Benford's Law

[eve11]

Deviation from Benford's Law is not proof of election fraud. The above article does a good job of explaining this: the bounds on precinct populations and counties do not span the orders of magnitude needed for the logarithmic distribution of first digits to take effect.

Comments

[wolfy_writing]

Very informative! Thanks for sharing!

[penwiper]

I don't know f I've ever said this, but I love you sharing your analytical knowledge. :)

[profrobert]

Wow, the Trumpies must be desperate if they are actually resorting to math!

[nonelvis]

That was really interesting! I've never heard of Benford's Law, and it's fascinating that this kind of distribution even exists.

[eve11]

There is a cool episode of the show "Connections" on Netflix that talks about Benford's Law.

[bexxa]

Thanks for the link! I'm math-impaired but I had wondered about the apparent discrepancy in the vote totals for the candidates.

Is it fair to say, though, that a more in-depth review of the apparent discrepancies would put to rest any concerns that shenanigans were going on?

[eve11]

I don't know; I think any statistical analyses like this is based on an underlying model that would need to be accurate and correct for the conclusions to be valid. Things that might raise red flags would be if eg, a whole stack of ballots all came in, in alphabetical order or something. But I suppose one might be able to run some simulations assuming no cheating and/or fraud and see where various summary statistics would show up relative to the extremes of those distributions?