this seems to be a good analysis of the topic. he explores the various factors which don't appear to be significant and using his experience from a past election he polled, explains his reasoning to account for the results. mr. kohut acknowledges where he's guessing but he has a very educated guess.

As he puts it, "my guess is...."

Mr. Kohut does not state that this was definitely the reason for the discrepancy, but he makes a very strong argument.

Aha ! Yet another "Bradley Effect" or maybe a "Wilder Effect" theory for the foul up by the NH primary polls and commentators' prognostications. But this one (perhaps, but one cannot be sure) bears only subtle differences from the other two, so let's label this one the "Kohut Effect" theory. But gosh, it's still as foggy in reasoning as the whackier ones presented a day or so ago. Here's an idea: since any one is as good, or as purely speculative as the others, why not ignore all three ? In a small sample such as in NH with not very much to distinguish various classes of voters, although the differences in percentages may look impressive, the actual vote counts deal with relatively small numbers, so maybe it's due to ... More »