Replies

6 posted on 09/25/2012 8:02:07 AM PDT by The_Reader_David (And when they behead your own people in the wars which are to come, then you will know...)

18 posted on 09/25/2012 9:13:04 AM PDT by Doctor 2Brains (If the government were Paris Hilton, it could not score a free drink in a bar full of lonely sailors)

Reweighting polls with skewed samples (that provide enough internals to allow this) can lead to a more accurate result. However, I’m not certain that unskewedpolls.com’s use of Rasmussen’s August 2012 party affiliation data as weights isn’t too extreme a correction, leading to too rosy a picture of GOP prospects this fall.

I agree. It's a little suspicious when Rasmussen part-ID weighting when applied to these other polls gives Rasmussen a significantly larger advantage than Rasmussen's own poll, which recently has had had Romney down a point or two.

It would seem that these "unskewed" polls are doing a bit of over-correction.