Developing a Sense of Virtual Community Measure
Post on 10-Mar-2017
CYBERPSYCHOLOGY & BEHAVIORVolume 10, Number 6, 2007 Mary Ann Liebert, Inc.DOI: 10.1089/cpb.2007.9946
Developing a Sense of Virtual Community Measure
ANITA L. BLANCHARD, Ph.D.
Sense of virtual community is an important feature of virtual communities. This study de-velops a sense of virtual community (SOVC) measure, building off the strengths of a widelyused measure of sense of community (SOC) for face-to-face communities. Although there isoverlap between the senses of community for face-to-face and virtual communities, there aresignificant differences. The new SOVC measure is compared to the SOC measure on 265 mem-bers of seven online groups, explaining at least 7% more of the variance from exchangingsupport and member identification. This study represents an important step in developing avalid measure of SOCV.
DEVELOPING A MEASURE OF SENSE OFVIRTUAL COMMUNITY
SENSE OF VIRTUAL COMMUNITY (SOVC) is recog-nized as a significant feature of virtual commu-nities.14 SOVC is defined as members feelings ofmembership, identity, belonging, and attachment toa group that interacts primarily through electroniccommunication. SOVC assesses the community-ness of virtual communities; it distinguishes vir-tual communities from other types of virtualgroups. However, there is not currently a good mea-sure of it.
The purpose of this study is to present a measureof SOVC based on the current measure of face-to-face sense of community (SOC) and taking into ac-count the unique features of sense of virtual com-munity. SOC consists of feelings of membership,feelings of influence, integration and fulfillment ofneeds, and shared emotional connection.5 However,research demonstrates that members of virtual com-munities may have less pronounced feelings of in-
fluence than do members of face-to-face communi-ties.1 Additionally, they feel that they know the per-sonalities of others and experience and observemore personal relationships than do members offace-to-face communities. Thus, there are differ-ences between SOVC and SOC.
These differences are significant. Most re-searchers currently adapt the most current measureof SOC, the sense of community index (SCI),6 to vir-tual groups.2,4 This practice creates a problem be-cause the SCI may not have sufficient content va-lidity as a measure of SOVC. That is, there may beitems in the SCI that do not have relevance in vir-tual communities, and the SCI may be missing partsof the SOVC construct domain. Thus, the SCI wouldbe a less sensitive measure of SOVC. An improvedmeasure of SOVC would take into account theunique components of sense of community in vir-tual communities and therefore represent an im-provement in measuring SOVC.
This study examined potential antecedents ofSOVC, exchanging support1,711 and identifica-
Department of Psychology, University of North Carolina, Charlotte, North Carolina.
tion1,1219 to determine if the newer SOVC measureis an improvement over the SCI.
Participants were 265 members of seven onlinegroups: three listservs and four Usenet newsgroups.The topics of the groups ranged from pet lovers tohuman resource professionals. Five of the groupshad mainly social topics and two had mainly pro-fessional topics. To be considered for this study, thegroups had to be active with messages posted daily.The mean age of the participants was 47 (SD 11.7).Of the participants, 62% were female.
Sense of virtual community. Twenty-two itemswere used to assess sense of community (see ap-pendix). Twelve items from the SCI were adaptedby changing the words block to group and neighborsto members.6 Ten additional items were created toassess the unique components of SOVC and to im-prove the clarity of the SOC construct, a problemnoted by other researchers.2022 These items werepilot tested for clarity and appropriateness by mem-bers of another online group. Minor changes inwording addressed the pilot groups feedback. Re-sponses to the SOVC ranged from 1, strongly dis-agree, to 4, strongly agree.
Exchanging support. To assess exchanging sup-port, three measures were created: one to assess ob-servation of others exchanging support (two items),one to assess the participants exchanging supportthemselves through e-mail (four items), and one toassess participants exchanging support by posting
to the group (four items). Responses for all itemsranged from 1, never, to 4, quite a lot, with internalreliabilities of 0.85, 0.88, and 0.85 respectively.
Identification. Two items were used to assessidentification. Responses ranged from 1, stronglydisagree, to 4, strongly agree. The reliability coeffi-cient of this scale was 0.80.
First, the SOVC items were factor-analyzed usingmaximum likelihood factoring and a promax rota-tion. Three factors were extracted to determine ad-equate fit. All the items loaded adequately, and thethree-factor structure seemed appropriate. Com-munalities were also examined, and two items fromthe SCI (Q2 and Q11) were eliminated because ofvery low communalities. The remaining items metall the criteria for inclusion in the factor analysis.The correlations between the three extracted factorswere all above 0.64, indicating the factors could becombined for one overall measure of SOVC.
Next, to ensure that the items developed for thisstudy all factor onto their respective scales, anotherfactor analysis was conducted with the SOVC itemsand the exchanging support and identificationitems. Two items from the original SCI, Q6 and Q8,inappropriately loaded onto the exchanging behav-ior by posting factor. These items were thereforedropped from the analysis. (The results from thesetwo factor analyses are available from the author.)All other items loaded onto their appropriate scales.The internal reliability coefficient for the SOVC scalewas 0.93.
Table 1 contains the comparison of the regressionof the antecedents on SOVC and SOC. All the vari-ables are strongly related to both SOC and SOVC.However, the antecedents explain 53% of the vari-
TABLE 1. REGRESSIONS COMPARING SOVC TO SOC
B SE B SE
Observe support 0.21 0.05 0.25* 0.15 0.04 0.21*E-mail support 0.19 0.03 0.32* 0.14 0.03 0.27*Post support 0.20 0.04 0.29* 0.19 0.04 0.32*Identification 0.24 0.05 0.25* 0.22 0.05 0.26*R-squared 0.53 0.46
ance of SOVC and only 46% of the variance of SOC.Thus, SOVC is a more sensitive measure than SOC,accounting for 7% more of the variance, suggesting,therefore, improved content validity.
The purpose of this study was to demonstratethat a new measure of sense of community in vir-tual communities has increased content validity andsensitivity over the traditional measure of face-to-face SOC (i.e., the SCI) in virtual groups. Althoughthere is overlap between the SOC and SOVC con-structs, there is also evidence of significant differ-ences.1 In this study, 22 items were assessed for in-clusion in the new SOVC measure. Four items fromthe SCI were eliminated because of poor factor load-ings, poor reliability (i.e., communality), and inap-propriate loading onto other study variables. Theresulting 18 items in the SOVC measure are evenlydivided between the SCI and newly developed mea-sures. Some of the newly developed measures ad-dress specific differences of the SOVC construct ascompared to the SOC construct, and others addresssome of the shortcomings of the SOC.2022
The SOVC measure is a more sensitive measureexplaining at least 7% more of the variance than thetraditional SOC. Although simply adding variablesto a measure will increase the amount of varianceexplained, these new items fit well conceptually andempirically into the new measure of SOVC. Also,because the new SOVC measure added over 7% ofvariance explained to an already substantial 46% ofthe variance, the new items are not likely to be triv-ial components of SOVC. Greater sensitivity is im-portant, particularly in understanding the mecha-nisms that create SOVC. Using a more sensitive andcontent-valid measure, researchers will be able todevelop models of the antecedents, mediators, andmoderators of SOVC.
The validation of a measure is a lengthy processrequiring multiple studies over a period of time,which allows for an integrated evaluation of the em-pirical evidence and theoretical rationale of the ad-equacy and appropriateness of the measure.23 Thisstudy is not purported to be the last word on de-veloping an appropriate measure of SOVC. How-ever, a dialogue must be started among virtual com-munity researchers about the appropriateness ofsimply transposing the SCI from face-to-face to vir-tual communities. This study is a significant firststep in that process.
Q1. I think this group is a good place for me to bea member.
Q2. Members of this group do not share the samevalues.*
Q3. Other members and I want the same thing fromthis group.
Q4. I can recognize the names most members in thisgroup.
Q5. I feel at home in this group.Q6. Very few other group members know me.*Q7. I care about what other group members think
of my actions.Q8. I have no influence over what this group is like.*Q9. If there is a problem in this group, there are
members here who can solve it.Q10. It is very important to me to be a member of
this group.Q11. Members of this group generally dont get
along with each other.*Q12. I expect to stay in this group for a long time.Q13. I anticipate how some members will react to
certain questions or issues in this group.Q14. I get a lot out of being in this group.Q15. Ive had questions that have been answered by
this group.Q16. Ive gotten support from this group.Q17. Some members of this group have friendships
with each other.Q18. I have friends in this group.Q19. Some members of this group can be counted
on to help others.Q20. I feel obligated to help others in this group.Q21. I really like this group.Q22. This group means a lot to me.
* Deleted from the final SOVC measure
1. Blanchard AL, Markus ML. The experienced senseof a virtual community: characteristics and processes.The DATA BASE for Advances in Information Sys-tems 2004; 35:6579.
2. Forster PM. Psychological sense of community ingroups on the Internet. Behaviour Change 2004;21:1416.
3. Koh J, Kim Y-G. Sense of virtual community: a con-ceptual framework and empirical validation. Interna-tional Journal of Electronic Commerce 2003; 8:75.
4. Obst P, Zinkiewicz L, Smith SG. Sense of communityin science fiction fandom, Part 2: comparing neigh-
VIRTUAL COMMUNITY MEASURE 829
borhood and interest group sense of community.Journal of Community Psychology 2002; 30:10517.
5. McMillan DW, Chavis DM. Sense of community: adefinition and theory. Journal of Community Psy-chology 1986; 14.
6. Chavis DM et al. Sense of community throughBrunswicks lens: a first look. Journal of CommunityPsychology 1986; 14:2440.
7. Baym N. (1995) The emergence of community in com-puter mediated communication. In: Jones SG, ed. Cy-bersociety: computer mediated communication and com-munity. Thousand Oaks, CA: Sage.
8. Bergquist M, Ljungberg J. The power of gifts: orga-nizing social relationships in open source communi-ties. Information Systems Journal 2001; 11:30520.
9. Rothaermel FT, Sugiyama S. Virtual Internet com-munities and commercial success: individual andcommunity-level theory grounded in the atypical caseof TimeZone.com. Journal of Management 2001;27:297312.
10. Turner JW, Grube JA, Meyers J. Developing an opti-mal match within online communities: an explorationof CMC support communities and traditional sup-port. Journal of Communication 2001; 51:23151.
11. Weis R et al. Communities of care and caring: the caseof MSWatch.com. Journal of Health Psychology2003; 8:13548.
12. Douglas KM, McGarty C. Identifiability and self-pre-sentation: computer-mediated communication andinter-group interaction. British Journal of Social Psy-chology 2001; 40:399416.
13. Joinson AN. Self-disclosure in computer-mediatedcommunication: the role of self-awareness and visualanonymity. European Journal of Social Psychology2001; 31:17792.
14. Kiesler S, Seigel J, McGuire T. Social psychological as-pects of computer-mediated communication. Ameri-can Psychologist 1984; 39:112334.
15. Ma M. (2004) An identity based theory of informationtechnology design for sustaining virtual communi-ties. In Proceedings of the Twenty-fifth International
Conference on Information Systems. Washington,DC.
16. McKenna KYA, Bargh JA. Coming out in the age ofthe Internet: identity demarginalization throughvirtual group participation. Journal of Personality andSocial Psychology 1998; 75:68194.
17. Postmes T, Spears R, Lea M. The formation of groupnorms in computer-mediated communication. Hu-man Communication Research 2000; 26:34171.
18. Postmes T et al. Individuality and social influence ingroups: inductive and deductive routes to groupidentity. Journal of Personality and Social Psychology2005; 89:74763.
19. Riva G. The sociocognitive psychology of computer-mediated communication: the present and future oftechnology-based interactions. CyberPsychology &Behavior 2002; 5:58198.
20. Chipuer HM, Pretty GH. A review of the sense ofcommunity index: current uses, factor structure, reli-ability and further development. Journal of Commu-nity Psychology 1999; 27:64358.
21. Long DA, Perkins DD. Confirmatory factor analysisof the sense of community index and development ofa brief SCI. Journal of Community Psychology 2003;31:27996.
22. Obst P, White KM. Revisiting the sense of communityindex: a confirmatory factor analysis. Journal of Com-munity Psychology 2004; 32:691705.
23. Guion RM. (2002) Validity and reliability. In: Rogel-berg SR, ed. Handbook of research methods in industrialand organizational psychology. Malden, MA: Blackwell.
Address reprint requests to:Dr. Anita L. Blanchard
Department of PsychologyUniversity of North Carolina Charlotte
9201 University City BoulevardCharlotte, NC 28223