Skip to main content

View Diary: Daily Kos Elections Polling Wrap: Did Mitt Romney really fumble away the election yesterday? (254 comments)

Comment Preferences

  •  Rasmussen re-weights for party id each month (0+ / 0-)

    That means that the last time he calibrated his weighting was immediately after the GOP convention when GOP enthusiasm as at an all time high, and Dems, who normally don't really start focusing until the Dem Convention, were not fully jazzed up.

    I would bet that Obama is leading by 3 in Rasmussens sample on October 4-6.

    People wonder why Rasmussen tends to swing into accuracy just in time for election day? This is why. His party weighting is most accurate at the beginning of November.

    If he didn't weight for party id at all, like most pollsters, he'd be accurate throughout the year.

    •  The monthly re-weighting thing (0+ / 0-)

      would seem to produce artificial bounces and dips. Notice there was a bounce for Romney in Rasmussen after his convention, right when Ras applied the +4 R sample.

      •  not sure if he's ever used the August numbers (0+ / 0-)

        According to the Argo journal's reports it's "only" a 2.2 Republican advantage in his tracker samples; should be more like +5 if he was using his August partisan trends numbers. Maybe even Scotty thinks that might be taking it too far.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site