Mediation Analysis and Warranted Inferences in Media and Communication Research: Examining Research Design in Communication Journals From 1996 to 2017

Abstract

The number of studies employing mediation analysis has increased exponentially in the past two decades. Focusing on research design, this study examines 387 articles in the Journal of Communication, Human Communication Research, Communication Research, Journalism & Mass Communication Quarterly, and Media Psychology between 1996 and 2017. Findings show that while most studies report statistically significant indirect effects, they are inadequate to make causal inferences. Authors also often infer that they uncovered the “true” mediator(s) while alternative models and mediators are rarely acknowledged. Future studies should pay more attention to the role of research design and its implications for making causal inferences.