The effects of controlling for distributional differences on the Mantel-Haenszel procedure Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 22, 2019
  • Bowen, Daniel F.
    • Affiliation: School of Education
  • Propensity score matching was used to control for the distributional differences in ability between two simulated data sets. One data set represented English speaking examinees and the second data set represented Spanish speaking examinees who received the Spanish language version of the same test. The ability distributions of the two populations were set to mirror the data from a large-scale statewide assessment delivered in both English and Spanish. The simulated data were used as a benchmark to compare the Mantel-Haenszel procedure for identifying differential item function (DIF) under the two scenarios with the hope that controlling for distributional differences would reduce the errors made by the Mantel-Haenszel procedure in identifying DIF. With the increase in high-stakes testing of diverse populations in the United States as well as elsewhere in the world, it is critical that effective methods of identifying DIF be developed for situations where the reference and focal groups display distributional differences. However, the evidence of this study suggests that the matched samples did not help the Mantel-Haenszel procedure identify DIF, and in fact increased its errors.
Date of publication
Resource type
Rights statement
  • In Copyright
  • "... in partial fulfillment of the requirements for the degree of Master of Arts in the School of Education."
  • Cizek, Gregory J.
Degree granting institution
  • University of North Carolina at Chapel Hill
Place of publication
  • Chapel Hill, NC
  • Open access

This work has no parents.