Empirical Evaluation for Gender-Based DIF in an International Assessment: The Case of PIRLS

Abstract

This study investigated gender differential item functioning (DIF) in the Progress in International Reading Literacy Study (PIRLS) in the 2016 cycle. This is because there is no empirical review of gender DIF in this test. The DIF analyses were conducted within and across three countries which speak a common language. The study employed logistic binomial regression in order to detect uniform and nonuniform DIF items within and across the three randomly selected countries. The findings clearly indicated the presence of uniform and nonuniform DIF within and across the analyzed countries. The magnitude ranged from small to large effect size. This study endeavored for alerting psychometrists, professionals, and all other stakeholders about the prescience of gender-based DIF.