Abstract
In order to motivate individuals to share their knowledge in online communities, the use of extrinsic rewards and goals is a typical approach. However, extrinsic motivation may have unintended consequences. Although past studies have examined the direct effect of extrinsic motivation on intrinsic motivation, no research to date has investigated how extrinsic motivation moderates the impact of intrinsic motivation on knowledge sharing, or how the effect of extrinsic motivation on intrinsic motivation is contingent upon whether a member is active or not. Drawing on attribution theory and theory of planned behavior, the study was conducted with data collected from a large social Q&A site consisting of multiple online communities with millions of registered users; the data were analyzed with moderated regression and structural equation modeling. Results show that the effect of enjoyment in helping others on attitude toward knowledge sharing is undermined by virtual organizational rewards, while the effect of knowledge self-efficacy on attitude toward knowledge sharing is undermined by reciprocity. The results also show that the effect of virtual organizational rewards on enjoyment in helping others is contingent upon whether members are active or not. Specifically, for active members, virtual organizational rewards undermine enjoyment in helping others; for inactive members, however, virtual organizational rewards increase enjoyment in helping others. These findings enrich the research on unintended consequences of extrinsic motivation specifically, and the theory of motivation in general. Additionally, these findings provide practical insights on how and when to use extrinsic rewards/goals to motivate individuals to share knowledge in social Q&A sites.
Notes
1. The community allows members to use pseudonyms in order to attract more people to participate in the community. Hence, some members are anonymous in the community whereas others reveal their real identity.
2. The survey items were translated into Chinese by a native speaker of Chinese who is bilingual. The survey items were translated following Brislin’s [Citation11] guidelines. To verify the accuracy of this translation, the Chinese survey was translated back into English by another bilingual person with native English-speaking fluency and the two English surveys were compared. After comparing the two versions of items several times, any discrepancies found in the wording of the translated surveys were discussed and resolved.
3. Given that any stranger can create an account to register and leave online communities freely, the site did not have accurate demographic information, thereby making it impossible to compare respondent demographics with the demographics of the site population. The researchers consulted a manager of the site, who said that more than 90 percent of members are male. Thus, this sample reflected the male-dominated site population.
4. The statistical significance of the difference between the slopes of two regression lines is indicated by the significance of the interaction coefficient produced by the moderated regression.
5. Further, we used another SEM approach, Mplus, to test this structural model, which produced similar results. (See Appendix G for the results of the Mplus test.)
6. In statistics, each predictor is assigned only the separate portion of the criterion variable’s variance it accounts for (i.e., its unique contribution), whereas the same portion of the criterion’s variance predicable from different predictors is not assigned to any of the predictors [Citation72].
Additional information
Notes on contributors
Li Zhao
Li Zhao ([email protected]; corresponding author) is a faculty member in the Department of Information Processing Science at the Faculty of Information Technology and Electrical Engineering at the University of Oulu. He holds a Ph.D. in information systems from McMaster University in Canada. His research interests include knowledge management, security, online games, social networks, social media, and e-marketing. His research has been published in leading journals and conferences in the information systems field.
Brian Detlor
Brian Detlor ([email protected]) is chair and associate professor of information systems at the DeGroote School of Business at McMaster University in Canada. He received his doctoral education at the University of Toronto’s iSchool. His research interests lie in the interaction of users, information, and technology. His scholarly work bridges the fields of information systems (IS) and library and information science (LIS), and he has published extensively in leading journals, conferences, monographs, and textbooks in these fields. His research projects investigate topics such as digital storytelling, information literacy, and the adoption and use of information systems.
Catherine E. Connelly
Catherine E. Connelly ([email protected]) is a tier II Canada Research Chair and associate professor of organizational behavior at the DeGroote School of Business at McMaster University. Her research focuses on the attitudes and behaviors of nonstandard workers, employees’ use of communication technologies, and knowledge sharing and hiding in organizations.