Comparability of student evaluations of teaching across Arabic- and English-medium higher education tracks: measurement invariance evidence from the UAE

Introduction Universities increasingly use student evaluations of teaching across parallel language tracks, but meaningful interpretation requires evidence that the survey functions comparably across groups. Methods This study examined the measurement comparability of the 49-item Course Evaluation Survey (CES) across Arabic-medium and English-medium tracks in a parallel monolingual teacher-preparation program in the United Arab Emirates. The CES was administered to 528 students (229 Arabic-track and 299 English-track). The tracks followed equivalent learning outcomes, content, assignments, assessments, and objectives, while differing in language of instruction and survey administration. Multi-group confirmatory factor analysis tested measurement invariance, and differential item functioning (DIF) analysis examined item-level differences. Results Full metric invariance was not supported (Delta CFI = .019; Delta SRMR = .062). Seven of 49 items showed DIF, mainly involving relevance, flexibility, discussion, teacher-student relationships, and responsiveness to feedback. The DIF impact curves showed greater divergence for the flagged items than for the full item set, indicating localized rather than uniform non-comparability. Discussion The findings do not imply that all cross-track comparisons are invalid. However, observed CES score differences should not be interpreted as direct evidence of differences in teaching quality or faculty performance without additional evidence of measurement equivalence. The results have implications for multilingual quality assurance, program review, benchmarking, and institutional use of course-evaluation data.