How to assess the impact of fellowships on academic careers? Latent transition analyses for funding programmes of the Alexander von Humboldt Foundation

Abstract Although fellowship programmes continue to be seen as an effective means of advancing young researchers' academic careers, the impact of fellowship programmes on fellows' career development is still unclear. The central question of this article concerns the evaluation of fellowship programmes: What methodological challenges does the evaluation of fellowship programmes pose with regard to career development, and how these can be addressed in the context of evaluations? Specifically, there are three key methodological challenges facing research evaluation in the context of career development, which can be described by the terms 'impact', 'validity and fairness', and 'tailored programmes'. A career is understood as a transition between positions over time; career stages can be understood as latent categorical variables, i.e. types of career stages (temporary, full-time). Transition is modelled statistically using latent transition analyses within a person-centred approach. Covariates, such as funding programmes, can impact both the initial configurations (i.e. the frequency of fellows in different career stages) and the transition itself. A funding programme is fair if all fellows, regardless of their characteristics (gender, career stage, cohort), have the same chances of success. Different types of fellows with different career trajectories indicate heterogeneous subpopulations that require tailoring of funding programmes. The approach is illustrated with data on the career development of 1418 fellows from three Alexander von Humboldt Foundation programmes. The majority of fellows benefit in their academic careers from the funding, but the null hypothesis of no specific effects (e.g. programmes, age, gender) could not be rejected (endogenous and homogeneous trajectories).