Michael Herron has posted this draft in progress. Here is the abstract:

The aftermath of the 2018 Midterm Election in North Carolina has been marred by allegations of absentee ballot fraud in the state’s 9th Congressional District, in particular in Bladen County, one of the eight counties in North Carolina that intersects this district. Consistent with these allegations, we show that Bladen County’s election returns in 2018 are anomalous in that Congressional candidates in the county had mail-in absentee support rates inconsistent with their election day support rates. Moreover, across the North Carolina Midterm and General Elections of 2010, 2012, 2014, 2016, 2018, Bladen County in 2018 exhibits the most anomalous mail-in absentee versus election day difference of any year–county pairing. A similar conclusion holds when comparing Bladen County’s 2018 mail-in absentee and election day returns to comparable returns in Arkansas, Georgia, and Oklahoma, three states that tabulate election results in a way that facilitates comparisons with North Carolina’s. In addition, Congressional election returns in Bladen County exhibit anomalous mail-in absentee voting patterns in the 2016 General Election and in the 2018 Republican Primary. In short, either Bladen County is idiosyncratic in a heretofore unexplored fashion, or the recent mail-in absentee ballot fraud allegations involving this county merit serious attention.