Procedure to Perform Item Response Analysis by Hierarchical Generalized Linear Model

Recent research demonstrated that multilevel modeling could be a useful approach to conduct item response data analysis. Although the literature provides extensive discussions on modeling, it may not provide sufficient information on how to conduct such a data analysis for practitioners. This paper presents detailed descriptions of practical procedures to conduct item response data analysis using HLM software. Two-level Rasch equivalent model, two-level model with a person-characteristic variable, three-level Rasch equivalent model, three-level model with a person-characteristic variable, and three-level model with person- and group- characteristics variables are presented with step-by-step descriptions of data analysis, including preparation of data files and model specifications for HLM software. A part of Third International Math and Science Study (TIMSS) data are used for the illustrative purpose