Quote:
Originally Posted by AllTheCheese
She was leading the whole way in the polls, but they still dipped post Comey. After checking, the dip in the averages was less than I remember it, but still enough to have tipped the election. Trump gained a point nationally and HRC lost one.
Change each state -1 Trump and +1 HRC, and Clinton wins Wisconsin, Michigan, Pennsylvania, and even Florida.
This is horrible logic.
1) Regardless of the swing, she was still leading going into the election. Even assuming polls are good predictors of election day voting behavior (which isn't necessarily a safe assumption), that's actually a fairly minor change.
2) There's no way of knowing what caused the change in the polls. It could have been people reading about the Comey letter; it could have been random noise caused by differences in who got called by pollsters.
3) Why do you think Comey-letter vote-switchers were particularly likely to be concentrated in the swing states you mentioned?
4) Regardless of the cause of the change in polls, why do you think the people responsible for the 1% change in both directions were likely to effect the region you mentioned?
In other words, this seems to be ruling out the possibility that the polls were inaccurate all along as a result of sampling error, outdated methods of data collection, and plain-old faulty model construction on the part of pundits.
Last edited by DrModern; 04-09-2017 at 07:40 PM.