Political scientist and Stan user Josh Alley points to this new paper on using hierarchical (multilevel) models to estimate heterogeneous effects. I like it. The only thing I’d add is an emphasis that group-level predictors can be important. Jennifer and … Continue reading …