|Polling Place||Room Capacity||Non-voters in Room||Voter Processing Points||Inside Queue Capacity||Number of Check-In Stations||Average time for check-in||Arrival rate||Target wait time||Alert||Delete Row|
|# of people||# of people||# of people||# of people||# of stations||minutes||voters per hour||minutes|
|Please add rows by pressing the +Add Polling Place button|
|Precinct||Average Total Wait Time (minutes)||Average Total Queue Length (people)||Average Outside Queue Length (people)||Percent of time room is full||Chance Voter Waits Longer than Target||Alert|
|Table has no data|
This tool uses queueing theory to model a single process step at a polling site for which there are limits on the number of voters that may be present at any one time.
The tool relies on a classic queuing model, known as an “M/M/k system.” For this model of a polling site, we assume that voters arrive randomly with a constant arrival rate, and join a single queue that is being served by a set of parallel stations or servers. In a polling site, we assume that we are modeling the bottleneck step in the process; this might be the check in stations, or the voting stations, or possibly the ballot scanning station.
When the bottleneck is the check-in station, then the model reports only on the amount of time a voter would wait in line prior to check in; in effect, we assume that there is minimal waiting at the downstream steps in the process of voting and ballot scanning. When the bottleneck is the voting machines or booths, then the model reports on the amount of time a voter would wait in line prior to going to a voting station; in practice due to space limits and regulations at many polling sites, much of this waiting would typically occur in the line for check in so as to allow voters sufficient privacy when casting their ballots.
In addition for this model, we assume that the polling site has a maximum occupancy that will limit the number of voters that can be in the site at any time. We capture this in the model by imposing a limit on the number of voters that can be in line waiting inside the facility at any point in time. As a conservative estimate, we assume that all stations in the vote process (that is, the number of check in stations, plus the number of voting booths, plus the number of scanning machines) always have one voter and thus reduce the number of voters allowed to wait in line.
This version of the model assumes that when the facility is full, the waiting line to vote will extend outside the facility. We then assume that any arriving voter will join this queue and wait.
In addition to using the tool on this web page, you can download an Excel spreadsheet that will perform the same calculations. One advantage of the spreadsheet is that it is easier to analyze multiple precincts at one time.
This tool is designed for election officials hoping to understand line lengths and wait times with simple social distancing measures in place. This tool is meant to answer questions like:
This tool comes in an online form and an excel-based form. We recommend using the online tool to try a few different polling place setups, understand what the tool outputs represent, and what inputs impact the results most. Once comfortable with these dynamics, the excel-based tool (downloadable online) is best for running tens or hundreds of scenarios, saving results, and seeing more detailed outputs.