Comparing Maximum Likelihood Estimation Methods in the Rasch Model With Sample Sizes and Test Lengths

• In Event: Rasch Modeling: Methodology and Application

Abstract

This study investigates the effects of sample sizes, test lengths and different distributions of item difficulty in the sense of the robustness of item parameter recovery for joint, conditional and marginal maximum likelihood estimation methods for the Rasch model with mean absolute bias (MAB) and root mean squared error (RMSE) with a simulation study.

Author

• Jiaqi Zhang, SKT Education Group