A stepwise test characteristic curve method to detect item parameter drift

An important assumption of item response theory (IRT) based equating is that the item parameters should be invariant over different testing occasions. Sometimes, however, item parameters do not remain invariant due to factors other than sampling error, and this is termed item parameter drift (IPD). Several methods have been proposed to detect drifted items. However, most of the existing methods aim at detecting the drift in individual items, which may not be ideal when only the overall test characteristic curve (TCC) is of interest to the users. One such occasion in common practice is IRT-based true score equating, where the goal is to create a conversion table to make the two TCCs as close as possible. This paper introduces a stepwise test characteristic curve (Stepwise TCC) method to dynamically detect item parameter drift based on TCC without requirement to set any critical values. Comparisons were made between the new method and two commonly used existing methods under the three-parameter logistic model. Results show that the new method performed well in IPD detection.

Name thesis:degree_name

M.A.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Psychology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2014

Author dc:creator

• Guo, Rui

Contributors dc:contributor

• Chang, Hua-Hua

• item parameter drift
• stepwise selection
• test characteristic curve
• item response theory
• true score equating

Statement dc:rights

• Copyright 2014 Rui Guo

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/49806

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/49806