Gender-related differential item functioning and academic predictors of study completion expectations in sports science students

Gender-related differential item functioning and academic predictors of study completion expectations among sports science students remain underexplored despite their theoretical relevance within the Social Cognitive Career Theory (SCCT) framework. This study developed and validated the Study Completion Expectation (SCE) instrument, initially comprising 10 items across two subfactors, among 458 Indonesian sports science students (304 male, 154 female; 116 athletes, 342 non-athletes) through EFA, CFA, and Rasch analysis. Exploratory factor analysis reduced the instrument to a 9-item, two-factor structure with strong psychometric properties (Robust CFI = 0.974; α = 0.911–0.922; WLE reliability = 0.774–0.856 for the AOE and PSB subscales and 0.872 for the composite SCE score). A preliminary multi-group CFA, conducted on subsamples below recommended size thresholds, did not indicate non-invariance across gender or athlete status, though this exploratory finding requires replication with larger samples. Rasch theta-based regression confirmed GPA as a significant predictor across all three models (p < 0.05; range p = 0.002–0.022), consistent with SCCT predictions regarding performance accomplishments. Thesis readiness showed a positive trend toward perceived barriers but did not reach significance (B = 0.548, p = 0.058). The most notable finding was that female students consistently demonstrated higher outcome-expectation beliefs on three Academic Outcome Expectation items than male students, indicating gender-related differential item functioning (DIF) detectable only through Rasch item-level analysis and invisible at the construct level; a smaller DIF effect was similarly observed on one Perceived Surmountable Barriers item for athlete status. These findings indicate that comparing AOE scores across gender risks obscuring meaningful differences and warrants the development of gender-specific norms.