Weiqing Zhang will be the Jones Graduate School of Business at Rice University as an Assistant Professor in Marketing. His research interests focus on media content creation and consumption, small businesses’ welfare on e-commerce platforms, and dynamics of customer product returns.
Research Summary: We investigate the value of content inclusiveness, wherein content that deviates from a social media platform’s preferred stances is retained, from the perspectives of users, content creators, and the platform. Specifically, we study (1) user preferences on content deviation, which refers to the semantic gap of a publication and its host platform’s preferred stances, (2) content creators’ creation decisions on content deviation, and (3) the effect of the strictness of the platform enforcement on its stances.
In this paper, we use various NLP methods including topic modeling and transformer models to identify the platform’s preferred stances and construct content deviation of publications from content creators. Subsequently, we build an empirical framework that combines both user preference on media consumption and content creators’ content deviation decisions. On the demand side, we model how users make consumption decision based on the level of content deviation of a publication and of overall deviation in the environment. On the supply side, we model how content creators determine the level of content deviation, given demand and the strictness of platform enforcement on its stances, as an equilibrium outcome and provide the sufficient conditions for the existence and uniqueness of the equilibrium.
We estimate the proposed model using data on news creation and consumption onWeChat. Our findings indicate that users prefer publications with high levels of content deviation but reduce the consumption in a deviation overload environment. We also use the empirical framework to derive structural estimates that reflect the strictness of platform enforcement on its stance (SPEIS) over time. These estimates enable us to quantify its impact on platform-level content consumption. Motivated by real-world practices, we perform various counterfactual experiments, including adjusting the levels of SPEIS, suspending highrisk accounts, and demonetizing certain controversial topics. Through these experiments, we find that the interplay of the effects regarding user preferences on content deviation and environment deviation determines the market outcomes.
