More and more empirical studies address doctoral candidates' health. Yet, the mechanisms linking supervision and doctoral candidates' health often remain unclear. We start to fill this research gap with classifications of supervisors produced by latent class analysis, which were introduced into structural equation models with motivation towards the dissertation research as a mediator to predict doctoral candidates' health satisfaction. We used data from more than 200 doctoral candidates from a German university. Three types of supervisor support were extracted (poor support: 18.4%; good support: 26.4%; very good support: 55.2%). Poor support was significantly negatively associated with doctoral candidates' levels of motivation and health satisfaction. The relationship between poor support and health was partly mediated by motivation. By means of the advanced statistical models, mechanisms linking supervision and doctoral candidates' health could be identified and research on the dimensions of (very) good supervisor support could be expanded.

Kunz, C., Schneijderberg, C., & Müller, L. (2023): Well-supervised, highly motivated, and healthy? Using latent class analysis and structural equation modelling to study doctoral candidates' health satisfaction. In: Higher Education Quarterly, 00, 1–16. doi.org/10.1111/hequ.12479