Jerone Stephens says:

I simply do not see any value in your models or elaborate statistical “models”. You were quite wrong this year on the Senate, but will cover it up by not noting that 54 Republicans are actually going to win, and I think you predicted this would happen in 17% of the time. You are better than Nate Silver who is often wrong, and cannot pick a sports winner or Senate seats for that matter; he is nothing but a commercial hack that gets on comedy shows where they cannot ask any questions of value. I never bother with the models except to laugh at them. Silver consulted 1524 polls he said at the one point where I bothered to look. The polls with running averages, using the mean rather than the median when they do what is still a useless exercise, but confuses every one who follow these things daily to get reinforcement for their beliefs. A few days before elections, I looked a a few polls, looked at the political culture, went back to Robert Dahl’s book on politics and looked at his comments on turnout. I examined an old copy of the 1953 article on why the president in power always loses House seats, and it is turnout. Reagan lost 8 Senate seats in 1986 even though he was popular. Ike lost House seats from 1954 through his presidency. I called my predictions the 2 minute predictor, and got all right with the exception of NC, but got Kansas right. There is simply no need for the type of analysis “model builders” do; you seem like a nice person and sincere but you have a viable academic specialty and are wasting your time on political analysis, and I do not think any serious person is going to pay the slightest attention to model builders in 2016, at least I hope so and we can get back to real polls but only toward the end. This is likely to be a false hope given the vicious 24 hour “news” cycle, but hope springs eternal to coin a phrase. sincerely, Jerone Stephens