In Michigan, a Republican in a local election went from barely losing to winning clearly after picking up ~1200 votes in a recount. That seems like a lot to me, which got me wondering. What is the most a vote has changed as a result of a recount?
5
-
2Was there actually a recount in that example? The current article doesn't use the term, and the Breitbart article only mentioned it in a quote of the winner (then thinking he lost) saying he wouldn't seek a recount.– JJJ ♦Commented Nov 15, 2020 at 18:39
-
@JJJ I don't know how else they would have caught the mistake. There must have been some kind of debugging process following the election or the erroneous result would have stood.– Ryan_LCommented Nov 15, 2020 at 18:48
-
3@Ryan_L the article mentions the following statement from an Oakland County Clerk: “This is proof that our process of checks and balances works. A methodical canvass is an essential tool to ensure an accurate count and precise results.” I don't know what that "methodical canvass" is, but I think it could be an interesting new question.– JJJ ♦Commented Nov 15, 2020 at 18:52
-
You seem to mean change in absolute number of votes, not pecentage.– Keith McClaryCommented Nov 16, 2020 at 20:15
-
@KeithMcClary I think both are interesting.– Ryan_LCommented Nov 16, 2020 at 20:56
Add a comment
|
1 Answer
Ballotpedia has an article on Noteworthy recounts in the United States. For 2000-2015, they cite a report from FairVote:
From 2000 to 2015, there were 27 recounts in [4,687] statewide races across the country. Of those 27, there were three that resulted in a change of the election result. The average vote shift across all 27 recounts was 282 votes, which accounted for 0.019 percent of the vote in those races
That page also has a section on "Notable recounts" from a few different years (including the 2000 Bush/Gore recount). The biggest swing they have listed is the 2018 Florida Senate race, where the Republican challenger's vote lead decreased from ~15,000 votes to ~10,000 votes after both machine and hand recounts. Notably, this change in votes is larger than that of the three races where the result actually changed, because they were all closer than this to begin with.
It's certainly possible that there have been larger recount swings before 2000, or in single-district races which FairVote didn't look at, but it's unlikely that anything smaller than a state-wide election had a vote shift larger than that 5000 vote Senate one, simply because of the smaller number of votes involved.
