Reply #7: It would be interesting to have the transcript of what was really said

[karynnj]

Going to your link, the date of that article was January 26, the day before the NH primary. The author is not clear whether he is speaking of a national poll or a NH poll. The Iowa caucuses were on January 19.

The problem I have is that the author injects his own idea of the motives without any back up. A possible alternative view was that he was saying that he did not believe Dean's jump. It could well have been that that specific sample, by chance, produced an atypically high point estimate for Dean (and of course, an atypically low point estimate for Kerry - to be expected as the estimates are not independent.)

If it was a NH poll, I doubt he would choose the day before the election to put out a fake poll - getting the "wrong" result hurts his company. If he was being insidious, a fake poll a week or two earlier would be better driving people and if that didn't happen, at least it would not be proven that it was not so at that point in time.