Abstract

This study investigates the predictors of dropout intention in higher education through a systematic literature review complemented by bibliometric and network analysis. Drawing upon seventy-four peer-reviewed publications from the Scopus database between 2010 and 2025, the study adopts the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) framework to ensure methodological transparency and rigor in selecting relevant sources. Bibliometric techniques—employing VOSviewer software—enabled the visualization of keyword co-occurrence, co-authorship patterns, citation networks, and the geographic distribution of research output. Bibliometric analysis documents a post-2020 boost in dropout intention research, prompted by the COVID-19 pandemic, with a notable thematic shift toward psychosocial and crisis-related themes (e.g., loneliness and financial insecurity) from structural-academic themes (e.g., integration). The field is theoretically scattered, based on foundational models like Tinto’s, but not unified. Authorship is disseminated, international collaboration limited, and geographic focus skewed toward European perspectives, ignoring Global South contexts. Consequently, we propose an integrative conceptual model—grounded in SD-R and social cognitive theories—at micro-individual (motivation, mental health), meso-contextual (instruction support, institutional climate), and macro-structural (policy, economic barriers) levels. It invites mixed-methods, longitudinal, and cross-national investigations to capture temporal and contextual heterogeneity. Practically, it informs strategic interventions that combine academic, financial, and psychological support, particularly to vulnerable groups. Limitations are database (Scopus) and language bias, which underline the need for increased inclusivity in future scholarship to create equitable, evidence-based retention practice globally.