Copy the page URI to the clipboard
Huo, Qunying
(2017).
DOI: https://doi.org/10.21954/ou.ro.0000cdea
Abstract
One challenge faced by online forums is the provision of a sustainable supply of contributions of knowledge (Wasco et al., 2009). Previous studies have identified online trust and perceived critical mass as antecedents of online knowledge contributions. However, the dynamic aspects of antecedents are little investigated. Moreover, how the dynamics together impact on members’ willingness to contribute knowledge is an open question to be further investigated.
To examine the dynamic antecedents of online knowledge continuance, this thesis seeks to develop a holistic approach through three studies. Drawing on a decomposed theory of planned behaviour (Taylor and Todd, 1995), study one identifies dynamic antecedents of intentional online contribution behaviours. Covariance-based structural equation modelling analysis of 910 responses obtained shows that perceived critical mass and trust in online forums that mediates trust in members are the highlighted antecedents in the context of online forums. The development of trust in online forums is investigated through a time series approach in study two. Findings using webnographic and machine learning analysis show that the cognitive dimension of institutional trust is essential in initial trust building. Study three uses network analysis techniques to explore the role of critical mass members. Results indicate that only 5% of critical mass members can sustain online forums. However, critical mass members compete for their connections, inferring the importance of brand building in the beginning of online forums development. A summary of findings from the three studies suggests that the structure assurance of online forums can mediate the effects of interactions between members to a coalition of membership over time. The study provides further knowledge on the voluntary contribution within online forums by taking a dynamic approach, while previous studies in this field are predominantly cross-sectional and un-prophetic.