antecedents of virtual community satisfaction and loyalty: an empirical test of competing theories

CYBERPSYCHOLOGY & BEHAVIOR

Volume 11, Number 2, 2008© Mary Ann Liebert, Inc.DOI: 10.1089/cpb.2007.0003

Antecedents of Virtual Community Satisfaction andLoyalty: An Empirical Test of Competing Theories

HSIU-FEN LIN, Ph.D.

ABSTRACT

The rapid growth of virtual communities (VCs) on the Internet raises important research ques-tions regarding the antecedents of satisfaction and loyalty in VCs. This study compared thetechnology acceptance model (TAM), Seddon Information Systems Success Model (Seddonmodel), and an integrated model (combining the TAM and Seddon model) to examine whichmodel best helps to predict user satisfaction with VCs. Using a structural equation model,data collected from 198 community members were used to compare the three models in termsof overall model fit, explanatory power, and path significance. The findings show that twoTAM components (perceived usefulness and ease of use) are determinants of user satisfac-tion with VCs. This study also found the influence of quality-perception dimensions (e.g.,system and information quality) on user satisfaction to be significant. Additionally, this studyfound general support for user satisfaction as a determinant of loyalty in VCs. Finally, an in-tegrated model, combining the TAM and Seddon model, provided better explanatory powerthan either the TAM or the Seddon model alone. Finally, this study discusses the implica-tions of these findings and offers directions for future research.

138

INTRODUCTION

VIRTUAL COMMUNITIES (VCS) are formed on the In-ternet and are expected to serve the needs of

members for communication, information, and en-tertainment. VCs are having a major impact on en-hancing Internet user online experiences. For ex-ample, activities conducted in VCs include chatting,making friends, exchanging ideas, and sharingknowledge on particular subjects. All these com-puter-mediated communications (CMCs) have ledindividuals to change their communication and col-laboration methods. Moreover, a successful VC de-pends on people visiting their sites, facilitating so-cial interaction among community members, andmost significantly, believing that the VC offers bet-

ter choices than the alternative Web sites.1 Hence,in order to develop an effective VC to facilitate so-cial interaction and design for sociability, it is im-portant to understand what drives users’ satisfac-tion and their decisions regarding participation inVCs.

This study aims to conceptualize, develop, andvalidate independent variables that result in usersatisfaction and loyalty for VCs. It examines howthe technology acceptance model (TAM) and Sed-don Information Systems Success Model (Seddonmodel) theoretically differ in explaining VC satis-faction and loyalty and compares whether a re-search model integrating the two models can pre-dict user satisfaction in the context of VCs betterthan using either of the two models alone. Using a

Department of Shipping and Transportation Management, National Taiwan Ocean University, Taiwan, Republic of China.

structural equation model, data collected from 198community members were used to compare thethree models (TAM, Seddon model, and the inte-grated model) in terms of overall model fit, ex-planatory power, and path significance.

The theoretical framework, presented in Figure 1,attempts to explain the antecedents that influenceVC satisfaction as well as the consequences of usersatisfaction. This study utilizes constructs from theTAM and Seddon model as a lens to examine usersatisfaction and loyalty in VCs. The theoreticalframework of this study proposes that the two be-liefs of TAM and Web site quality influence satis-faction and thus determine user loyalty to VCs.

THEORETICAL BACKGROUND

Partly as a result of the contemporary nature of VCs, few empirical studies have examined andcompared the validity and explanatory utility of ex-isting theoretical models in the context of VCs. Gen-erally, perceptions of VC success have been inves-tigated within two primary streams of research,including technology acceptance literature2,3 andWeb site quality literature.4,5 Both research streamsmake valuable contributions to improving under-standing of the manner in which system features ul-timately influence user satisfaction with VCs. Useracceptance of VCs is an important first step towardachieving success in VC; however, a high-qualitysystem is more critical in ensuring long-term VCsuccess.6 Literature review clearly identifies twodrivers of VC success: comfort and acceptance of theenabling Internet-based technology and the Website quality provided by VC providers, which inturn leads to VC member loyalty. These constructsare also proposed by TAM and the Seddon model.TAM is based on Ajzen and Fishbein’s7 theory ofreasoned action (TRA), which indicates that socialbehaviors are motivated by individual attitudes,and it is specifically designed to predict system us-age.8,9 The Seddon model assumes that system qual-

ity and information quality affect user satisfactionand system usage.10 Moreover, an inclusive and in-tegrated analysis can provide a multidisciplinaryperspective on user satisfaction and loyalty in VCs.This study reviews the theoretical foundations ofthe models and presents a case for integrating themin assessing VC satisfaction and loyalty.

Satisfaction and loyalty

Generally, satisfaction measures subjective userassessments of any outcome or experience regard-ing a specific information system (IS).11 In the con-text of VCs, Lin and Lee4 empirically validated threemeasures of user satisfaction: satisfying users’ so-cial needs, adequacy of the information content foruser needs, and user satisfaction with the VC. Sat-isfaction is a common measure of IS success becauseit affects users’ motivation to continue to use theIS.12 VCs are currently perceived as a new form ofmainstream personal communication, which aresufficiently potent to make Internet users changetheir information exchange methods and firms ad-just their marketing strategies. Hence, research onVC acceptance can enhance understanding of userbeliefs or motives to use VCs and show the effect ofthese factors on user satisfaction and loyalty in VCs.Furthermore, based on the human–computer inter-action perspective, researchers noted that Web sitequality is a critical factor for predicting members’intention to use the VC.13 A high-quality Web sitecan generate a comfortable virtual environmentwhere users can easily identify functional groupsand navigation aids and ensure efficient informa-tion exchange. Therefore, this study examines VCsatisfaction on the basis of an integrated analysis us-ing multiple constructs of technology acceptanceand Web site quality.

Satisfying Internet users is strategically crucial forall online businesses. Among its most strategic con-sequences, satisfaction leads to improved customerretention, positive word of mouth, and increasedcustomer loyalty. Oliver14 also suggested that satis-faction feeds back into the system and influences

VIRTUAL COMMUNITIES: TEST OF COMPETING THEORIES 139

VC technology

acceptance

VC Web site

quality

VC satisfaction VC loyalty

FIG. 1. Theoretical framework.

subsequent interaction and behavior. In online en-vironments, as overall satisfaction with Web sitesincreases, the likelihood that users will use the Website again is increased.15

Technology acceptance model

TAM has been shown to be a powerful and parsi-monious framework for explaining user acceptance oftechnology.8,9 TAM identified that perceived useful-ness (PU) and perceived ease of use (PEOU) deter-mine attitude toward an IS. The attitude, in turn, leadsto user intention to use an IS and the eventual accep-tance of technology. Regarding intention in TAM, itis argued that if no other variables intervene, as is thecase in many IS applications, then intention can be ex-cluded.16 Nevertheless, satisfaction is an attitude con-struct that is captured in several items relating to userloyalty. Recent findings indicate that user satisfactionin online environments is significantly higher than intraditional IS because of the ease of information ac-quisition.16 Thus, based on recent research regardingonline environments, this study applies TAM con-structs to examine user satisfaction with VCs.

Seddon IS Success Model

Seddon10 respecified and slightly extended a ver-sion of DeLone and McLean’s IS success model11 toprovide a clearer, more theoretically sound concep-tualization of the relationships between the variousIS success constructs. The five dimensions for IS suc-cess are system quality, information quality, per-ceived usefulness, user satisfaction, and IS use. Inthe Seddon model, system quality and informationquality affect PU and user satisfaction. PU impactsuser satisfaction. The model contains a direct pathleading from user satisfaction to IS use.

In the field of VCs, specific Web site quality fac-tors are believed to be critical in affecting the usageof VCs.5,13 However, frequency has received onlylimited attention in the empirical Seddon model forexplaining the determinants of VCs’ success. Over-all, the three dimensions of VC success—systemquality, information quality, and PU—measure dif-ferent aspects of online communication efficiency.While system and information quality measure thefunctionality and information content of VCs, rela-tive PU measures the benefits of VCs. Thus the Sed-don model extends TAM constructs to the Web sitequality factors.

Integrated model

Recently, Agarwal and Karahanna17 related TAMas an individual’s perception of a new technology

based as the instrumental perspective, focusing onmotivational drivers such as usefulness and ease ofuse. In the context of VCs, PU refers to the degreeto which users perceive a subjective probability thatthe VC will increase its performance in informationexchange, while PEOU represents the degree towhich a VC is perceived to be easy to understand,learn, or operate. Expanding Agarwal and Kara-hanna’s17 proposal that TAM explains user accep-tance of a Web site, the Seddon model measuresWeb site quality features, which include Web sitefunctionality and user-based information contents.Further, Web site quality refers to the results thatusers want, which can be obtained with interactiveprocesses between online businesses and users.Web site quality features can expand the usefulnessconstruct beyond individual performance duringVC use. Thus, the Seddon model complements TAMin capturing user perceptions of VCs.

The integration of constructs can also be viewedfrom the perspective of online social interaction inVC participation. A VC is an online communicationchannel, typically involving several activities. Thefirst step often involves the community membersobtaining and sharing information in the VC andlearning what (presumably credible) others think.This step often generates a comfortable virtual en-vironment in which users can easily identify func-tional groups and navigation aids, thus ensuring ef-ficient information exchange. The next step usuallyinvolves the community members accomplishingspecific tasks, such as problem solving, generatingnew ideas, influencing the opinions of others re-garding hot topics, and validating previouslyreached decisions, through the VC. This online com-munication channel is captured appropriately byconstructs in the TAM and Seddon model. Severalprevious studies have applied these constructs,though in a unidimensional approach. For example,Teo et al.3 found that willingness to use VCs deter-mined user beliefs (PU and PEOU). Moreover, thelink between online quality factors and user satis-faction is well established in the VC literature.4 Inreviewing the theoretical development of related lit-eratures, this study integrates these two models andexamines their intrinsic relationships with user sat-isfaction and loyalty in VCs.

METHODS

Participants and procedure

Data for the study were collected using a paper-based survey administered in class to 236 students

LIN140

registered in five business courses at a large uni-versity in the north Taiwan. Of the 236 question-naires distributed, 230 completed and usable ques-tionnaires were received. The first question askedrespondents whether they had any experience in us-ing VCs. Those who responded no to this questionwere not included in the analysis, since the studyfocused only on VC members who shared informa-tion and knowledge with other members. As a re-sult, 32 questionnaires were excluded from the dataanalysis stage. Of the entire 198 respondents, 60.1%were female, 39.9 were male. The majority of re-spondents were 21 to 25 years old (74%). Over one-half of the respondents had used the Internet formore than 3 years. More than 75% had been com-munity members for more than 1 year. Moreover, asample size exceeding 190 is acceptable, since sam-ple sizes ranging from 150 to 20018 or more19 are al-ready sufficient to generate statistically reliable es-timates of the causal paths among constructs.

Measures

PU and PEOU were measured by four items eachfrom the previously validated inventory8,9 andmodified to suit the context of VCs. Satisfaction withVC was assessed using members’ opinions regard-ing the VC. Member loyalty was measured usingtwo items taken from Yoo et al.5 that measured theextent to which members were involved in the VC.System and information quality were developed tomeasure Web site quality from Lin and Lee’s4 mea-sures. All items were measured using a 5-point Lik-ert-type scale (1, strongly disagree to 5, strongly agree).

Statistical analysis

Data analysis utilized a two-step approach as rec-ommended by Anderson and Gerbing.18 The firststep involves the analysis of the measurement

model; the second step examines the model fit re-sults of the proposed theoretical models (TAM, Sed-don model, and the integrated model).

RESULTS

Measure reliability and validity

The research instrument used confirmatory factoranalysis (CFA) to examine reliability and validity. In-ternal consistency reliability to test unidimensional-ity was assessed by Cronbach’s alpha. The resultingalpha values ranged from 0.79 to 0.88, which wereabove the acceptable threshold 0.70 suggested byNunnally and Bernstein.20 Convergent validity can beassessed by factor loadings from the CFA; all factorloadings should be exceed 0.6.21 All of the factor load-ing items in the research model ranged from 0.67 to0.91. Therefore, the measurement model had ade-quate reliability and convergent validity.

Structural model results

Figure 2 displays all of the structural relationshipsamong the studied constructs in TAM. Path coeffi-cients and their significance and the variance (R2

value) for each dependent construct are also presentedin this figure. Figure 3 shows information for the Sed-don model. Figure 4 displays information for the in-tegrated model. Additionally, Figures 2, 3, and 4 showthe fit statistics for each structural model. Overall, thethree structural models displayed a good fit with thedata, compared with the suggested fit criteria.18

DISCUSSION

The main objective of this study was to empiri-cally compare two of the most dominant contem-


0.23**

Perceived usefulness

Perceived ease of use

Satisfaction with VC

Member loyalty

χ2/df = 473.69/233 = 1.96; AGFI = 0.86; CFI = 0.95; NNFI = 0.92; RMSEA = 0.062

0.32**

0.30***

(0.37)

(0.52) (0.45)

0.45***

FIG. 2. Results of TAM. Path significance: *p � 0.05; **p � 0.01; ***p � 0.001. Parentheses indicate R2 value.

porary models used in research on IS success, theTAM and Seddon model, and then extend thesemodels to produce an integrated model to better un-derstand the phenomena of our interest: satisfactionwith and loyalty in VCs. As shown in Figures 2, 3,and 4, all three models achieve comparable fit to thedata. Consequently, it is reasonable to examine themodels in terms of their path significance and ex-planatory power. First, based on the theory of com-parison approach,22,23 this study adopted reason-able fit and explanatory power to evaluate threecompeting models and identify the best. In fact, thepercentage of variation in satisfaction explained bythe integrated model is 59%, which is significantlyhigher than any of the models examined separately.Second, TAM is well established in the IT litera-ture.8,9 The findings generally support the results ofTAM-based online service studies1–3 and indicatethat satisfaction with VCs leads to member loyalty.

The findings of this study support that PU andPEOU are key determinants of user attitude, as-sessed by their satisfaction with VCs (Figure 2).PEOU also exerted an indirect effect on satisfactionvia PU of VCs. Third, after users deem the VC to beacceptable and select to continue to use the VC; ef-fective online support offered by VC providers in-fluences their continued satisfaction. The Seddonmodel and the integrated model found the influenceof quality-perception dimensions (e.g., system qual-ity and information quality) on user satisfaction tobe significant. This result implies that when usersfind VCs with high levels of valuable functions anduseful contents, they will be satisfied with their gen-eral effectiveness and efficiency. Another interest-ing observation is that comparison of the separatemodels revealed increased strength of certain con-structs in the integrated model. Specifically, in theintegrated model, system and information quality

LIN142

Member loyalty

System quality




0.30**

0.20** (0.48) (0.45)

Information quality

0.31*** 0.15*

0.45***

0.22**

(0.35)

Member loyalty

System quality

Perceived ease of use



0.27**

0.23**

(0.59) (0.45)

Information quality

0.21**

0.30*** 0.45***

(0.40)


0.34***

0.17*

0.26**

FIG. 3. Results of Seddon model. Path significance: *p � 0.05; **p � 0.01; ***p � 0.001. Parentheses indicate R2 value.

FIG. 4. Results of the integrated model. Path significance: *p � 0.05; **p � 0.01; ***p � 0.001. Parentheses indicate R2

value.

received increased support in explaining VC satis-faction. This result further supports the validity ofthe Web site quality factors. In summary, users willfind the VC experience gratifying only if the VC isuseful, easy to use, and has a high-quality Web sitedesign. The results of the theory-driven analysissupport the robustness of the VC metrics drawnfrom the TAM and Seddon model, and provide abasis for future research, in addition to offeringguidance to VC providers regarding the antecedentsof VC satisfaction and loyalty.

The implications of the findings of this study canbe broadly divided into two categories: theoreticalcontribution to VC researchers and implications forVC providers. As a theoretical contribution, this studysimultaneously examines TAM, the Seddon model,and an integrated model (combining the TAM andSeddon model) in the VC domain, specifically userexperiences and attitudes toward VCs. This work pro-vides a theoretical basis for further research on theboundaries and extensions of these theoretical mod-els. Furthermore, this study found that success crite-ria used in IS research can also be applied to the VCenvironment. For example, the TAM components PUand PEOU are important determinants of satisfactionand loyalty in VCs. Additionally, this study has en-riched the VC satisfaction model by introducing con-structs based on Web site quality factors. Method-ologically, this study compared the TAM, Seddonmodel, and an integrated model to examine whichmodel best helps to predict user satisfaction with theVC. The results indicated that the integrated modelprovided better explanatory power than either theTAM or the Seddon model alone.

From a managerial perspective, the new digitaleconomy comprises an increasing number of VCs,and online businesses are encouraged to establishVCs to cultivate more personal relationships withconsumers. To sustain a successful VC, VC pro-viders need to focus their attention on designingboth useful and easy-to-use Web sites. Moreover,the results of this study found that system and in-formation quality are critical antecedents to VC sat-isfaction, which in turn influences member loyalty.Online businesses and VC providers thus shouldconsider various design and implementation strate-gies for attracting more members. For example, theWeb site should provide for online communicationtechniques among community members and re-spond rapidly to the search and browsing needs ofusers. A good Web site should not only be user-friendly but should also provide high-quality in-formation. The information provided in the VCmust be accurate, complete, current, customized tothe user, and presented in an easy-to-use format.

This research has several limitations. First, otherindividual differences (such as age, level of educa-tion, and VC experiences) may affect user satisfac-tion and loyalty in VCs. Future studies can incorpo-rate these variables into the research model. Second,the data were gathered from a relatively homoge-nous demographic group, namely college students.Caution is necessary when generalizing the findingsof this study to a broader population group. How-ever, it may be appropriate to generalize the findingsto other similarly aged Internet users.

REFERENCES

1. Ahn T, Ryu S, Han I. The impact of the online and of-fline features on the user acceptance of Internet shop-ping malls. Electronic Commerce Research and Ap-plications 2004; 3:405–20.

2. Lin HF. Understanding behavioral intention to par-ticipate in virtual communities. CyberPsychology &Behavior 2006; 9:540–47.

3. Teo HH, Chan HC, Wei KK, Zhang Z. Evaluating in-formation accessibility and community adaptivityfeatures for sustaining virtual learning communities.International Journal of Human–Computer Studies2003; 59:671–97.

4. Lin HF, Lee GG. Determinants of success for onlinecommunities: an empirical study. Behaviour & Infor-mation Technology 2006; 25:479–88.

5. Yoo WS, Suh KS, Lee MB. Exploring the factors en-hancing member participation in online communities.Journal of Global Information Management 2002;10:55–71.

6. Cothrel J, Williams R. Understanding online commu-nities. Strategic Communication Management 1999;3:16–21.

7. Ajzen I, Fishbein M. (1980). Understanding attitudes andpredicting social behavior. Englewood Cliffs, NJ: Pren-tice Hall.

8. Davis FD. Perceived usefulness, perceived ease of use,and end user acceptance of information technology.MIS Quarterly 1989; 13:319–40.

9. Davis FD, Bagozzi RP, Warshaw PR. User acceptanceof computer technology: a comparison of two theo-retical models. Management Science 1989; 35:982–1003.

10. Seddon PB. A respecification and extension of the De-Lone and McLean model of IS success. InformationSystems Research 1997; 8:240–53.

11. DeLone WH, McLean ER. Information system suc-cess: the quest for the dependent variable. Informa-tion Systems Research 1992; 3:60–95.

12. Zviran M, Erlich Z. Measuring IS user satisfaction: re-view and implications. Communication of the AIS2003; 12:81–104.

13. Preece J. Sociability and usability in online commu-nities: determining and measuring success. Behaviour& Information Technology 2001; 20:347–56.


14. Oliver R. A cognitive model of the antecedents andconsequences of satisfaction decision. Journal of Mar-keting Research 1980; 22:460–69.

15. Bhattacherjee A. An empirical analysis of the an-tecedents of electronic commerce service continuance.Decision Support Systems 2001; 32:201–14.

16. Moore GC, Benbasat I. Development of an instrumentto measure the perceptions of adopting an informa-tion technology innovation. Information Systems Re-search 1991; 2:192–222.

17. Agarwal R, Karahanna E. Time flies when you’re hav-ing fun: cognitive absorption and beliefs about infor-mation technology usage. MIS Quarterly 2000; 24:665–94.

18. Anderson JC, Gerbing DW. Structural equationmodeling in practice: a review and recommendedtwo-step approach. Psychological Bulletin 1988; 103:411–23.

19. Kline RB. (1988). Principle and practice of structuralequation modeling. New York: Guilford Press.

20. Nunnally JC, Bernstein IH. (1994). Psychometric theory.New York: McGraw-Hill.

21. Chin WW, Gopal A, Salisbury WD. Advancing thetheory of adaptive structuration: the development ofa scale to measure faithfulness of appropriation. In-formation Systems Research 1997; 8: 342–67.

22. Chau PYK, Hu PJH. Information technology accep-tance by individual professional: a model comparisonapproach. Decision Sciences 2001; 32:699–719.

23. Taylor S, Todd .A. Understanding information tech-nology usage: a test of competing models. Informa-tion Systems Research 1995; 6:144–76.

Address reprint requests to:Dr. Hsiu-Fen Lin

Department of Shipping and Transportation Management

National Taiwan Ocean UniversityNo.2, Beining Road

202-24 Keelung, Taiwan, Republic of China

E-mail: [email protected]

LIN144

antecedents of virtual community satisfaction and loyalty: an empirical test of competing theories

Documents