We’ve been discussing how to use population margin information about poststratification variables (see Modeling Complex Contingency Tables and SynthMargins & Bayes-Raking). I want to connect two terms for this we’ve mentioned a lot in this series: the logit shift and …