Conditional Process Analysis: Moderated Mediation Model of Perceived Ethnic Discrimination and Binge Drinking Among Recent Latino Immigrant Youth

• Activity Number: 88 - SPEED: Causal Inference and Related Methodology Part 2
• Type: Contributed
• Date/Time: Sunday, July 28, 2019 : 5:05 PM to 5:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Zoran Bursac* and Miguel Angel Cano and Seth J Schwartz
• Companies: Florida International University and Florida International University and University of Miami
• Keywords: Mediation; Moderation; Conditional process analysis

Abstract

A model that includes both a mediation and moderation component is known as a conditional process model, in which the direct and/or indirect effect of independent variable X (e.g., perceived ethnic discrimination) on dependent variable Y (e.g., binge drinking) though mediator M (e.g., self-esteem) is moderated by or conditioned on one or more variables, for example W or Z (e.g., gender). A conditional process model can include moderation of more than one path in the causal sequence. The effect of X on M and the effect of M on Y can be moderated by a common variable W. We extend this scenario to multiple mediators Mi, moderators Wi, and combination of two. We present conceptual models, statistical models, and apply it to the motivating data set via regression-based approach and SAS PROCESS macro.