Are you aware of any examples of election forensics in practice? Or at least any applied research on real large-scale datasets (e.g. govermental elections)? Thanks.
2 Answers
$\begingroup$
$\endgroup$
Walter Mebane (University of Michigan) has written a lot on using Benford's Law distributions of digits as tests for electoral fraud, both as a methods problem and in a series of applications. There are several working papers listed here: http://www-personal.umich.edu/~wmebane/
$\begingroup$
$\endgroup$
A quick Google led to a paper by Cantu and Saiegh on Argentinian elections, which appears to have some good references as well. The website FiveThirtyEight does a lot of election prediction and perhaps has some good links. And I seem to remember Andrew Gelman doing some forensics work, but couldn't find a link.