Re:Looking for Voter Fraud (In Old Elections) with Data Visu...
by aceon48 on Sunday November 08, 2020 @09:40PM
(#25030945)
This is too high level. You would have to look at individual state levels, particularly swing states.
[ Parent
| on HN ]
Re:Looking for Voter Fraud (In Old Elections) with Data Visu...
by timtylin on Sunday November 08, 2020 @10:36PM
(#25031230)
The more I think about it, the more I believe Benford's Law can be violated too easily within more urban areas, especially when precinct sizes are artificially limited to a fixed size (500~1000 people) without geographically-induced truncations. As long as voter turnout is consistently above 10-20%, with the 2-party system you're simply not going to get the vote count number to span the several magnitudes needed to induce Benford's Law. (Exception of course for smaller/independent candidates, and perhaps if one of the major candidates really sucked and kept getting 10% of the votes)
Re:Looking for Voter Fraud (In Old Elections) with Data Visu...
by mikexstudios on Sunday November 08, 2020 @10:57PM
(#25031318)
https://github.com/cjph8914/2020_benfords/issues/9 for more context.