[POA2]

Oh, BTW, speaking of ignoring the obvious, do you really think that Newsweek pollsters, even with all their lackluster intellectual merit and bias, are too stupid to see the problem you identified by comparing the face value of two numbers?

Last time - Go look at the history of NewsWeek polls! - Their History is awful - Period.

Newsweek does not poll for ACCURACY - they poll for spin and to create news stories - (this is their history) -

And no there is no reason to OVER Sample Republicans and UNder-sampel Dem's (or vise versa) if the goal is to predict an accurate representation of what the outcome of the election would be if held on that day! -

And that is what the "horse-race" numbers are for - To see who would win at that moment -

Now if their goal is to spin news a certain way - they sure....you under-sample and over-sample all the way to the bank - And this is exactly what NewsWeek does -

[ableChair]

And no there is no reason to OVER Sample Republicans and UNder-sampel Dem's (or vise versa) if the goal is to predict an accurate representation of what the outcome of the election would be if held on that day! -

Interesting semantics. No, but there might be a reason to call more Reps than Dems if other factors affect the results. Statisticians do this all the time. It's called normalization. Trust me, their statisticians are thoroughly familiar with how it is done and the need for it. This technique results in equal sampling, not over-under sampling. Yes, I'm sure they're biased but propagandists are far more sophisticated than you may naively think and I can assure you they would not be so stupid as to call more Reps for no reason, thus creating an error outside their own MOE and making it blatantly obvious they did so to people like yourself. They're smarter than that. They'll bias it in other, much more subtle ways. Given how obvious all this is, it's clear to me that you have an agenda to downplay Bush gains.