Jr. (1974), “Self-Concept, Ideal Self-Concept and Con-sumer Purchase Intentions,” Journal of Consumer Research, 1 (Sep-tember), 44-51.

Martineau, Pierre (1958), “The Personality of the Retail Store,” HarvardBusiness Review, 36 (January-February), 47-55.

Meuter, M. L., A. L. Ostrom, R. I. Roundtree, and M. J. Bitner (2000),“Self-Service Technologies: Understanding Customer Satisfactionwith Technology-Based Service Encounters,” Journal of Marketing,64 (July), 50-64.

Moore, Geoffrey A. (2002), Crossing the Chasm: Marketing and Sell-ing High-Tech Products to Mainstream Consumers. New York:HarperCollins.

Olson, Jerry C. (1972), “Cue Utilization of the Quality Perception Pro-cess: A Cognitive Model and an Empirical Test,” doctorial disserta-tion, Purdue University, West Lafayette, IN.

Onkvisit, Sak and John Shaw (1987), “Self-Concept and Image Congru-ence: Some Research and Managerial Implications,” Journal of Con-sumer Marketing, 4 (1), 13-23.

Parasuraman, A. (2000), “Technology Readiness Index (Tri): A Multiple-Item Scale to Measure Readiness to Embrace New Technologies,”Journal of Service Research, 2 (4), 307-20.

Perdue, B. C. and J. O. Summers (1986), “Checking the Success ofManipulations in Marketing Experiments,” Journal of MarketingResearch, 23 (November), 317-26.

Plouffe, C. R., M. Vandenbosch, and J. Hulland (2001), “IntermediatingTechnologies and Multi-Group Adoption: A Comparison of Con-sumer and Merchant Adoption Intentions toward a New ElectronicPayment System,” Journal of Product Innovation Management, 18(2), 65-81.

Rogers, E. M. (1995), Diffusion of Innovations, 4th ed. New York: FreePress.

Schlosser, Francine K. (2002), “So, How Do People Really Use TheirHandheld Devices? An Interactive Study of Wireless TechnologyUse,” Journal of Organizational Behavior, 23, 401-23.

Schmitt, Bernd H. and Clifford J. Shultz II (1995), “Situational Effects onBrand Preferences for Image Products,” Psychology & Marketing, 12(5), 433-46.

Shah, James (2003), “The Motivational Looking Glass: How SignificantOthers Implicitly Affect Goal Appraisals,” Journal of Personality andSocial Psychology, 85 (3), 424-39.

Sirgy, M. Joseph (1982), “Self-Concept in Consumer Behavior: A Criti-cal Review,” Journal of Consumer Research, 9, 287-300.

(1985), “Using Self-Congruity and Ideal Congruity to PredictPurchase Motivation,” Journal of Business Research, 13, 195-206.

and M. K. Ericksen (1992), “Employed Females Clothing Prefer-ence, Self-Image Congruence, and Career Anchorage,” Journal ofApplied Social Psychology, 22, 408-22.

, D. Grewal, and T. Mangleburg (2000), “Retail Environment,Self-Congruity, and Retail Patronage: An Integrative Model and aResearch Agenda,” Journal of Business Research, 49, 127-38.

Snyder, M. (1979), “Self-Monitoring Processes,” Advances in Experi-mental Social Psychology, 12, 86-128.

and S. Gangestad (1986), “On the Nature of Self-Monitoring:Matters of Assessment, Matters of Validity,” Journal of Personalityand Social Psychology, 51, 125-39.

358 JOURNAL OF SERVICE RESEARCH / May 2005

Steiger, James H. (1980), “Tests for Comparing Elements of a CorrelationMatrix,” Psychological Bulletin, 97 (2), 245-51.

Tabachnik, B. G. and L. S. Fidell (1996), Using Multivariate Statistics.New York: HarperCollins.

Tajfel, Henri and John C. Turner (1986), “The Social Identity Theoryof Intergroup Behavior,” in Psychology of Intergroup Relations,S. Worchel and W. G. Austin, eds. Chicago: Nelson-Hall.

Tornatzky, L. G. and K. J. Klein (1982), “Innovation Characteristicsand Innovation Adoption-Implementation: A Meta-Analysis ofFindings,” IEEE Transactions on Engineering Management, 29 (1),28-45.

Venkatesh, Viswanath and Fred D. Davis (2000), “A Theoretical Exten-sion of the Technology Acceptance Model: Four Longitudinal FieldStudies,” Management Science, 46 (2), 186-204.

Viquez, Frank (2001), “Mobile Commerce: Analyzing Emerging Oppor-tunities in Technologies, Applications, Markets and Trends,” AlliedBusiness Intelligence Inc. Strategic Research Report, 1-225.

Williams, E. J. (1959), “The Comparison of Regression Variables,” Jour-nal of the Royal Statistical Society, Series B, 21, 396-99.

Mirella Kleijnen currently is an assistant professor at the VrijeUniversiteit Amsterdam, the Netherlands. She is in the last stagesof her Ph.D., which she has written during her time as a Ph.D. stu-dent with the Faculty of Economics and Business Administrationat Maastricht University, the Netherlands. Her research focuseson the factors influencing consumer adoption of innovative ser-vices, and mobile services in particular. Her work has been pub-lished in the Journal of Interactive Marketing, the Journal ofFinancial Services Marketing, Mobile Commerce: Technology,Theory, and Applications, the International Journal of ServiceIndustry Management, and Total Quality Management.

Ko de Ruyter is a professor of international service research atthe University of Maastricht, the Netherlands, and director of theMaastricht Academic Center for research in Services (MAXX).He has published six books and numerous scholarly articles in,among others, the Journal of Marketing, the International Jour-nal of Research in Marketing, the Journal of Retailing, DecisionSciences, Marketing Letters, the Journal of Management Stud-ies, the Journal of Business Research, the Journal of EconomicPsychology, the Journal of Service Research, the InternationalJournal of Service Industry Management, the Journal of Busi-ness Research, Information and Management, the EuropeanJournal of Marketing and Accounting, and Organisation andSociety. He serves on the editorial boards of various internationalacademic journals, among which the Journal of Service Re-search and the International Journal of Service Industry Man-agement. His research interests concern international servicemanagement, e-commerce, and customer satisfaction anddissatisfaction.

Tor W. Andreassen is an associate professor of marketing and adoctor of economics from Stockholm University, School of Busi-ness (1997). His works have been published in journals such asthe Journal of Economic Psychology, the Journal of Public Sec-tor Management, the Journal of Service Research, the Journal ofServices Marketing, the European Journal of Marketing, and theJournal of Service Industry Management. His research interestsare frontline employees’ contribution to customer value, cus-tomer (dis)satisfaction, branding service organizations, and cus-tomer relationship management.

Kleijnen et al. / IMAGE CONGRUENCE 359