nimh wrote:One's courtesy of Fox News. Now Fox is partisan, all that. But that doesnt explain why Bush does so much better in the Fox poll this week than he did in the Fox poll two weeks ago:
Bush 47 (+5)
Kerry 40 (-2)
Nader 3
Fox didnt just release this very Bush-friendly (even to its own standards) national poll, but also a batch of state polls.
In Pennsylvania, where the last three three-way polls (Quinnipiac, SurveyUSA and Rasmussen) all showed a 1% lead (twice for Kerry, once for Bush), a new Fox poll shows a 5% lead for Bush.
In Michigan, where the last three polls (Epic/MRA, Rasmussen and Mitchell) showed Kerry with a lead of 2-6%, Fox now has Bush with a 2% lead.
In Ohio, where the last two polls (ARG and LATimes) had Kerry with a lead of 3-6% and before that, Rasmussen had Bush in the lead by 2%, Fox now has Bush with a 4% lead.
In Florida, where the last few polls have been erratic (ARG gave Kerry a 1% lead, SurveyUSA gave Bush one of 7% and Rasmussen had them equal), Fox has Bush in the lead by a whopping 10%.
Now, eh ...
Does anyone else feel that the above list says more about Fox than about Bush?
But there might yet be a rational explanation - more rational than Bushie Fox honchos gathering in the attic to doctor polls, that is. Note that the Florida poll, for example, had Bush in the lead by 48% to 38% (Kerry polling 5-9% less than in other polls). That must be an awful lot of undecideds. More even in Michigan: Bush 42%, Kerry 40%.
OK, so thats the obvious bit that
Dales' blog pointed out. The Fox polls feature more undecideds because probably they push for an answer less (eg by not asking, "if you're not sure, who do you
lean to", for example). And considering Bush's support is "harder" while Kerry relies more on "leaners", that will skew results Bush's way. Without any doctoring, though its anyone's guess whether the choice of methodology might have been inspired by the expected effect on the outcome ... :wink:
Interesting stuff, this. Lot to learn <nods>