Election Night - Counting Bias by State

[Vosem]

Virginia's very strong R bias on Election Night is well-known. North Carolina typically has a noticeable D bias, though not as strong as Virginia's. Indiana also sometimes has a mild D bias, or at least it had one in 2016.

In terms of the distinction between results on Election Night and the final results, CO and CA have very strong D biases (of several points; Republicans generally have to be up around ~52-48 on Election Night in CA to have actually won the race). Some other, much weaker (won't swing the result by points, but if the election is within a few hundred votes, noticeable) patterns I've noticed include consistent swings toward the Democrats after Election Night in AK and OH, and swings to the Republicans in AZ and NY.

The primary pattern is that the more ideological (Sandernista or Freedom Caucus) candidate nearly always gains in post-Election Night counting; in CA Sanders achieved a 7-point swing post-Election Night (the state went from Clinton+14 to Clinton+7) and in one of the more dramatic primaries in 2016 Andy Biggs went from being down like 900 votes on Election Night to ultimately winning by 16.