Data analysis and simulation: Optimizing voter wait times

摘要

The bipartisan election commission formed after the 2012 election recommended that no American should wait longer than 30 minutes to vote. However, in every presidential election year, stories surface of voters having to wait several hours. Long lines disrupt voters' schedules and hinder economic activity, but can also discourage voters from remaining in line to vote. One way to decrease the average and maximum voter wait times is to better prepare polling locations by staffing optimally and having enough voting booths available. Data was collected from a Williamsburg polling location in Virginia during the off-year November 2015 delegate election. Simulation analysis found that in order to have maximum wait times of less than 30 minutes in this Williamsburg precinct during a presidential election then at least 4???5 poll workers to check in voters and 12???15 voting booths or machines are needed. Data on the number of voters that arrive per hour and the amount of time it takes to check in and vote are often collected by the state or by certain polling places. A general, free version of this discrete-event simulation was created in Java. This resource allocation tool takes previous data as an input and estimates the number of voting booths and staff needed in order to keep approximately 99% of wait times less than 30 minutes. Simulation and statistical analysis are used to determine the number of resources necessary.