CBS Slants Bush Poll in Favor of Democrats
Posted by Greg Sheffield on February 28, 2006 - 08:57. In its classic "fair and balanced" tradition, CBS slanted in favor of Democrats its poll that found Bush has a 34 percent approval rating and a 59 percent disapproval rating, an all-time high for a CBS poll.
On the bottom of the PDF version of the poll (page 18) it says how many Democrats versus Republicans were contacted.
"Total Republicans" contacted: 272 unweighted and 289 weighted.
"Total Democrats" contacted: 409 unweighted and 381 weighted.
"Total Independents" contacted: 337 unweighted and 348 weighted. Brent Baker also noted how CBS failed to highlight a key portion of its poll on the Feb. 27 "CBS Evening News." 66 percent of respondents thought the media devoted "too much time" to Cheney's hunting accident.
[/align][/align]
I'm not surprised. I think when this happens, the public should backlash big time so CBS loses serious money for deceiving the people to try and foist its own political agenda on America! The problem is, most of the country won't hear about how the poll was rigged, they'll only hear the low approval rating!
They put out BS deception about the port deal, then quickly poll democrats before the facts come to light! Unbiased media? Yeah, right!
No big suprise here! I think this happens with every poll. I could take a poll right now of 300 people, and Bush's approval rating would be 100%. That's exactly why I don't trust or pay attention to polls or ratings- they can say whatever you want them to say.
__________________
Member of Team "REAPERS OF THE HARVEST"
A good reason we should all have a working understanding of how sampling, polling, and the statistics related to those activities work. Most Americans have no idea how bias can affect polls. Half the polls I receive are leading.
In this case, "weighting" means that they take a certain numbers and then adjust them according to the U.S. Census Bureau data to be more accurage.
In this case, it is a phone survey. Pollsters know that they are more likely to contact older people who tend via phone because they are home more, etc. So you have raw data, then they adjust it accordingly to match the demographics...an unperfect system, but short of polling everyone in the U.S., it is meant to be representative.
They also adjust per voter registration...i.e, if they pulled random numbers and happen to get more democrats in one data set, they "adjust" to be representative of U.S. population trends. All pollsters have their own "weighting" system.
Is the poll skewed? Maybe, it could also be random error or just chance. Most polls have a + or - 4% error. That could mean that they believe the true number is really 37% or even 31% -- it is range.
If you go to pollingreport.com you will see a summary of all of the polls for Bush over the last five years. His support has been steadily declining, and an average of several polls had him about 40%. The CBS poll is the first poll out since the Dubai deal, and the brink of civil war in Iraq, so the best bet to get a true idea would be to wait a week or two, see what other polls are saying from the website I just gave you, and then take an average of six or seven using different weightings and interview types as well as questions (sometimes the wording of a question can sway results).
Personally, I doubt he dropped six points, but I bet he dropped a couple at least.
i've never conducted nor had to interpret the raw data from a national poll, but 2 observations come to mind:
1.-1000 participants seems like a small sampling for a nationwide poll
2. the D/R/I breakout percentages should reflect either a) the accurate categoricalproportion of registered voters nationwide, or b) the accurate categorical proportion of supervoters (people who have voted in, say 3 of the last 4 presidentials) nationwide.
i usually only care about the latter, because if you don't vote, i don't really care what you think, and you have no impact on the outcome of an election.
the proportionreflected here (D/R/I of 37.4/28.2/34.4) is almost certainly NOT supervoters. i also suspect it doesn't reflectsimple registered voter proportions, although i could be wrong on that. i am not familiar with what the latter figure would be, for the reason stated above. i'd be curious to know the rationale behind the sampling draw, through.
I too am not an expert with polls but I have noticed (and also heard) that an average national poll will be made with approximately 700-1,000 people polled. There are a variety of things they calculate into the polls to ascertain within how many points such polls are accurate (usually within 5 points or less). From what I've read on polling, if the pollsters write the questions well, choose an appropriate spectrum of people and do things honestly, they get a very accurate result. Unfortunately for cBS, they have a reputation for shading things and not being completely honest. It will be interesting to see what #'s other polling firms come up with. From past experience, I would expect less than a 10 point difference on just about any responsible and reliable poll.
__________________
Jesus Christ--The reason for the season!
If you can read this, thank a teacher. If you can read this in English, thank a veteran.
If you're certain you know everything, there's little opportunity to learn anything.
CBS Poll false!!! (again)
In its classic "fair and balanced" tradition, CBS slanted in favor of Democrats its poll that found Bush has a 34 percent approval rating and a 59 percent disapproval rating, an all-time high for a CBS poll. 
