Anyone who believes that voter fraud played a significant role in the 2020 elections is not a clear thinker.
From the Hoover Institute: “We focus on fraud allegations with the appearance of statistical rigor. Trump and allies used statistics to claim some election facts would be unlikely if there had been no fraud. The claims fail either because sometimes the “fact” is inaccurate or it is accurate but not surprising. For example, a viral anonymous report claimed Dominion machines added 5.6% to Biden’s vote share. But, we show that the purported Dominion effect disappears as soon as we control for 2016 results, or make any number of other sensible design choices. Other times this is because accurate claims about the 2020 election simply are not that surprising. Trump and his allies claimed it was suspicious that Biden lost 18 of 19 counties that had correctly picked the winner since 1980. But we show that bellwether counties are bad at predicting future winners. Since these counties went for Trump in 2016, Biden’s low haul of bellwether counties isn’t suspicious at all. Likewise, in a lawsuit filed against PA the Texas Attorney General claimed that Biden had a “one-in-a-quadrillion” chance of winning. The probability comes from a report filed by Charles Cicchetti who examined election-to-election changes and the shift from early-to-late votes. We show Cicchetti’s tests are riddled with errors and vastly understate the probability of change. We apply his test historically and show that vote changes he said had a “one in almost infinite chance” of occurring actually happened in 6% of US elections. Our work is intended to help assess the security of US elections. We think it is important that non-partisan election experts evaluate fraud claims–to either identify suspicious results or reassure the public about the safety of US elections.”
Justin Grimmer weighed in on the comments section at Hoover:
* Crowd size is a poor predictor of final votes, there were not more mail in ballots (we deal with that explicitly, you should read the paper!), the thousands of affidavits are largely nonsense or unrelated to fraud (you should read the news reports about this!), can you link to the video of ballots being destroyed? (the GA video was debunked), ballot run off happens all the time and isn’t evidence of fraud. Yes, there was much higher turnout this election, so Biden got more votes, but a smaller vote share than Obama. The Biden vote increases were not statistically impossible (you should read the paper, we cover this!). We also cover this bellwether argument, which is not a very good one and Nixon won Florida, Iowa, and Ohio but lost the presidency in 1960. I don’t know what to do with this last argument, but if we assume a different state of the world things would be different. But that isn’t what happened!
* Zimny-Schmitt and Harris are describing the characteristics of bellwether counties, rather than measuring how predictive they are. So even though bellwethers might tend to have certain characteristics, what really matters is how well they can predict future elections. We show that bellwethers tend to be *worse* at predicting elections than counties with similar election results in the previous election. On the number of losses: it actually isn’t that surprising. If you “rerun” the 2016 election results, you’d expect Biden to lose 18 or more bellwether about 20% of the time. You can see in Figure 3 that Trump didn’t just win these bellwethers in 2016, he won them by a lot. In fact, this reflects a trend where Republicans do better in many smaller population counties, while Democrats do better in fewer high population counties. The result of this is that Biden loses bellwether counties (which exist by statistical accident). Again, if you look at our Figure 3, you’ll see that there really isn’t anything necessary about flipping the bellwether counties for Biden’s win, because he tended to improve his margin in larger counties. That said, Biden did win more counties than Clinton.
* I’d be curious about the correlation between crowd size and support for a candidate. There are related factors (like yard signs) which tend to not be particularly predictive. But I’m not sure of a systematic study of crowd size and candidate performance in an election. One reason to be doubtful of the relationship is that crowds in a presidential election will still be much smaller than the number of votes needed to win. What’s more, the Biden campaign actively pursued smaller crowds because of COVID.
Comments at Andrew Gelman’s blog:
* I didn’t see where the Hoover article covered the discussions I have seen significant reductions in various places in the % of mail in ballots rejected (compared to prior and sometimes very recent elections) as invalid. From what I have read, most of the rejections have typically been for failure to sign the mail in ballot.
I personally found it very comforting that our electorate improved its ballot signing performance in such a dramatic way… if indeed this was as widespread as some have suggested. I haven’t seen this addressed elsewhere, and maybe I missed it in the Hoover piece.
Justin Grimmer: The state where this comes up is Georgia. Gabriel Sterling , a republican election official, explains this as a result of a law change in GA that gives citizens a chance to fix their signature if the election workers find a problem.
“The decrease in rejections is attributable to a recently passed law that gives Georgians a chance to correct problems, such as a rejected signature, with their ballots. Both parties had teams roaming the state and contacting voters whose ballots were at risk of rejection, but Mr. Sterling said the Democrats were simply more prepared for the task.”
One thing that went along with the increase in mail in voting were changes in laws that made it possible for people to fix ballots that would be potentially rejected. This made it possible for the parties to connect voters to potentially rejected ballots and have them corrected. So on the one hand, it does seem intuitive that an increase in mail in ballots might increase signature issues. But on the other hand, the ability to fix signatures suggests that it should decrease substantially.
