A Fit Propensity Effect Size Approach to Improve Model Selection

• In Event: Electronic Board Session II

Abstract

Despite their limitations, latent variable model selection relies on fit indices. Fit propensity (FP) is advanced to overcome limitations, where good model fit and low FP compared to alternative models are preferred. An effect size approach to characterizing FP is investigated using a simulation study. Results demonstrate improved model selection.

Authors

• Holmes Finch, BALL STATE UNIVERSITY

• Brian French, WASHINGTON STATE UNIVERSITY

• Jason Immekus, UNIVERSITY OF LOUISVILLE