The internationalization of postgraduate education has heightened issues about the quality, efficiency, and equity of supervision as the key factors affecting success and completion rates of students. This study sought to establish how quality, efficiency, and equity in postgraduate supervision affect student success and supervision outcomes among MPhil Education students at the University of Cape Coast, Ghana. The study was based on the Social Cognitive Theory and examined the direct impact of the characteristics of supervision on student achievement and supervision outcomes, and finally examined the success as an intervening variable (operationalised as the mediating mechanism of research self-efficacy, skill mastery, and scholarly identity formation through which supervisory characteristics are theorised to shape supervision outcomes) in these relationships. Data were drawn from a census of 123 MPhil Education students at a Ghanaian public university, using a cross-sectional survey design analysed through partial least squares structural equation modelling (PLS-SEM). The model explained 67.2% of the variance in success and 51.1% in supervision outcomes, with acceptable predictive relevance on both. Quality was the dominant predictor throughout the model, but its effect ran opposite to theoretical expectation: higher perceived quality was associated with lower success (β = -.564, p < .001) and less favourable supervision outcomes (β = -.603, p < .001). Efficiency similarly reduced success (β = -.254, p = .018) but had no relationship with supervision outcomes, and equity predicted neither. Success did not mediate any of the three supervisory dimensions' relationships with supervision outcomes, so the hypothesised mediation was not supported. The negative quality and efficiency paths are interpreted cautiously: they may reflect genuine expectation-disconfirmation among more discerning students, but discriminant-validity checks also show quality and success are not cleanly separated in this dataset, so a measurement artefact cannot be ruled out. The study recommends against revising supervision policy on the strength of these results alone, and instead calls for routine, validated monitoring of supervision, structured supervisor training in feedback practice, continued attention to equitable access, and qualitative or longitudinal follow-up to clarify the mechanisms involved.