Monday Nov 28

Fall 2022

Election Maintenance is Either Suppression or Protection

Kicking eligible voters off the rolls is a time-honored tool of voter suppression, but it’s seen a revival since the Supreme Court gutted sections of the 1965 Voting Rights Act in 2013. The Brennan Center tracked a significant uptick in the number of voters removed from rolls between 2008 to 2018, as well as a significant uptick in the number of counties removing high percentages of voters from their rolls. They also found an increase in the number of activist organizations seeking more aggressive voter purges.

Legally, whatever the state or county, voter purges ought to only remove the names of people no longer eligible to vote in a jurisdiction: the deceased, those who have moved, etc. But as the 2020 election aftermath showed many Americans, our electoral system is not a unified entity. It’s a dizzying patchwork of state and local officials, boards, judges, and volunteers, each of whom with their own policies and procedures.  After finding an erroneous purge of 40,000 in Ohio in 2019, the Voter Purge Project was formed to closely monitors how states and counties comply with the law.

Federal law, like the National Voter Registration Act, puts some restrictions on purge processes. But as the Brennan Center and the Voter Purge Project have found, several states and localities have done overtly illegal purges, others have implemented rules that violate federal law, and still others have just aggressively executed purges rather than explicitly changing rules.

Going into the midterm elections in 2018, purges occurred in North Dakota, Ohio, Missouri, Kansas, and New Hampshire.  Kris Kobach, then Kansas Secretary of State, recruited other states to use a database in order to purge their lists –a database that was found be wildly inaccurate. The Secretary of State in Georgia, now the governor, purged 1.5 million voters in that state, including 500,000 in 2017 alone. Texas was barely prevented from implementing a massive, unjustified purge.

Aggressive purging has acquired new legitimacy after the Supreme Court ruled by a 5-4 majority in 2018 that Ohio’s process that year was legal.  Voters could not be purged for not voting, but they could be purged if they had not voted and did not respond to a mailed notice seeking to confirm whether they had moved. Ohio has been among the most aggressive states in this area, sending a notice to all voters who do not vote in any federal election cycle.  Georgia has a similar system. All states purge voters at some level as part of their list maintenance process, but under the cover of the Supreme Court decision, more have interrupted the decision as a greenlight to move more aggressively in this direction. 

This was all before Trump’s claims of election fraud in 2020 turned elections themselves into the Republican Party’s defining issue.

Voter purging is an area of election procedure extremely vulnerable to those who’d seek to unfairly disenfranchise voters, especially given the wide variety in how each state maintains their voter rolls, and their dependence on registrars and officials at the county level for accurate and true numbers.  In the Ohio case in 1999, the Voter Purge Project found the problem also cuts the other way, since the majority of the errors in that purge involved the state ignoring or overlooking accurate figures on voting records being received by the two largest counties in the state, which also happen to have been Democratic strongholds being monitored by a Republican state government.  There’s a major need for this to be monitored.

In 2020, we carried out just such a monitoring and alert program in the critical battleground state of Georgia among many others.  The Voter Purge Project developed a unique system for acquiring, querying, and analyzing state voter rolls to detect patterns of disenfranchisement. We drilled down to the individuals we believed to be improperly purged, reached out to them to let them know, and had respondents re-register to vote.

Further, in some formerly preclearance states, registrars still collect racial data for each registered voter, in addition to their address and voter district. This additional data is critical, because it allows researchers to investigate whether voters of any one demographic are purged more often than others.

By integrating and analyzing voter roll data with other relevant census and contextual data, we’ve sought out patterns of wrongful disenfranchisement and discrimination related to race, age, gender, and other classes, within a state, region, and/or nationally.

Since 2019, the Voter Purge Project partnership has been collecting voter lists from states and processing to monitor purges and other data practices.  We were able to process more than a dozen states on a regular basis and a few less frequently, until we cemented a partnership with Catalist, the well-known, long experienced data operation based in Washington, D.C.  We are on our way to assembling and processing all fifty states and the District of Columbia, but it is a huge undertaking, requiring the collection, handling, and sorting of lists from each jurisdiction multiple times.  Currently, we are at twenty-nine states and counting.  By the end of November, we will have all fifty in our system.  With the midterms looming, we wanted to share observations and analysis of the states we have on hand, because they tell a critical story, even if not yet the whole story.

State by State Analysis

Each state approaches its purges differently, and each tracks different information.

Mississippi, which tops our list for the highest purge rate, engages in what we call “soft purges,” meaning, basically, the state can purge a voter by changing their voter status to "purged". They are still in the voter file but can't vote.  The state does not automatically reassign them.  They have to register again.  This produces more voter records than voters, and in the case of Mississippi, more voter records than citizens.  After a period of time, they remove the records with a status of purged.  This occurred in the past year.  We believe this process adds an unnecessary stage to the process that opens the door for additional errors and erroneous purges, but more investigation would be required to highlight where those are occurring. Nevada is the only state in the country that allows registrars to purge voters based on their personal knowledge of a voter’s death, which perhaps contributes to its high rate. Again, further study would be needed.

In the table below, we note the state’s overall purge rate, which party affiliation’s voters saw the highest purge rate (Democratic, Republican, minor parties, or no party affiliation), and the age group and gender that were purged at the highest rate.

Arizona, Nevada, Georgia, and Florida, all seen as battleground states and bellwethers for each control of Congress and the White House were in the top ten states for purges.  In the anarchy that characterizes the state-by-state election system, even for federal races, different states collect different data or no data at all.  Although our methodology of comparing apples to apples as we receive the list from each state on a frequent basis allows us to tabulate the purge rate, other data that depends on something other than the pure data file is harder to compare conclusively.  Some states don’t share age data.  Many don’t collect racial data.  Others don’t note the gender.  Some states make the process easier and some harder. 

Age is an interesting barometer for purges.  Not surprisingly, of the states we’ve shared, the 65-and-over cohort is usually the largest, as most would suspect, because that is where we would assume mortality is the largest factor.  Yet, even among these thirty states, we note that the highest level of purges also can sometimes be found in the younger cohorts.  Arizona purges 45% in the 18-24 grouping and California purges over 6% in that age group, leading all others joined by North Carolina and New York State, where this is also the leading age group being purged.  We might speculate that these states simply experience higher mobility rates among younger voters triggering purges for changes of address, and that might be accurate.  We are surprised at how often the purge rates are high in this age group when many are registered by still using parental home addresses and voting in their “home” states even while at or leaving university or in their first jobs and still settling.  Actual field tests on the doors too frequently have indicated to VPP canvassers errors in this grouping.

Looking at just a sample of the states that provide racial demographic data in their voter files, the Voter Purge Project offers North Carolina, Georgia, and Florida as cases in point, using just the pure voter list they provide.

If non-white voters are aggregated, their purge rate is dramatically higher than the purges for white voters in North Carolina.

Once again, the aggregate on non-white purges is higher than whites.  Interestingly, the purges implemented by the Secretary of State’s office for Blacks were slightly lower than the statewide percentage of Black residents of the state.  Though it is impossible to establish categorically, the project wondered if the scrutiny over purges, particularly of Black voters, after the race between Representative Stacie Abrams and then Secretary of State and now Governor Brian Kemp, who initiated the purges, was as factor in such a close alignment. 


The story of Florida is much the same as Georgia and North Carolina.  Nonwhite voting groups were purged much higher that whites, but the numbers are more disturbing.  The purges of Black voters were almost identical to the purges of white voters in percentage terms with Hispanic voters dropped at a slightly lower rate. Together, Black and Hispanic purges were of more concern in absolute terms.  Whites are 61.6% of the population with Blacks at 12.4% and Hispanics at 18.7%.  Given the seniors and snowbirds that have surged in Florida for decades, demographically we might have expected the purge rate among whites to be higher.  The significant publicity and legislative action to bar the formerly incarcerated from the roles after the popular referendum mandate enfranchising them may be a factor in these numbers as we dig deeper. 

What’s In Store for Our Future: 

Democratic Voters Purged at Higher Rate in Berks, Fayette, and Lancaster Counties, Where Trump Acolytes Have Captured Election Office

Observers far and wide are concerned about the political polarization in the United States, and have speculated on whether or not the claims and misinformation around election security could be a significant threat to coming elections and the democratic norms that have existed the country.  In the 2022 primaries in Pennsylvania for example, this concern has already manifested in disturbing ways.  Three counties, Berks, Fayette, and Lancaster, delayed certification of election results, delaying the final certification statewide by the Secretary of State based on allegations about the election system and voting mechanism. 

In trying to anticipate future problems, the Voter Purge Project examined how voter purges were being conducted in these states to see if there is reason for concern.  Our analysis and comparison of voter rolls in three Pennsylvania counties between January and July 2022 showed that the percentage of purged versus total voters was higher for Democrats than Republicans in each. The raw total of Democrats was also higher in Berks (53% of all purged, 1590) and Fayette (58%, 656). This contributed to the overall loss of registered Democrats in both counties (Berks saw a net loss of 201 Democrats and Fayette saw a net loss of 964).

Also notable was the high number of voters between the ages of 15-34 purged, especially in Berks and Lancaster Counties.

In Miami-Dade County, the Number of White Registered Voters is Rising While Other Demographics Fall

Pennsylvania isn’t the only battleground state where greater scrutiny is likely needed.  Florida which has become ground zero for two 2024 presidential aspirants, claiming both former President Trump and current Governor Ron DeSantis, who is seen as rising in the Republican polls.  Miami-Dade County is a contested area in any Florida election.  The Voter Purge Project looked closely at how voter purges are being handled there.

American Indians were purged at a greater rate than other race groups, followed by those who identified with more than one race and Black Americans. Furthermore, American Indians, Black Americans, and multiracial Americans suffered a net loss of voters–meaning more voters were dropped than were registered in this period.

Democrats are Purged at a Higher Rate than Republicans in Miami-Dade County

Summary to Date

These are just snapshots on how purges are being handled as part of the larger mission of the Voter Purge Project to rank all fifty states on how they handle and maintain voter lists.  In a report in preparation looking at purges, access, transparency, and other issues, we will be ranking all of the states so that the best and worst on these metrics will be available in fuller detail.  We are also drilling the data down to the county and parish level so that the anomalies will be more self-evident and can be addressed more specifically or, where needed, can be remedied more accurately.  In short, stay tuned, more is on the way!

Having processed and parsed voter lists for an increasing number of states over the last three years the one conclusion that we can offer categorically is that purges are important and need to be monitored extensively if we are to maintain even the semblance of an election system with the pretense of being fair and democratic, regardless of who is in place to do the job, or what party might prevail in any particular state.


*The Voter Purge Project is a join partnership of ACORN International, the Labor Neighbor Research & Training Center, and the Ohio Voter Project.

The Voter Purge Team: David Thompson is the research director of ACORN International, Jessica Buttermore manages digital and other communications for the ACORN family of organizations, Joseph Woods is a data processor for the Voter Purge Project, James Goedert is a data and geocoding consultant for the VPP, Mark Madere is the data manager for VPP, and Steve Tingley-Hock is a data engineer who directs the Ohio Voter Project.

Joomla! Debug Console


Profile Information

Memory Usage

Database Queries