New Research Pinpoints the Ideal Mentoring Dose for Undergraduates
Zhi, R., & Pan, Q. (2026). Too much of a good thing: The impact of mentoring on innovation capacity among undergraduate students—Evidence from a provincial University in China. Acta Psychologica, 268, 107284. https://doi.org/10.1016/j.actpsy.2026.107284
Introduction
As universities push to cultivate innovation capacity earlier in students’ academic careers, faculty mentoring is often treated as a straightforward lever: more contact with a mentor should translate into more research skill and creative output (Zhi & Pan, 2026). But the authors note that this assumption, largely validated in graduate and doctoral populations with established research identities, may not transfer cleanly to undergraduates, who are still forming disciplinary identities and building basic research confidence. Drawing on motivational synergy theory and self-determination theory, the authors ask whether mentoring’s benefits might taper off, or even reverse, once meeting frequency crosses some threshold.
Methods
The researchers surveyed 1,066 undergraduates at a Chinese provincial university with a “full-coverage” mentoring system, in which every student from sophomore year onward is assigned a faculty mentor. Innovation capacity was measured through both self-assessed research literacy and objective output (competition awards, publications, project leadership). Mentoring frequency was operationalized as average group meetings per semester. The team used quadratic regression to test for a curvilinear relationship, then applied bootstrap-based mediation analysis to examine four potential mechanisms: mentor pressure, peer influence, academic identity, and self-efficacy.
Results
The data revealed a clear inverted U-shape: innovation capacity rose with mentoring frequency up to approximately 13.35 meetings per semester, then declined beyond that point (Zhi & Pan, 2026). All four motivational pathways showed the identical rise-then-fall pattern, suggesting mentoring’s psychological character shifts from supportive to controlling as frequency increases. The effect was strongest among students with moderate baseline innovation capacity, in STEM disciplines, and in smaller mentoring cohorts. The average student in this sample received only 4.27 mentoring sessions per semester—meaning under-mentoring, not over-mentoring, is the more pressing concern for most students.
Discussion
With a mean of 4.27 meetings per semester against an estimated inflection point of 13.35, most students in this system are being under-mentored rather than over-mentored, and the practical priority for the majority is increasing contact. The risk of overload becomes relevant only for a minority already receiving intensive supervision. Notably, the two intrinsic pathways (academic identity, self-efficacy) hit their inflection points earlier (around 11–12 meetings) than the two extrinsic pathways (mentor pressure, peer influence, around 13–14 meetings). The authors interpret this ordering as consistent with self-determination theory’s prediction that controlling pressure erodes intrinsic motivation. In other words, a student’s sense of ownership over their research starts diminishing before the external structure itself stops being useful.
Well-aligned mentoring relationships have more room to run before hitting diminishing returns, since a mentor whose expertise closely matches a student’s interests keeps providing genuinely useful information even at high contact frequency, delaying the shift into controlling territory. Conversely, in large cohorts, the authors suspect mentor attention gets diluted and interactions become standardized in ways that resemble surveillance rather than individualized coaching, while in STEM fields the more structured, milestone-driven nature of lab-based research may make it easier for informational input to remain informational longer.
Implications for Mentoring Programs
Programs should set meeting frequency deliberately, tailored to discipline and developmental stage. In practice, this could mean providing denser, more structured contact for students just entering research work, while intentionally tapering that contact as students mature and shifting the relationship toward lighter-touch guidance that supports independent inquiry. The authors suggest pairing extrinsic supports, such as research opportunities, with conditions that preserve a student’s sense of ownership: genuine latitude in choosing research directions, and feedback that affirms competence rather than simply correcting output. Lastly, the authors argue that mentoring resources should be allocated with an eye toward where they’ll do the most good: students with moderate, still-developing innovation capacity. Students with little prior research exposure may need foundational training and structured problem-solving practice before they’re positioned to benefit from intensive one-on-one mentoring, while already-advanced students may be better served by greater autonomy, exposure to frontier problems, and leadership opportunities within research teams. Standardized mentoring distribution, applied identically regardless of a student’s starting point, risks underserving students at both ends of the capacity distribution even as it helps those in the middle.
Read the full article here
