Which Size Matters? Effects of Crowd Size on Solution Quality in Big Data Q&A Communities

Abstract

Question Answering websites have evolved into one of the most important platforms for knowledge sharing and problem solving online. Despite widespread adoption of Q&As by technical communities as well as an abundance of domain experts, many questions fail to attract a sufficient audience to obtain a good solution or any solution at all. We investigate the effects of crowd size on solution quality in Stack Exchange Q&A communities on topics related to big data. We find that three distinct levels of group size in the crowd (topic audience size, question audience size, and number of contributors) affect solution quality. Therefore, we argue that group size in the crowd is not unitary, but rather a multi-level construct. This work advances a theoretical model of group size in the crowd and the relation between crowd size and performance. The work also has practical implications for system designers trying to route crowds to problems efficiently.