Great Question! Question Quality in Community Q&A

Abstract

Asking the right question in the right way is an art (and a science). In a community question-answering setting, a good question is not just one that is found to be useful by other people: a question is good if it is also presented clearly and shows prior research. Using a community question-answering site that allows voting over the questions, we show that there is a notion of question quality that goes beyond mere popularity. We present techniques using latent topic models to automatically predict the quality of questions based on their content. Our best system achieves a prediction accuracy of 72%, beating out strong baselines by a significant amount. We also examine the effect of question quality on the dynamics of user behavior and the longevity of questions.

Discovering Response-Eliciting Factors in Social Question Answering : A Reddit Inspired Study

“…This gives us good ground to understand the issues with unanswered questions. On the contrary, 99.4% of questions studied in (Ravi et al 2014) and 95% of questions in (Shah and Pomerantz 2010) received at least one answer, often involving multiple responders. • The notion of quality is natural and intuitive in our dataset, where a single responder handpicks a few questions that he/she wishes to answer.…”

“…Prior research on Community Question Answering (CQA) has addressed issues as diverse as predicting whether a particular answer will be chosen by the inquirer or not (Adamic et al 2008), predicting answer quality (Shah and Pomerantz 2010;Jeon et al 2006) and quantity (Dror, Maarek, and Szpektor 2013), understanding question type (Allamanis and Sutton 2013) and quality (Ravi et al 2014;Yang et al 2011;Li et al 2012), analyzing inquirer's satisfaction (Liu, Bian, and Agichtein 2008) and responders' motivation, finding similar questions (Li and Manandhar 2011) and expert potential responders (Li and King 2010). Although the aforementioned studies are helpful, we are more curious about the factors of a question that are more likely to generate a response, especially in the MISR setting where there is just one responder but multiple inquirers.…”

“…The task of predicting question quality is studied in (Ravi et al 2014) where latent topics were found to be useful esti- mators for the task. Our work, although similar in spirit, goes one step beyond to identify latent topics that are not only effective but also interpretable so that one can understand the qualities that response-eliciting questions share along with issues that are common across questions that fail to generate a response.…”

“…single user may be responded to by multiple other responders. Prior research has used datasets from these sites to predict whether a question is of high quality (Ravi et al 2014;Li et al 2012); or guess if a given question will be answered (Yang et al 2011;Dror, Maarek, and Szpektor 2013); or estimate the number of answers a question will receive (Dror, Maarek, and Szpektor 2013); or select the best response to a given question (Adamic et al 2008). However, analyzing factors of a question which are likely to elicit a response has been outside the scope of such prior work.…”

Expertise and Dynamics within Crowdsourced Musical Knowledge Curation: A Case Study of the Genius Platform

“…In this paper, we also analyze Genius in the context of other well-studied crowdsourced information sites, such as Stack Overflow (Ravi et al 2014;Posnett et al 2012;Tian, Zhang, and Li 2013), Quora (Wang et al 2013;Maity, Sahni, and Mukherjee 2015), Yahoo! Answers (Adamic et al 2008), Amazon (Bai et al 2018), and Wikipedia (Beschastnikh, Kriplean, and McDonald 2008;Mesgari et al 2015).…”