Duo Research Archive (University of Oslo) · 2017 · 22 citations · 0 references
EducationEducational AssessmentPsychologyDifferential Item FunctioningItem Response TheoryPsychometricsDifferential Test FunctioningItem AnalysisNonuniform Differential Item FunctioningPsychological TestingSocial ScienceTest TheoryDichotomous Differential Item FunctioningDichotomous Item Differential FunctioningDifferential Item Functioning ScreeningItem-level Differential AnalysisPolytomous Differential Item FunctioningPolytomous Item Differential FunctioningTest Item DesignValidity Evidence Based On Response Processes
To investigate the domain-specificity of research competencies, higher education students from the social sciences were assessed with a standardized test in four disciplines: (a) sociology, (b) political science, (c) educational studies, and (d) psychology. The measure covered declarative and procedural knowledge of research methods, methodology, and procedures. Quantitative and qualitative research traditions were represented equally by test items. The domain-specificity of the measure was examined by detecting and explaining differential item functioning (DIF) between the disciplines. It was hypothesized that due to differences in opportunities to learn (OTL), students from different disciplines responded differently to subgroups of items. As expected based on the OTL-patterns, research traditions significantly explained variance in DIF. While psychology students were more likely to correctly answer items addressing quantitative methods than students with the same overall ability level but from different disciplines, students of all other disciplines were more likely to solve items addressing qualitative methods. These differences coincided with differences in OTL. Overall, the findings suggest that research competencies are similar across the social sciences, but differences between disciplines exist in their focus on quantitative or qualitative methods. (DIPF/Orig.)