Disentangling doctoral well-being support in progress-focused workshops: Combining qualitative and quantitative data in single-case learning analytics

Doctoral education (DE) suffers from widespread well-being issues. Recent evidence from short-term training actions shows potential to address them, but also large variability. Further, DE practitioners face challenges in understanding whether (and for whom) such interventions work, due to small sample sizes, short intervention durations, and the inherent uniqueness of each dissertation. This methodological paper proposes a novel, practice-oriented, and idiographic approach to such understanding, supported by learning analytics of quantitative and qualitative data. To illustrate this approach, we apply it to two datasets from six authentic doctoral workshops (N = 105 doctoral students), showcasing how it can provide individualized practice-oriented insights to doctoral students and help trainers better understand their interventions, while coping with typical limitations of data from doctoral training. These findings exemplify how the triangulation of simple, interpretable analytics models of mixed longitudinal data can improve students, practitioners', and researchers' understanding, re-design, and personalization of such training actions.

Educational relevance and implications statement

Collecting data about the context and process of a doctoral training action can help practitioners and students understand who benefits more (or less) from such training. The individualized analysis of such data, obtained with even very simple technologies, can also help students understand their processes and contexts, to better address progress and well-being issues. The use of student-authored short narratives (e.g., diaries), along with longitudinal quantitative data, plays an important role in these personalized analyses, and the promise of automated qualitative coding makes this approach increasingly feasible.