Bush Bias in Rasmussen's poll analysis??

[Smash255]

Granted the Rasmussen polls have been more in Kerry's favor than the other polls since the RNC and Rasmussen stated the biggest reason for that has to do with Party id in the poills.

However one thing I have noticed when rasmussen reports the polls is a bit odd.  Both Bush & Kerry have days in which they have polled quite a bit higher than any of the surrounding days.  when kerry has a day like that Rasmussen always points it out, however when Bush has a day like that he never points it out,  For example one day over last weekend Kerry had an abnormal high day where he polled much better than the surrounding days, Rasmussen mentions it.  Now from the day by day polling I have seen in the last 4 days, Kerry was ahead 2 days by a small amount, Bush ahead by a razor thin margin 1 day, and Bush up by 5.5 another day.  Yet Rasmussen doesn't mention that Bush had a day wher e he was well above all the surrounding days.  Vorlon correct me if I'm wrong with the #'s over the last few days.  But it looks to me that everytime Kerry has an odd day in his favor Rasmussen mentions it, but when Bush has a odd day in his favor Rasmussen never mentions it

[○∙◄☻¥tπ[╪AV┼cVê└]

Well, Rasmussen is a Republican.

[The Vorlon]

Please note if you take my one day totals and average then into 3 day totals they do NOT exactly match what Rasmussen publishes (they are close)  

Each day gets weighted independantly, as do the 3 day samples, so because the weights vary from day to day, the do not exactly match up.

They are close, within a few tenths.

For what it's worth, these are the daily Bot samples....

The date on the chart is the date of publication, actualy sample was conducted the night before

[Sibboleth]

Ah... the infinate wonders of statistical noise...

[Smash255]

Please note if you take my one day totals and average then into 3 day totals they do NOT exactly match what Rasmussen publishes (they are close)  

Each day gets weighted independantly, as do the 3 day samples, so because the weights vary from day to day, the do not exactly match up.

They are close, within a few tenths.

For what it's worth, these are the daily Bot samples....

The date on the chart is the date of publication, actualy sample was conducted the night before

Vorlon thanks forr the graph prertty much shows what I had.  Anyway point remains anytime Kerry has a larger lead than the surrounding days Rasmussen mentions it as he did last week, when Bush has an average than larger lead as shown on the sample from yesterday rasmussen never mentions it as being out of thre ordinary

[The Vorlon]

Ah... the infinate wonders of statistical noise...

And because it's a tracking poll... you get two days worth of statistical noise (ok..1.414 days worth if we want to be picky) each and every day as the noise of the sample rolling off gets joined up with the noise of the sample rolling on....

And to add more noise we then reweight the entire sample too..

Ain't life grand

FWIW - I "think" Rasmussen will show Bush up 3ish tomorrow as a sample with a lot of Dems unloads, which shifts his weightings a fair bit, which adds about 1.5% to Bush... unless it doesn't...