IRT Models: Comparison of 1, 2, and 3 Parameter Models

This module provides an overview of Item Response Theory (IRT) models, specifically comparing the 1-, 2-, and 3-parameter logistic models. It details the mathematical formulas for each model and illustrates the impact of varying item parameters like difficulty, discrimination, and the pseudo-guessing parameter. A key focus is on how these parameters influence item characteristic curves and how they relate to examinee ability. Furthermore, the text addresses parameter estimation procedures using programs like LOGIST and BILOG and discusses methods for evaluating how well these models fit observed data. The module also acknowledges the assumptions underlying IRT models, such as unidimensionality, and provides annotated references for further study. Finally, teaching aids designed to complement this module, including data sets and parameter values, are available from the NCME.

keepSave to notecopy_alldocsAdd noteaudio_magic_eraserAudio OverviewschoolBriefing doc