Read on Twitter
Millions of Americans still believe the election was stolen. In a new paper we ( @aeggers, Garro, and I) analyze Trump claims about fraud in the 2020 election. We show these claims suffer from statistical errors that render the claims demonstrably false.
https://bit.ly/3cCcAT3
https://bit.ly/3cCcAT3
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: Trump and his allies claimed it was suspicious that Biden lost 18 of 19 counties that had correctly picked the winner since 1980. Viral facebook memes called into question how Biden could have won 81 million votes, but only a small number of counties.
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.
Further, support for Democratic candidates has concentrated in fewer, more highly populated counties over time. Biden won more counties than Clinton and the bigger counties by a larger margin than Obama. So, Biden’s performance is in line with recent trends.
Similarly, 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.
His test also fails as a diagnostic of shifts in vote totals among early- and late-counted ballots. We apply his test to Arizona and show it would flag the shift towards Trump in late-counted ballots as just as surprising as the shift towards Biden in Georgia.
Other claims are simply false. A viral anonymous report claimed Dominion machines added 5.6% to Biden’s vote share. And while Trump’s lawyers spun fantastical tales of long-dead communists subverting global democracy, this report had the appearance of statistical rigor.
But the statistical analyses supporting this claim are based on fishing and poor research design. The purported Dominion effect disappears as soon as we control for 2016 results, or make any number of other sensible design choices.
Other claims rest on poor statistical analyses. @JohnRLottJr claimed to show that suspicious counties in GA and PA padded Biden’s vote total and suspicious counties in swing states had high turnout. These claims were tweeted by the president as solid evidence of voter fraud.
But neither claim is true. As we explain in this thread ( https://twitter.com/justingrimmer/status/1346144084902072320?s=20), Lott’s analysis of Fulton and Allegheny counties depends on the arbitrary order in which “control” counties were included.Once we address this issue,there is no evidence that Biden overperformed.
Similarly, we show that Lott’s accusations about high turnout in suspicious counties are entirely due to including lower-turnout states that had no suspicious counties. Once this is addressed, there is no evidence of higher turnout in “suspicious” counties
Debunking some results required only arithmetic. Others required relatively subtle data analyses. All claims fall apart under scrutiny.
Carefully evaluating voter fraud claims is essential to addressing public concern about election security. We hope our paper provides a resource for refuting claims made about the 2020 election and a road map for evaluating fraud claims in future elections.
