A statistical model where the model parameters are random variables. In case of clinical trials, the reason for heterogeneity among the included studies in a meta-analysis is explained by not only variation within each primary study but also between the primary studies.