Assessing the Scope and Alignment of Undergraduate Research Training: A Computational Analysis of Thesis Topics in a Medical University Context.

Abstract

Undergraduate research training is a fundamental component of health professions education, intended to develop critical thinking and scientific inquiry. However, systematic evidence regarding the scope and alignment of student research topics with disciplinary expectations and educational program objectives remains limited. This study examined patterns in undergraduate thesis topics within a Chinese medical university context, using a computational framework to assess their alignment with institutional positioning, established disciplinary research, and peer-recognized undergraduate achievement. A total of 392 undergraduate theses from an applied psychology program (2022-2025) at a Chinese medical university were analyzed using a combination of expert evaluation and BERTopic modeling. To benchmark the currency and diversity of these topics, student work was mapped against a joint model of 2,790 peer-reviewed psychology journal articles and award-winning theses from Chinese peer institutions. The findings indicate a pronounced misalignment between the clinical resources available at the institution and the research practices reflected in student theses; despite the university's positioning as an academic medical center, fewer than 3% of student theses recruited patient populations, while 81% relied on convenience sampling of fellow students via questionnaires. Topic modeling further showed substantial concentration, with 69% of theses clustering into just four narrow domains that occupy only 30% of the available disciplinary topic space. This pattern reflects limited engagement with methodologically demanding, experimentally oriented topics that characterize much of the published disciplinary research. This concentration was associated with organizational dynamics; specifically, a 51% decline in faculty availability and the departure of medical specialists from the institution coincided with students clustering into low-resource, supervisor-centric research domains. These findings illustrate how institutional positioning and faculty constraints can create self-reinforcing bottlenecks that limit student exposure to clinical environments, even in the presence of formal goals for clinical research productivity. This study offers a scalable computational framework for examining research training and highlights the need for programs to structurally align their clinical assets with undergraduate curricular pathways.