Evaluating Model Replication Efforts Through Prior Predictive Similarity Checking

• In Event: Evaluating Fit

Abstract

In model-based research, goodness-of-fit (GOF) of the original model (e.g., CFA) to the replication data is routinely provided as evidence of replication. We demonstrate that GOF obscures important differences between the original and replication studies. We present Bayesian prior predictive similarity checking: a tool for rigorously evaluating model-based replication.

Authors

• Sonja D Winter, UNIVERSITY OF MISSOURI, COLUMB

• Wes Bonifay, University of Missouri

• Hanamori Skoblow, University of Missouri

• Ashley Watts, Vanderbilt University