Yesterday we looked at the polls coming out of Michigan and noticed some evidence of herding. In other words, the polling doesn’t show the normal distributin you would expect in which some polls show one candidate up and another down almost at random. Polls should create a bell curve which gives you an idea where the race actually is, i.e. somewhere near the peak of the distribution. What they should not do is all fall within a point of one another.
Most of the recent polls in Michigan show a tied race with less than a point between the candidates, but a few recent polls also show Harris up 4-5 points over Trump. And that’s it. There hasn’t been any other variation this month, which is odd. The herding theory suggests that pollsters have made so many adjustments to their data to avoid the mistakes they made in 2016 and 2020 that they have essentially ironed…
