“Optimal vs. Rational: Bias in and Debiasing Solver Behaviors in Crowdsourcing Ideation Contests”

Crowdsourcing ideation contests allow idea-seeking organizations to solicit solutions from external solvers to address specific problems. In this study, we present a tension between optimal solver behaviors and rational solver behaviors, specifically in the context of solvers’ use of developmental feedback from different sources (seeker/peer) and of different practicality (high/low). In Experiment 1, we show a source bias in solvers’ use of feedback, where they base their ideas more on seeker feedback than on peer feedback. Such behaviors, possibly considered as rational in ideation contests, could lead to suboptimal behaviors such as an underuse of high practicality peer feedback and overuse of low-practicality seeker feedback. In Experiment 2, we test the effectiveness of a feedback rating mechanism in debiasing the source bias in solvers’ use of feedback. We show that the mechanism can correct the suboptimal behaviors in solvers’ use of feedback, particularly for perceptive solvers. In a way, the mechanism can cause solvers to behave less rationally (in terms of their self-interest) but more optimally (in terms of submitting better ideas) in contests. This study contributes by providing insights into solver behaviors in ideation contests and proposing a bias-mitigating mechanism design for contest platforms to enhance the ideation process.
Keywords: crowdsourcing, ideation contests, feedback, source bias, feedback rating mechanism, debias, experiment

“The Effect of Identity Disclosure on Online Community Participation: A Natural Experiment” with Bingjie Qian and Michael Zhang

When participants of online communities click the “like” button, are they trying to maintain a social relationship or their own social image? Since we cannot directly observe user motivations, these two effects are usually confounded. In this study, we leverage a natural experiment to examine these two forces. Our identification strategy relies on an exogenous policy change of a Chinese online community, douban.com. Users who “liked” an article were anonymous before the change; the identities of such users were publicly disclosed after the change. We identify three motivations from prior literature that are relevant in our context: intrinsic utility, relational, and image. We leverage that these three motivations predict different user reactions to the policy change to identify the dominating motivation. We find that users gave significantly fewer “likes” after the policy change, which is consistent with the image motivation prediction. The analyses of the content and user characteristics provide additional support for the dominating role of image motivation. This paper provides insights into user behavior in online communities and has practical implications for both platforms and content creators.
Keywords: identity disclosure, online community participation, social image, social relationship, intrinsic utility

“Seeker Exemplars and Solver Behaviors in Crowdsourcing Contests: Impacts on Quantitative Ideation Outcomes” with Muller Cheung

Idea seekers in crowdsourcing ideation contests often provide solution exemplars to guide solvers in developing ideas. We examine how such seeker exemplars affect the quantitative ideation outcomes in the scanning, shortlisting, and selection of ideas by solvers; these ideation activities relate to the Search and Evaluate stage of the Knowledge Reuse for Innovation model. We theorize that solvers’ prior belief, aim of satisfying seekers’ preferences, and effort and time concerns in contests shape their use of local (problem-related) and distant (problem-unrelated) seeker exemplars in the respective ideation activities. Consequently, solvers demonstrate a focus-switching behavior during ideation, whereby they attend more to local seeker exemplars when searching and shortlisting ideas but actively consider both local and distant seeker exemplars when selecting ideas to submit for the contests. The results from an online ideation contest experiment indicate that solvers could search for, shortlist, and/or submit fewer ideas when certain seeker exemplars are shown. Thus, although showing seeker exemplars in ideation contests is common and facilitates idea generation, doing so can also negatively affect idea quantity, which, according to prior research, could impair idea quality. We discuss the theoretical and practical implications of our findings.
Keywords: crowdsourcing, ideation contests, confirmation bias, online experiment, Knowledge reuse for innovation