Effects of Ability Scale Purification on the Identification of dif

Abstract

Summary: Research related to the detection of item bias or differential item functioning (dif) has proliferated in psychometric and applied psychological literature over the last 25 years. In fact, debate has been heated on the nature and, more particularly, on the methods for bias/dif detection. Today, conditional methods have obtained wide acceptance. However, these methods present a problem of circularity: If the test contains biased items, then a biased measure of the matching variable will be used for investigating dif. Thus, we investigate dif with a biased measure. The aim of this paper is to investigate the feasibility of improving item bias identification through “purification” of the ability measure used. Several bias-detection techniques are analyzed: Mantel-Haenszel statistic, logit model, logistic regression procedure, restricted factor analysis, and two item response theory (IRT)-based indices. Results show that purifying the ability scale improves item bias detection greatly, providing rates of correct identification close to 100% with all these techniques. IRT-based indices showed the greatest improvement.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 60 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.