I recently read Kubsch and colleague’s (2021) paper in Practical Assessment, Research, and Evaluation, and it is revelatory. It not only introduces Bayesian statistical methods in the context of science education research, it shows: a) a case in which prior knowledge can be meaningfully brought to bear upon the analysis and b) how the use of a prior can be critiqued/compared with the effects of a less informative prior. It also compares the analysis to the frequentist approach nicely. I think this could be a foundational piece in science education research.
https://scholarworks.umass.edu/pare/vol26/iss1/4/
One question I have for the authors: I believe the outcome was created as a composite of individual items. Is it possible to model (measurement-related) uncertainty in this (latent) factor/construct using Bayes?