Open Collections

The concept of drift and operationalization of its detection in simulated data

University of British Columbia

In this paper, the phenomenon of changes in item characteristics over time (often referred to as drift) is discussed from several theoretical perspectives, and a new procedure for the detection of Item Parameter Drift (IPD) is proposed. An initial evaluation of the utility of the proposed procedure is conducted using simulated data modeled by the 2-Parameter Logistic (2PL) Item Response Theory (IRT) model. In addition to the proposed procedure, an IPD analysis of the simulated data is conducted using two known methods: Kim, Cohen, and Park's (1995) extension of Lord's (1980) Chi-square test of Differential Item Functioning (DIF) to multiple groups, and logistic regression. The results indicate high agreement and accuracy in the detection of true IPD using the two known methods, but poor performance of the proposed procedure. Possible explanations of the findings and future directions are discussed.

Text

2017-09-29

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/63135

Measurement, Evaluation and Research Methodology

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/