Consequences of Ignoring Guessing Effects on Multiple-Group Factor Analysis

• In Event: Hilton Roundtable Session 4
  In Roundtable Session: Division D Section 1 Roundtable Topic: Applications of Item Response Theory

Abstract

This study investigated the consequences of ignoring guessing effects on both model-data fit evaluation and parameter estimation when conducting measurement invariance analysis using multiple-group factor analysis. A simulation study was conducted using 3-parameters logistic item response theory model for data generation. The manipulated factors included distribution of the abilities, the pseudo-guessing parameter values, and sample size. The results showed that when the guessing effect was present, the parameter estimates were biased. However, the fit indices in general indicated a good model-data fit across all three levels of measurement invariance analysis. The results suggest that the model-fit indexes are not useful for detecting model misspecification with respect to the ignorance of guessing parameters in testing measurement invariance.

Authors

• Ismail Cukadar, Florida State University

• Yanyun Yang, Florida State University

• Insu Paek, Florida State University