Scientific Measurement in Redistricting Research (Princeton University)

We discuss the essential requirements for the measurement of any quantity of interest as applied to redistricting research. Most importantly, a quantity of interest must be defined separately from its measure, without which empirical estimates cannot evaluated or improved. Only with such a standard can we learn about an electoral system or understand fundamental concepts in the field such as legislative compactness, partisan bias, electoral responsiveness, among others (with or without differentially private noise applied to census data), all of which we will illustrate.