Replication Data for: Forecasting US Voter Turnout

Voter turnout is a crucial indicator of democratic health, yet forecasting turnout remains an understudied area in political science. This article presents two pioneering models for predicting U.S. presidential election turnout: The National Model and The State Model. The National Model, using data from 1868-2020, employs lagged turnout as its sole predictor. The State Model, covering 1984-2020, incorporates demographic and institutional variables to forecast state-level participation. The National Model predicts 65.3% turnout for 2024, while the State Model forecasts increased turnout in 41 states compared to 2020. The models' ability to generate early predictions offers valuable lead time for planning and resource allocation, which has implications for election administrators and political campaigns as well as for the vibrancy of civic engagement in America.