On March 22, 2022 the Nevada County Board of Supervisors was handed a report during the public comment period, “Analysis of the 2020 General Election.” The report was prepared by Paul Gilbert of the Citizen Auditors of Nevada County. Because the contents of the report are critical of the Nevada County Elections Office and the processes employed to maintain accurate voter rolls, and because the author is a candidate who seeks to become the next County Clerk-Recorder, we requested a copy of the report for our review.
After a close look at the report, Nevada County Elections is of the opinion that the report does not describe any discernable problems with our office’s voter list maintenance activities. The report rather unfortunately demonstrates a fundamental misunderstanding of voter behavior and fails to employ basic best practices for analyzing data. The report also fails to accurately articulate the laws governing voter registration list maintenance and bases its findings on that lack of understanding.
There are several parts of the analysis that compare data from a voter roll produced by the Nevada County Election Management System (EMS) on June 4, 2021 to the tabulated election results for a single contest—the November 3, 2020 presidential contest—generated by the Hart Verity system. This analysis is problematic in at least two ways.
First, the data being compared represent different snapshots in time. Pulling a voter roll, which represents the county’s current active registered population, eight months after the November 2020 election, and comparing those ballots cast to a single contest, which is representative of who was registered and voted at the time, will always show a difference in totals. The report asks you to be worried about the difference between two dissimilar things. The difference is nonsensical, and this office contends that there is no reason to worry. Registration data change daily, and in fact it is normal. For more about voter registration trends in Nevada County, our office encourages everyone to read our “Election 2020 Year End Report,” found at https://www.mynevadacounty.com/DocumentCenter/View/40714/2020-Year-EndReport_FINAL.
With respect to the November election, the number of people registered when the November election took place was 75,123. The number of active registered voters when the voter roll in question was pulled was approximately 72,000. Again, these are two different moments in time. Nowhere in the report does it occur to the Citizen Auditors that this office’s diligence in its voter list maintenance duties may explain the difference in registered voters. People re-register with the Department of Motor Vehicles, move, turn voting age, and die every day. It is only logical to conclude that there will be changes to the voter roll over eight months. The Citizen Auditor report takes none of these nuances into account, but rather says the differences cannot be explained. Clearly, they can be explained.
Second, the report conflates ballots cast with tabulated results without any explanation why these datasets are comparable or any reference to overvotes or undervotes not included in the vote total. Strangely, the author instead subtracts one number from the other and thinks that the difference between these two dissimilar things represents a problem that requires investigation. This is a careless treatment of data. To be clear: an overvote cannot be counted because the voter marked more choices than is allowable under law and the voter’s intent cannot be ascertained by the physical marks left on the ballot. An undervote occurs when the voter makes fewer than the maximum number of choices allowable in a contest and it’s clear that the voter intended to do so, or when the voter chooses to write-in a name that does not appear on the certified list of qualified write-in candidates (e.g., “Mickey Mouse”). Even in a high stakes presidential election, there are always voters who choose to skip the top-of-ticket contest, cast invalid write-in votes, and overvote. It is entirely possible for a voter to get credit for voting and skip the contest in question—in fact, it is very common for a voter to skip contests. The report takes none of this well-known voter behavior into account.
Our office was also struck by the report’s blatant failure to accurately describe the laws and processes governing voter list maintenance. The report repeatedly refers to the Help America Vote Act of 2002 (HAVA) without mentioning other, more relevant laws at play. Specifically, the National Voter Registration Act of 1993 (NVRA) provides the basis by which all states implement their list maintenance programs and Chapter 3, Division 2 of the California Elections Code, which articulates the specific state and county responsibilities to maintain the voter roll.
The report’s other major miss is the failure to understand that removal from the voter roll is predicated on evidence that the voter has moved, died, or has another change in status that would disqualify them from registration. Failure to vote in itself is not a reason to remove a qualified elector—in fact, that restriction is specifically articulated in the NVRA (52 USC § 20507(b)(2)) and implied in state law (Cal. Elec. Code §§ 2200-2201). The report tables that specifically refer to evidence on non-voting in Nevada County do not take the needed steps to understand how many of those non-voters might have received notifications from our office that their residential address changed (and importantly, how many of those notifications were returned by the voter), how many are new registrants, and how many died or were otherwise disqualified from voting.
Gregory J. Diaz
Nevada County Clerk-Recorder
Registrar of Voters
Voter registration lists are complex and dynamic systems that depends on state and local election officials and voters to maintain. Maintaining voter rolls present ongoing challenges to all of America’s states and localities, and over the years the process has been imperfect. In no way does Nevada County Elections claim to have a perfect list maintenance process, nor does our office have any opinion as to the current public policies governing list maintenance. We do, however, think it would be unwise to rely on the Citizen Auditors’ “Analysis of the 2020 General Election.”
