“During graduate school, it was easy to feel lost in the process. I often felt uncertain about where I stood or what was expected of me.” Dr. Chris Vaaga said that not as a cautionary tale but as a memory that actively shapes how he mentors his Biomedical Sciences students at Colorado State University today. His experience is far from unique, and the consequences are serious.
Up to 50 percent of graduate students report symptoms of depression, anxiety or burnout during their training, which is more than six times the rate of the general population. Attrition across PhD programs runs between 36 and 51 percent, with nearly half of doctoral students who reach the dissertation phase never finishing. Behind every one of those numbers is a person who arrived with real potential and left without the support they needed.
The conventional institutional response treats mentorship as a faculty disposition rather than an institutional responsibility—something good professors do naturally and something administrators quietly hope for. However, that model fails students who are first generation, international, neurodivergent or from nontraditional backgrounds. It also fails faculty like Dr. Vaaga, who want to do right by their students but have never been given the tools to do so consistently. At CSU, we decided to build something different.
Infrastructure, Not Just Programs
The instinct in higher education is to respond to problems with programs (e.g., a workshop here, a resource guide there). What we built at CSU through the CSU Mentor Well initiative is better described as infrastructure: a connected system of training, community, peer support and governance that holds itself together and builds over time.
The foundation is the Mentor Well faculty development program, grounded in evidence-based practices from the Center for the Improvement of Mentored Experiences in Research (CIMER). However, we did not simply adopt the CIMER curriculum and call it done. We expanded it with modules on student well-being, belonging and support outside STEM fields because graduate education’s mentoring crisis is not a STEM problem but a graduate education problem.
What followed was a recognition that one-time workshops do not change culture, so we built ongoing touchpoints: Mentor Mondays, which create regular community spaces for faculty to share what they are encountering and how they are responding; Mentor Tips, which deliver brief practical resources on a rolling basis; and traveling presentations that bring targeted support directly to departments facing specific challenges, from supporting students through mental health crises to delivering effective micro-affirmations.
On the student side, the Graduate Peer Mentoring Program (GPMP) connected 65 mentoring pairs in its first year. What stood out in that data was not only the academic support those pairs provided but the breadth of what students were actually seeking help with. Personal challenges, mental health, social isolation and professional navigation all appeared alongside coursework. Graduate students need mentors who are prepared for that full range, and GPMP also gave graduate students the chance to develop their own mentoring skills.
When Mentoring Becomes Policy
The piece that makes CSU’s model distinctively durable is the Graduate Center for Inclusive Mentoring (GCIM): an interdisciplinary network of approximately 230 faculty, representing nearly 10 percent of CSU’s faculty, who have completed Mentor Well. GCIM is not a committee but is a faculty governance group, and that distinction matters enormously. That number continues to grow, and with it the prospect of reaching the kind of institutional critical mass where mentoring culture shifts on its own.
Dr. Michael Pante, Chair of Anthropology and Geography, describes graduate mentorship as “the hardest part of the job.” What GCIM gave faculty like him was a seat at the table where the rules get made. Members worked through shared governance to revise faculty evaluations and promotions, embedding mentorship quality into these standards. Department head guidelines now explicitly address a faculty member’s ability to foster professional growth, provide timely support and attend to students’ psychosocial needs.
Dr. Tatiana Nekrasova-Beker, an associate professor in the English department, puts it plainly: “The mentorship needs of MA students, international students and nontraditional students differ greatly from traditional STEM PhD students. Having mentorship recognized formally in policy allows us to adapt our approaches flexibly and thoughtfully.” When mentoring is left to individual discretion, it gets applied unevenly. When it is embedded in policy, it becomes a shared expectation. As Dr. Pante describes it: “Before this structural change, mentoring was something done quietly and without much institutional acknowledgment. Now, it’s visible, rewarded and systematically supported. That makes a profound difference.”
What Other Institutions Can Take From This
CSU’s experience does not offer a plug-and-play template, but several principles translate directly:
- Build for the whole student, not just the academic one. GPMP’s first-year data made plain that graduate students seek support across personal, mental health, social and professional domains. Institutions that train mentors only for academic advising are preparing them for a fraction of what students need.
- Make expectations explicit on both sides. Faculty consistently told us that Mentor Well helped them differentiate advising from mentoring and articulate that distinction to their students. Structured programs create the conditions for those conversations. Without them, students are left guessing.
- Put mentoring advocates in governance. Training faculty is necessary but not sufficient. The mechanism that made change stick at CSU was giving trained mentors a role in the decisions that govern faculty evaluations and departments expectations. Without it, mentoring improvement stays an individual choice rather than a shared norm. At CSU, the goal is to reach 20 percent faculty participation—the threshold organizational research associates with a cultural tipping point, where norms begin to shift on their own rather than requiring constant institutional reinforcement.
- Invest in ongoing community. The faculty most engaged with mentoring improvement at CSU are those who have a continuing community around it. Mentor Mondays are not a luxury because they are what keeps the learning alive and the community in contact after initial training ends.
The Honest Case for Acting Now
Institutions that wait for the perfect program design before acting are making a choice, and it has real costs. Every year of inaction is another cohort of graduate students navigating their programs without the support they need and another cohort of faculty left to figure it out alone. Dr. Vaaga still remembers what it felt like to be lost. What has changed for his students is that he now has the tools, the community and the institutional support to make sure they do not have to feel that way. That is what this kind of infrastructure makes possible: not perfection, but a system that actually tries.
Graduate education has long treated mentoring as something that happens in the spaces between everything else. CSU’s experience suggests it can be something more deliberate than that. The question is whether other institutions are willing to build for it.