In a recent paper @JohnLottJr purports to show evidence of fraud in GA and PA in the 2020 election. In a new paper, we ( @aeggers, Haritz Garro, and me) show that Lott’s claims are based on a fatal flaw. Once corrected, there is no evidence for fraud.
https://www.dropbox.com/s/ps5oufevtka3nmf/Fraud.pdf?dl=0

Lott’s conclusions about fraud in these counties are entirely dependent on the completely arbitrary order in which pairs of precincts in other counties are entered in the dataset. The conclusion is reversed when an alternative and equally justified data entry rule is used.

Lott’s arbitrary ordering of precincts matters for his conclusions: any alternative rule would produce weaker apparent evidence for anti-Trump fraud in Fulton County and almost any rule would produce weaker evidence for anti-Trump fraud in Allegheny County

When we replace Lott's unusual specification with a more standard approach that does not depend on arbitrary coding rules, we find absolutely no evidence for fraud in either Fulton County or Allegheny County.

Engaging with this work is important. Lott’s paper has been retweeted by the president and received widespread attention. In fact, Lott appears to be circulating his results in meetings with state elected officials to persuade decertification of results. https://www.politico.com/news/2021/01/03/trump-georgia-election-454122

Trust in elections is fundamental to democratic representation. We can be reassured that even though this study purports to show evidence of fraud, its claims are baseless.

Adding the correct tag for @JohnRLottJr