19.2 Bayesian hierarchical models

From a practical standpoint, Bayesian hierarchical models are similar to the linear mixed models in Chapter 14. They allow us to account for individual- or group-level variability in estimated parameters of interest (ie. intercepts and slopes for regression). They also allow us to make both group-specific and sample-wide inference about trends of interest. What’s more is that the hierarchical structuring of parameters (e.g. individual within plot within site) allows us to “share” information between groups by informing group-specific parameters with global or population-wide “hyperparameters” (if that is not cool I don’t know what is).