APS
2026 APS Annual Convention
Rethinking Mediation: Evidence, Transparency, and Tools for Better Causal Inference
Mediation analysis remains central yet contested in psychology. This symposium reviews recent practices, evidential challenges, and risks of analytical flexibility, then introduces novel statistical methods addressing threats to causal inference. Talks span scoping reviews, replication concerns, mixed-model decompositions, and diagnostics for omitted moderators, highlighting implications for robust psychological theory testing.
Chairs & Discussants
- Hilary BergsiekerChair
University of Waterloo - Vincent YzerbytCoChair
Universite Catholique de Louvain - Hilary BergsiekerDiscussant
University of Waterloo
Presentations
- Prevalence, Transparency, and Specification of Mediation Models: A Scoping Review across Psychology SubfieldsConnery Knox, Shawn Yee, Bhumika Bhandari, Timothy Ireland, Agnes Zientarska-Kayko, Hilary Bergsieker
- Improving Quality of Evidence In Mediation Analysis: From P-Hacking to PreregistrationAmanda Montoya, Aaron Charlton, Antonia Krefeld-Schwalb, John Price, Joseph Hilgard
- Mediation Models for Experimental Designs with Crossed Random FactorsJoshua Correll, Vincent Yzerbyt, Charles Judd
- Illusory Indirect Effects: Diagnosing Causal Heterogeneity and (Re)Discovering Omitted ModeratorsShawn Yee, Connery Knox, Vincent Yzerbyt, Hilary Bergsieker
