Press "Enter" to skip to content

Finding Unusual Voters

The last two big U.S. elections have rekindled discussion of voter fraud. There have been claims that non-citizens are voting, of people voting more than once, and of dead people voting. I wondered if any of these claims could be tested by analyzing publicly available data on voter registrations and history. To this end I have built some tooling around the Florida voter data.

  1. I looked for voters who are unusually old. I built a report that shows voters who have voted in Florida as supercentenarians. This list looks a little suspicious to me because only one of these voters shows up on Wikipedia's list of American supercentenarians.
  2. I looked for duplicate voters. There are lots of ways to do this. As a first attempt, I built a report that simply looks for voters who have the same first and last name, street address, and birth date. This finds quite a few people who are registered more than once, if they are the same person. In many cases, the middle name is different. Some of these cases may be twins, but that seems unlikely to me. In some cases, the potential voter has voted more than once in the same election. If these are indeed the same person, these are invalid votes. In many of these cases, the voter was voting under one voter ID and then switched to another. This is likely due to overzealous or sloppy voter registration efforts. The voter may not even realize that he or she is registered multiple times. At the very least, the duplicated registrations should be investigated and cleaned up by the local supervisor of elections.
  3. I noticed that the total number of unique votes is different when looking at unique counties in the vote history. That should not be possible as one cannot legally vote in two counties at once. I made a report that shows these double votes.

At the time of this post, the reports are using the data published on November 13, 2018 by the Florida Division of Elections. The reports will automatically update when newer data is published.

Comments are closed.