Folks, the exit polls are garbage this year. They just can't not be. They'll try to fix this by weighting (as they tried in 2016; most of the numbers that got publicized that we remember were NOT the final numbers) but we can't weight our way out of this degree of non-randomness. https://twitter.com/bpmehlman/status/1324157591082192896

First, these are already *preliminary* (meaning they will be adjusted with weights) and I think it is irresponsible to publish them now. Of course people will think they mean something if they are published. Two, after adjustments, they should be published with giant caveats.

Three, feel free to ignore them. This year, we have precinct level vote counts. That's data. Most everything else is vaporware for purposes of *precision*.

Yeah everyone remembers the wrong figure even from 2016—same problem, preliminary data published and went viral— and I actually don't trust the final one either. 2016 was also a realignment year aka we don't know really the weights. https://twitter.com/cschneider8224/status/1324353953719635969