Suggested guidelines in reporting results from mediation analysis, standardized or unstandardized?

Abstract

Mediation analysis is commonly implemented in psychological, epidemiological, and social behavior studies to identify potential factors that mediate associations between exposures and physical or psychological outcomes. Various analytical tools are available to perform mediation analyses, among which Mplus is widely used due to its user-friendly interface. In practice, sumptuous results provided by Mplus, such as the estimated standardized and unstandardized effect sizes, can be difficult for researchers to choose to match their studies. Through a comprehensive review and utilizing findings from a proven study, we proposed guidelines and recommendations to help users select between standardized or unstandardized results based on data attributes and users' hypotheses. We also provided guidelines to choose from several types of standardized values based on the types of variables, including exposures, mediators, and outcomes.