Jul.23

electoral-vote.com

Dem 51

   

GOP 49

Today's Presidential Polls

These are old polls and so mean nothing now. It may take a week or more to start getting meaningful polls. One very big problem is that all pollsters correct the sample to match their idea of the electorate. So if the pollster believes 15% of the electorate will be young women and in the sample only 13% are young women, it will count every young women in the sample as 15/13 of a person (which is much better than the three-fifths of a person the Constitution speaks of for some people).

Now here's the rub. The "electorate" the pollsters mean is the set of people who will actually vote. It is not the voting-eligible population (i.e., citizens over 18 who are not excluded felons). The electorate is not fixed. With Harris in and Biden out, some double haters who were not planning to vote may have become single haters and are now planning to vote. This changes the electorate and thus requires a different correction to the polling sample. The good pollsters understand this, but figuring out the new electorate isn't so easy. It could take them a while and they could get it wrong. They won't really know for sure until election day and may be forced to use old models of the electorate that they have been using for a while. It is not a simple problem to solve quickly. The obvious fix is to go out and ask a whole bunch of people: "Are you going to vote?" but then we quickly get into the problem of differential nonresponse: shy Trump voters who won't answer the phone, enthusiastic 18-year-olds who don't know that you have to be registered in order to vote, etc. We hope that our method of aggregating polls from multiple pollsters smooths this out somewhat, but there is little else we can do right now. (V)