Comparison of R Packages for a Mixture Rasch Model (Poster 16)

• In Event: AERA Poster Session Two
  In Poster Session: Measurement Theory and Application 1

Abstract

The mixture Rasch model (MRM) can be a useful model for handling population heterogeneity in real data that violate the unidimensionality assumption in IRT. Different R packages have been used to estimate the MRM. The purpose of this study was to compare the results of two different R packages (mixRasch and TAM) for estimating the MRM on eTIMSS 2019 data. Detection of differential item functioning (DIF) between latent classes was used to compare item parameter estimates between the two packages. A 2-class model was selected by the mixRasch and a 3-class model was selected by the TAM based on the BIC index. Results clearly demonstrated that the two R packages provided different MRM solutions.

Authors

• Boram Lee, Ewha Woman’s University
  Non-Presenting Author

• Sunyoung Ahn, Jeongwha Arts College
  Presenting Author

• Youn-Jeng Choi, Ewha Womans University
  Non-Presenting Author

• Allan S. Cohen, University of Georgia
  Non-Presenting Author