Benford Analysis is dangerous to use for election results. The statistics it uses rely on numbers being distributed across samples sizes that span several orders of magnitude (an order of magnitude means multiplied by 10...so 1, 10, 100, 1000... each increment is one order of magnitude). Election districts tends to be fairly equally sized though, which makes Benford Analysis inconclusive.
It works for things like taxes because expenses tend to have a very large range, e.g. things that cost $1 up to things that cost $10,000. (4 orders of magnitude). It is a very reliable indicator of manipulated expenses in this realm.
But I have to concur it is the wrong way to approach exposing election fraud. It won't be convincing to anyone skilled in the art.
Benford Analysis is dangerous to use for election results. The statistics it uses rely on numbers being distributed across samples sizes that span several orders of magnitude (an order of magnitude means multiplied by 10...so 1, 10, 100, 1000... each increment is one order of magnitude). Election districts tends to be fairly equally sized though, which makes Benford Analysis inconclusive.
It works for things like taxes because expenses tend to have a very large range, e.g. things that cost $1 up to things that cost $10,000. (4 orders of magnitude). It is a very reliable indicator of manipulated expenses in this realm.
But I have to concur it is the wrong way to approach exposing election fraud. It won't be convincing to anyone skilled in the art.
This may help: https://www.youtube.com/watch?v=etx0k1nLn78