Background and Overview

During independent processing of Detroit's scanned ballot images from the 2020 Presidential Election, a striking pattern emerged: thousands of ballots had the write-in oval filled in for various contests, yet no candidate name was actually written in the space provided. These "ghost" write-ins—marked ovals with no accompanying text—raise questions about voter intent, especially given the absence of any ballot instructions suggesting that filling the oval alone expresses a protest or "none of the above" vote.

Scale of Write-In Votes

Across all processed ballots in the City of Detroit:

A 4.47% rate of write-in activity is unusually high for a presidential general election, particularly when the vast majority of these marks were blank (no name written). Standard voter behavior in general elections typically results in write-in rates well under 1%, and almost always accompanied by an actual name.

Comparison: Early In-Person vs. Election Day Precinct Voters

To determine whether this pattern might be explained by ballot design or normal voter practice, the ghost write-in rate was compared across voting methods:

Election Day in-person voters were more than 2.5 times as likely to mark a ghost write-in as early in-person voters. This substantial difference strongly suggests the behavior is not simply a response to ballot design or standard Michigan voter practice, as both groups used identical ballot styles.

Case Study: A Representative Ghost Write-In Ballot

The following ballot is a clear example of the pattern: 00001_00001_000221

This voter selected the Straight Democratic Party Ticket, which automatically votes for all Democratic candidates in partisan contests. Despite this, the voter overrode several straight-party selections by deliberately filling in blank write-in ovals:

• State Board of Education (vote for 2): Marked the first Democrat but skipped the second Democrat instead filling the blank write-in oval. So this process is removing votes from the voter's party.
• University Boards (Michigan, Wayne State, Regents): Ignored Democratic candidates entirely and marked blank write-in ovals—effectively withholding votes that the straight ticket would have provided.
• Prosecuting Attorney: Bypassed the Democratic incumbent and marked the blank write-in oval
  (a contest with notably high ghost write-in activity across Detroit).
• Sheriff: Voted Democratic as expected.
• Clerk: Voted Democratic as expected.
• Treasurer: Again chose blank write-in ovals instead of the Democratic candidates.
• Register of Deeds: Again chose blank write-in ovals instead of the Democratic candidates.
• County Commissioner: Voted Democratic as expected.

You have to wonder why someone who chose to vote straight party, would intentionally remove votes from a candidate.

The most striking behavior appears in the non-partisan judicial race for Circuit Court (15 positions):

• The voter selected only one actual candidate (Don Knapp).
• They then carefully skipped the first write-in oval and filled in the next 14 consecutive blank write-in ovals—precisely enough to cast votes in all 15 slots without choosing any other listed judicial candidates.

This deliberate and time-consuming pattern—counting and marking exactly 14 blank ovals—appears on many ballots and strongly suggests intentional action rather than random or mistaken marking.

The judicial races with large numbers of seats produced some of the most unusual voting patterns observed and will be examined in greater detail on a dedicated page.