Statistical Modeling of Reliability for Redistricting Plans

This report explores how statistical modeling can help assess the reliability of redistricting data produced under new privacy protections in the 2020 Census. By comparing multiple modeling approaches, the authors identify which geographic areas are most likely to retain accurate demographic counts despite added noise. Their findings offer a practical path for understanding the impact of differential privacy on population estimates used in redistricting plans. The work builds on previous research and applies new models to both historic and current census data.