Mentorship and money: what really gets Ghanaian graduate students to the finish line

A large survey of graduate students, supervisors, and administrators across five of Ghana’s leading public universities has found that the strongest predictors of dissertation success are not the intellectual brilliance of the students but the practical scaffolding around them: access to research grants, adequate research facilities, clear supervision guidelines, and, above all, a form of mentorship that combines emotional support, ethical guidance, and deliberate fostering of independence. The study, published in SN Social Sciences, offers one of the most detailed statistical portraits to date of how structural and interpersonal factors shape research productivity and assessment competency in a West African graduate education system that has long struggled with delayed completions.

The research team, led by Simon Ntumi of the University of Education, Winneba, together with Samuel Asare Amoah, Selete Kwasikuma, and Daniel Amissah, conducted a descriptive cross-sectional survey involving 550 respondents: 450 graduate students, 80 supervisors, and 20 administrators. The institutions included the University of Ghana, the University of Cape Coast, the Kwame Nkrumah University of Science and Technology, the University for Development Studies, and the University of Education, Winneba. By sampling three distinct stakeholder groups rather than students alone, the design allowed the researchers to compare perceptions across the entire ecosystem that surrounds a dissertation, from the person writing it to the people funding, supervising, and administering it.

The statistical core of the study was multiple linear regression, a technique that estimates how much each factor contributes to an outcome while holding the others constant. When the outcome was dissertation completion, administrative practices emerged as significant predictors. Access to research grants carried a standardized coefficient of 0.24, and the availability of research facilities a coefficient of 0.22, each explaining roughly three percent of the variance in completion outcomes as measured by the eta-squared effect size of about 0.030. Supervision guidelines, meaning formal written expectations for the supervisor-student relationship, matched grants at 0.24. These numbers may look modest individually, but in educational research, where dozens of variables compete to explain a messy human outcome, consistent small-to-moderate effects across related measures paint a coherent picture: money, infrastructure, and procedural clarity matter.

Quality assurance mechanisms formed a second cluster of significant predictors. Internal evaluation systems, in which departments periodically review student progress, showed a coefficient of 0.24, and active monitoring of dissertation progress showed 0.22, with effect sizes in the f-squared range of 0.030 to 0.040. In practical terms, students who worked in departments that tracked their milestones and evaluated them internally were more likely to finish on time than students left to drift. The finding aligns with a long international literature on doctoral attrition, including Vincent Tinto’s influential interactionalist theory of student departure, which holds that persistence depends on academic and social integration into the institution. The Ghanaian data suggest that integration is not only a matter of student psychology but of institutional machinery.

Perhaps the most theoretically interesting result concerns student satisfaction, which the analysis identified as a mediator between administrative practices and dissertation outcomes. A mediation analysis decomposes an effect into a direct path and an indirect path that runs through an intermediate variable. Here, administrative practices had a direct effect of 0.19 on dissertation outcomes, but the indirect effect running through student satisfaction was larger, at 0.22, with statistical significance at p below 0.001. In plain language, good administration does not merely remove logistical obstacles; it makes students feel supported, and that feeling itself propels them toward completion. The subjective experience of being well served by one’s institution is not a luxury but a mechanism.

When the researchers turned from completion to assessment competency, which they define as students’ preparedness to design, evaluate, and defend assessments of academic work, a different set of predictors dominated. Developmental mentorship, a model in which mentors invest in the whole person rather than only the research project, produced the largest effects in the entire study. Emotional support correlated with assessment competency at r = 0.60, ethical guidance at r = 0.63, and fostering independence at r = 0.62. Because the squared correlations all exceed 0.36, these relationships qualify as large effects by conventional benchmarks, meaning that more than a third of the variation in assessment competency tracked these mentoring qualities. Collaborative practices added further weight: participation in team research correlated at 0.55 and interdisciplinary interactions at 0.54.

Academic integration factors also predicted assessment preparedness in regression models. A student’s sense of belonging within their program carried a coefficient of 0.34, the strongest regression weight reported for this outcome, and participation in research seminars carried 0.30. The pattern echoes findings from doctoral education research worldwide, including studies of scholarly communities as learning environments, and it carries a practical implication: departments that run active seminar cultures and cultivate belonging are, in effect, delivering assessment training without calling it that. Students who present work, receive critique, and watch others defend ideas are absorbing the standards by which academic quality is judged.

The study also documented a competency gap that will discomfort university leadership. Assessment competency was significantly lower among graduate students than among supervisors and administrators, with a group difference corresponding to an eta-squared of 0.041. Correlational analysis reinforced the hierarchy of influences: the strongest associations with assessment competency were mentorship at r = 0.63, student engagement at r = 0.62, and administrative support at r = 0.51. Taken together, the numbers suggest that assessment skill is not transmitted through coursework alone but through relationships and institutional climate, and that students currently receive too little of both relative to the staff who evaluate them.

The authors frame their findings as data-driven guidance for reforming graduate education policy in Ghana, where concerns about delayed dissertation completion and inadequate assessment preparedness have grown acute as enrollment expands. The recommendations implied by the data are concrete. Universities should expand access to research grants and upgrade laboratory and library facilities, since these structural inputs carry independent statistical weight. Departments should formalize supervision guidelines so that both parties know what to expect. Quality assurance units should monitor progress at regular intervals rather than waiting for crises. And faculty development programs should train supervisors in developmental mentorship, emphasizing emotional support, ethical modeling, and the deliberate cultivation of student independence, because these interpersonal qualities showed the largest effects anywhere in the study.

The research carries caveats typical of cross-sectional survey designs. Because all data were collected at a single point in time, the correlations cannot definitively establish causation; it is possible, for example, that students who are already thriving seek out more mentorship rather than the reverse. The effect sizes, while consistent, are individually modest, reflecting the reality that dissertation outcomes are shaped by many forces no single survey can capture. The datasets are not publicly available due to confidentiality and ethical constraints, though anonymized data may be shared on reasonable request subject to approval by the University of Education, Winneba’s Ethics Committee. Ethical approval for the study was granted by the university’s Institutional Review Board under protocol UEW-EDU-2025-001, and all participants gave informed consent. Even with these limitations, the study stands as a rare quantitative map of the graduate education pipeline in sub-Saharan Africa, a region where higher education reform has been identified by the African Development Bank as central to economic development. Its central message is quietly radical: the surest way to speed dissertations and sharpen assessment skills is not to demand more of students, but to build institutions that fund, house, guide, and genuinely mentor them.

Subject of Research: Structural and interpersonal factors influencing dissertation completion and assessment competency in Ghanaian graduate programs

Article Title: Shaping dissertation success and assessment competency in Ghanaian graduate programs

Article References: Ntumi, S., Amoah, S. A., Kwasikuma, S., & Amissah, D. (2026). Shaping dissertation success and assessment competency in Ghanaian graduate programs. SN Social Sciences, 6(10), Article 457. https://doi.org/10.1007/s43545-026-01756-3

Image Credits: AI Generated

DOI: 10.1007/s43545-026-01756-3

Keywords: graduate education, dissertation completion, mentorship, assessment competency, Ghana, higher education policy, academic integration, student satisfaction, supervision, quality assurance, cross-sectional survey, regression analysis