First published: 29 September 2020
They’ve been researching this stuff for a while. I came across it due to a piece in The Economist:
Check out this nifty correlogram:
Then this distribution graph:
Now, I could come up with an alternative explanation – namely, laziness on the part of the election officials (just round off ballot counts, I guess). But it sounds like maybe they did have target percentage goals, but that they were spread out like that… no clue. That would take a lot of coordination, I assume. Would like to know where these polling stations are.
Maybe I’ll dip into the data later: