Fostering Sustainable Graduate Student Well-Being: Lessons and Approaches from the Well-Being Advocate Program:
The prevalence of mental health challenges among graduate students can be influenced by the distinct academic and professional environments in which they study and grow. These environments are further shaped by the unique cultural norms and values of individual graduate programs, creating complex and varied well-being climates. Declines in well-being can impact critical aspects of the student experience, including formation of a disciplinary identity, sense of belonging, and decisions about continuing in academia.
The University of Michigan Rackham Graduate School’s Well-Being Advocate Program is a cornerstone initiative advancing the principles of the Okanagan Charter within graduate education by creating sustainable well-being environments through systemic change. This innovative program integrates well-being into graduate program systems and policies by partnering with and empowering graduate programs and program leadership to help identify and implement tailored interventions that foster well-being cultures.
Exploring Graduate Student Well-Being: Insights from the Well-Being Advocate Program Survey Data:
Graduate student mental health and well-being have emerged as critical concerns within higher education, with research indicating increased rates of anxiety, depression, and burnout across disciplines. While numerous studies have examined well-being in specific graduate programs, there is a limited methodological use of synthesizing data collected through different survey instruments across varied academic contexts. Additionally, the application of advanced methodologies to analyze data is little to none.
Our project plans to analyze de-identified survey data from five to six graduate programs at the University of Michigan that were collected over the past year by the Well-Being Advocate Program. By utilizing advanced technology, we plan to process qualitative and quantitative responses from non-identical surveys that share the common goal of assessing student well-being. This methodological approach enables us to identify patterns and relationships between student experiences and various dimensions of well-being that transcend individual program boundaries.
Specifically, we utilize large language models (LLMs) to categorize and analyze free-text responses, supplemented by word embedding techniques that reveal semantic relationships within student narratives. This approach allows us to process data at scale while preserving the nuanced perspectives of individual students across diverse graduate environments. Additionally, the utilization of local LLM’s can ensure secure analysis of any potential confidential data.
We expect our analysis to reveal connections between academic structures and student well-being outcomes that may remain invisible when examining programs in isolation or examination of text in a traditional method. We argue that this technology-enhanced methodology not only provides actionable insights for improving graduate education at Michigan but also demonstrates a transferable approach for institutions facing similar challenges in understanding their graduate student populations holistically.
By showcasing how advanced AI technologies can advance well-being research, this work contributes to both the understanding of graduate student experiences and the methodological advancement of well-being assessment in higher education contexts.
