Lucianne.com links to a David Catron article for The American Spectator. Catron makes the argument that asking people who they believe will win a presidential election produces better predictive results than asking them for whom they plan to vote. He cites research which finds that superiority in 60 cases out of 77 observations where the prediction and intention differed.
Most of the polls you've seen quoted ask about "who you'll vote for." Catron argues this is because it produces results the people paying for the polls hope to see. These same intention polls predicted the Hillary Clinton win that failed to materialize.
The author cites an average of several prediction polls which show 55% of those polled believe Trump will win in November. I hope they're prescient.
Catron doesn't say so but I believe people are more likely to tell the truth when it isn't their own behavior they are predicting, but that of the crowd. What I believe "the voters" will do can't reflect badly on me.
