Investigating College Student Mental Health Needs and Barriers Using Natural Language Processing

Abstract

College students’s mental health has been declining, despite efforts by universities to improve outreach and programming. Though a variety of resources are available, students, especially those who experience marginalization, often do not engage with them. LGBTQ+ students in Florida, in particular, may face unique challenges. To understand marginalized college students’s mental health experiences, we conducted ten focus group interviews (N = 65) with students with various marginalized identities. Using Natural Language Processing (NLP) techniques, we conducted sentiment analysis with respect to students’s responses on interview questions (e.g., mental health needs, how needs are met/not met) among LGBTQ+ and other minority groups. Overall, LGBTQ+ students yielded a more negative sentiment than the other groups, with a marginally significant (at 0.1 level) difference suggested by Wilcoxon rank sum test. This project presents a novel approach to leveraging traditional qualitative methods and AI, provides deeper understanding of the mental-health-related needs of marginalized college students, and practical implications for prevention and intervention efforts to promote college students’s mental health.