Implication of Wald test anchor-all-test-all procedure for anchor selection

Abstract EN

Identifying differentiated functions across different groups is significant for establishing validity in the psychometric testing.

Differential item functioning (DIF) methods are formulated to address the comparison between different groups.

The study analyzes the alternative strategies for anchor selection that help overcome the limitations of IRT-LR (item response theory-likelihood ratio) all-others-as-anchors (AOAA) procedure.

IRT framework introduced different anchor-selection strategies for anchor testing, which need multiple model fittings.

Alternative anchor-selection strategies are needed to be implemented based on Wald X2 modified version in the IRTPRO and flex MIRT.

The comparisons in this study are based on ratio testing methods.

It showed that the anchor selection through the Wald test performs well in identifying the DIF items across other variables, including different numbers of groups.

The excellent performance of the Wald test detecting DIF is confirmed through reliable and effective anchor-selection methods