Polls
Yet another screaming headline. Bush at an all time low approval rating. One point to consider here:
Total Republicans 344 / 313 (unweighted/weighted)
Total Democrats 433 / 453
Total Independents 464 / 475
The sample data makes no sense whatsoever. The weighting reduced an already heavily undersampled population. The sample is skewed, blatantly. I wouldn't bet the farm on a poll that was this slanted.
Just pointing that out.
UPDATE: And the folks at the Harris Poll have some data that really puts paid to this sample. Their latest electoral breakdown shows 31% Republican, 34% Democrat and 24% Independent. That's completely in line with the trend they have over the years, too. The CBS people are really being dishonest here.
UPDATE: Brainster's got more. So Does Sister Toldjah, Stop The ACLU, Newsbusters, Flopping Aces and probably a bunch more people who can actually do a bit of math. CBS, your slip is showing.
Other Links to this Post
-
Sister Toldjah — Wednesday, 10 May , 2006 @ 8:39 am
-
A Blog For All — Wednesday, 10 May , 2006 @ 8:51 pm
By Black Jack, Wednesday, 10 May , 2006 @ 12:13 pm
Polls have long since abandoned any effort to identify the current opinions of American voters. Polls are now secretly twisted to lead opinion in a predetermined direction, provide reinforcement to the faithful, and give MSM talking heads cover to bash their opponents in the GOP.
Polls can be Ignored except for their one utility. The extent to which the sample has been skewed indicates how much the pollster fears the truth.
By Gaius, Wednesday, 10 May , 2006 @ 12:24 pm
Yeah, it’s pretty bad when it’s this obvious, though. I would have failed if I’d used samples like that in Design of Experiments.