Abstract
Multilevel regression with post-stratification, short MrP, has quickly become a standard model to measure public opinion across e.g. geographical units. This chapter shows how it outperforms other alternatives and illustrates MrP with concrete examples. Readers are walked step-by-step through the construction of a MrP model – code and data are available for replication. We also discuss and illustrate the relevance of context-level factors and how to generate uncertainty estimates of our measurements. The chapter ends with a discussion of extensions and ongoing research projects that promise to further improve MrP.
Leemann, Lucas