Abstract: 【Purpose/significance】Social Q&A communities play a key role in health information comunication, understanding user preferences of message type is crucial to improve information quality and facilitate effective communication with users.【Method/pro⁃cess】This study uses the multi-value set Qualitative Comparative Analysis (mvQCA) method to analyze relevant COVID-19 answers in Zhihu.We choose eight antecedent variables from the perspectives of message features and information sources.Then mine multiple concurrent causal relationships between variables to obtain the conditional configuration and characteristics about different user prefer⁃ences of answers.【Result/conclusion】The research results showed that:(1) There are three configurations for high preference of an⁃swers,among which more number of words and visualization are the necessary conditions.There are 4 configurations for low preference
of answers,among which less number of words and not containing experts are the necessary conditions.and other factors are different.(2) The topics of high preference of answers answers focus on several aspects:the current situation of each country about COVID-19.prevention and treatment ways of COVID-19,relevant information about doctor-patient special people【. Innovation/limitation】Differ⁃ent from previous research perspectives.the perspective of configuration can more effectively mine multiple concurrent causal relation⁃ships.and then get different answer types of user preferences.However,this study only conducts analysis on a small sample,and further studies can be considered on more samples.