Factors Influencing the Response Rate in Social Question & Answering Behavior

Despite the power of social networking sites (SNSes) in routing questions to users’ social networks, there is a relatively low response rate to social Q&A. We ask: are SNS’es really suitable to handle question answering?

 

examples of social Q&A  on Weibo

examples of social Q&A on Weibo

We investigated this question by examining factors that may influence the response rate on Weibo, China’s largest microblogging site.  Based on previous literature, we selected eleven extrinsic variables, consisting of both context features and content features. On Weibo, we found that the number of followers, the number of accounts @-mentioned, and the adoption of emoticons all positively affect the response rate to a question. To our surprise, we also found that questioners who update more than 10 times per day tend to receive fewer answers, and questions with @-mentions to verified accounts usually get very low or no responses. By classifying questions into categories based on their topics as well as their posting time periods, we noticed that location-based questions attract the most responses, whereas, questions posted during 0:00AM – 6:00AM received the least average number of replies.

With a deeper look into those factors with negative or no effect on response rate, we concluded two major problems in the the current design of SNS’es.  First, due to the heterogeneous hashtag vocabulary used in the context of social Q&A, hashtags hardly have any value in increasing the visibility of one’s question as compared to its performance in event tracking and broadcasting.  Second, due to the current lack of topical expertise judging mechanisms on SNS, questioners demonstrated difficulties in targeting the authorities who can offer them trustworthy answers. Without expertise evaluation, the value of verified accounts in social Q&A was small as compared to their important function in the process of information diffusion. Based on the present situation, we believe that the development of corresponding tools and mechanisms, such as interrogative hashtag vocabularies, expert authentication, recommendation, and compensation, could make the user’s expert seeking process easier and the experts more willing to share knowledge with others.

For more, see our full paper, Factors Influencing the Response Rate in Social Question and Answering Behavior

Zhe Liu, The Pennsylvania State University
Jim Jansen, The Pennsylvania State University