Building Loyalty to Online Communities through Bond and Identity-based Attachment to Sub-Groups Yla R. Tausczik Laura A. Dabbish Robert E. Kraut Carnegie Mellon University Sustaining Online Communities Wikipedia is not a social networking service like Facebook or Twitter. - Wikipedia Guidelines Social Design Features Social Attachment Community Loyalty Bond- based Identity-based Interpersonal Attraction Group Identification Bond- basedIdentity-based Small Group Attachment Sassenberg (2002) IRC channels Farzan et al. (2011) Groups of Tetris players Ren et al. (2012) MovieLens users Goals 1.Show compelling evidence that attachment to subgroups underlies the effect of social design features on site loyalty 2.To evaluate whether there are two distinct types of attachment Study 1 Study 1 - Hypotheses Hypothesis 1: Assigning Turkers to communicating groups will increase their loyalty to Mechanical Turk and their employer by increasing feelings of attachment to work groups. Hypothesis 2: Structuring group work and communication in pairs or as a group as a whole will differentially create bond or identity-based attachment respectively. Study 1 - Method Participants 606 participants 52% female 18 to 68 years old (M = 31.5) 76% US, 19% India, 5% Other 85.0% completion rate Study 1 - Method Procedure HIT Group Assignment TaskQuestionnaires Study 1 - Method Independent Variables Level of Interaction (Independent, Awareness, Communication) Work Group Type (Dyadic, Group) Dependent Variables Community Loyalty Self-Reported Attachment (Bond, Identity) Study 1 - Community Loyalty Log likelihood ratio test = 12.4, p = 0.002, N = 369, Groups = 123 ** Study 1 Mediation Analysis Employer Commitmen t Level of Interaction t(105) = 2.91, p < Subgroup Attachment Study 1 Mediation Analysis Subgroup Attachment Employer Commitmen t Level of Interaction t(105) = 9.06, p
