Reply #18: Many of us strongly disagree with Skinner on this

[unc70]

With the greatest respect, I disagree with Skinner's conclusion. The problem is that he relies on the exit polls to validate the voting. Unfortunately, the exit poll was apparently being re-weighted to compensate for differences between the sample versus the underlying population. For example, if the percent of women in the sample was less than the percent of women who voted, then you might multiply the repsonses of the sample women proportional to the undersampling.

Unfortunately, the exit poll results appear to have been also weighted based upon how they voted. While such adjustment allows better analysis of things like important issues, etc., they are not valid for use wrt election validation.

Last I heard, we don't have access to the raw polling data.