dif_irt: Item response theory DIF method

In knickodem/WBdif: Investigate DIF and generate reports

Item response theory DIF method

Description

Conducts DIF analysis in an item response theory framework using the iterative Wald procedure (Cao et al., 2017; Tay et al., 2015; Woods et al., 2013)

Usage

dif_irt(item.data, dif.groups, item.type, method, Wald)

Arguments

item.data	data frame of item responses with subjects in rows and items in columns
dif.groups	factor vector of group membership for which DIF is evaluated.
item.type	For IRT, the type of model to fit for each item. The default is "2PL" for dichotomous items and "graded" for polytomous items. See mirt for more options and details.

Details

The process begins by conducting an omnibus test of DIF by comparing the fit of no DIF, uniform DIF, and non-uniform DIF models. All models are run with multipleGroup. Specifically, in step 1, the no DIF model is run with parameters estimated freely between the dif.groups. The latent trait mean and variance for the focal group are saved. In step 2, the uniform DIF model is estimated while constraining the latent trait parameters for the focal group to those in step 1 and constraining the item slopes to be equal across dif.groups. The non-uniform DIF model is then estimated by adding equality constraints to the item intercepts or thresholds. Model fit is compared between the No DIF and uniform DIF, and between the uniform DIF and non-uniform DIF models with a likelihood ratio test. If no DIF is detected through the model comparisons, the process concludes. Otherwise, the specific item(s) with DIF are identified in steps 3 and 4 using DIF. This two-step process is sometimes referred to as the initial and refinement steps for identifying anchor items and items with DIF.

Value

A list containing

• DIF model comparisons

• item-level DIF tests

• integer vector of the items showing DIF (i.e., biased items)

• type of DIF

• IRT models needed for treatment effect robustness check

References

Cao, M., Tay, L., & Liu, Y. (2017). A Monte Carlo study of an iterative Wald test procedure for DIF analysis. *Educational and Psychological Measurement, 77*(1), 104-118. doi:10.1177/0013164416637104 Tay, L., Meade, A. W., & Cao, M. (2015). An overview and practical guide to IRT measurement equivalence analysis. *Organizational Research Methods, 18*(1), 3-46. doi:10.1177/1094428114553062 Woods, C. M., Cai, L., & Wang, M. (2013). The Langer-improved Wald test for DIF testing with multiple groups: Evaluation and comparison to two-group IRT. *Educational and Psychological Measurement, 73*(3), 532-547. doi:10.1177/0013164412464875

knickodem/WBdif documentation built on Feb. 3, 2024, 2:20 a.m.