QVF Change Report: September 2025 Edition

This is the first iteration of the new MFEI monthly QVF Change Report. This publication seeks to reveal features of the Michigan Qualified Voter File (QVF) through analysis of the month-by-month changes to the dataset. The series will feature high-level observations accompanied by documentation of the analyses performed. Each report is intended to be used both for data maintenance by Check My Vote users as well as reflection upon how the Michigan data infrastructure can be improved.

Check My Vote is a QVF analysis and maintenance service present in a growing number of states, providing free QVF access to residents of the state. Volunteer users may contribute digital research of individual voter registrations to a shared knowledge-base and coordinate to find and remove invalid registrations. Check My Vote seeks to create transparency in U.S. elections by accounting for every registration on file.

September 1, 2025

Following an analysis of the changes over the previous two month's duration, it was determined that, for a given monthly period, the recorded registration dates associated with newly added registrations poorly represent the period in which they were added to the QVF.

This report performs the same analysis on the changes to the QVF over the month of August 2025, comparing additions between August 1st and September 1st, using copies of the QVF obtained from the Michigan Secretary of State through Check My Vote via FOIA request. Registrations are considered additions if they exist only in the latter of the two copies. Using this subset of registrations, the recorded registration date has been analyzed for its distribution over time as well as by county and jurisdiction.

For the purposes of the study, two cutoff dates have been selected in order to categorize the differences between registration date and date of addition to the QVF. The first date is July 1st, 2025, offering a two-month window in which a registration may be delayed to be considered "on time". The second date is January 1st, 2025, intended to distinguish late additions from critically late additions which experienced over a eight-month delay before being added to the QVF. These categories were intended to account for the natural and expected delay between date of registration and date of addition to the QVF.

There are many possible reasons for this delay, from physical paperwork sitting in a bin over a weekend to waiting-periods between synchronization of the state QVF with county databases. The cutoff dates were selected to account for the natural delay built-into the pipeline. The durations used to select these dates may be adjusted in future editions of this report if doing so produces more meaningful categories.

The primary set of concern for this edition is the category of registrations for which a registration date before January 1st, 2025, was recorded, yet the addition of that registration was delayed until August, two thirds into the year. It can be expected that the registration-to-QVF pipeline accelerates in the months leading up to a general election to overcome a backlog of what is normally a low-priority task, however, these delays stand in contrast to the significant majority of registrations (~90%) having a turnaround of under two months.

Additionally, there are concerning outliers present in this data. Natural systemic factors explain the delay for the collection of registrations clustered around the sample duration (the period in which registrations were added to the QVF, this past month of August), however, in this data, there are recorded dates of registration at least as far back as two decades ago.

An explanation is required as to why data such as this is being added to the Michigan QVF on a monthly basis. If these registrations existed on-file at an earlier date and were removed, what condition has caused them to be returned to the QVF after several decades since their debut? Michigan's QVF is in-part populated by automated systems managed by NGOs. The addition of these registrations highlights the need for the inclusion of the method by which a registration is added in future QVF FOIA requests. With that added data column, it would be possible to explain the existence of any given registration.

Until the data column is added, and the data is either explained or necessary adjustments are made to the systems responsible, these reports will continue to highlight this apparent discrepancy.

The complete results of the analysis performed are attached below, sorted by county and jurisdiction. Although the primary purpose of this report is high-level (this is a state-wide secretarial issue), the results may be useful to CMV users, clerks, and data-analysts for further inquiry into Michigan's voter registration infrastructure and QVF maintenance.