Friday, September 16, 2005

The NYT/CBS notion of Balance and Demographics Weighting


Tom Elia at The New Editor asks “Is the Latest NYT/CBS Poll Biased?”

Off the top of my head, the answer is ‘yes’. On reflection, ‘yes’. On prolonged consideration and after examining the evidence and history, ‘absolutely yes’.

Fortunately, as I have mentioned before, the funny thing about the NYT/CBS Poll, is that for all their arrogance and bias, they are pretty forward about presenting their numbers, which makes it possible to reverse-engineer the poll and determine a true state.

The first thing which jumps out at me comes up right on the first page of the poll results. Right under the title and date is the following note:


White N = 877
Black N = 211”

In other words, 877 of the respondents were white (~75%), 211 of the respondents were black (~18%), and another 79 respondents were neither black nor white (~7%), possibly meaning Hispanic, though the NYT/CBS does not say.

The Census reports that whites comprise about 82% of the population, blacks about 13%, and other races about 5%. This demonstrates a demographic mis-weighting by the poll.

Continuing, the overall support was 12% by blacks and 47% by whites, both well within historical norms for Dubya during his Presidency, especially in this generally sour poll. 49% of whites approve of how Bush has addressed Katrina, very respectable considering the absolute smear job against him by the MSM, and 18% of blacks agree he’s done well in responding to Katrina, which is higher than the black support in general for Bush. When the question is phrased in terms of the “people affected by Katrina”, white support climbs to 50% and black support to 23%, again both very good in the context presented.

I thought it interesting that the poll focused mainly on four questions, treating them very much as related to each other. The poll addressed the response to Katrina, the condition of the economy, and the War in Iraq. The poll also paid attention to the confirmation hearings on John Roberts, even though nearly half the respondents admit they have paid almost no attention to them.

Returning to demographics, the NYT/CBS poll states that 22% of the respondents were 18-29 years of age. The Roper Center shows that in the 2004 election, only 17% of voters were 29 years old or younger, so again there is an obvious imbalance. And as Elia noted in his column, the NYT/CBS Poll shows a 36% Democrat/26% Republican/28% Independent/9% Don’t Know breakdown, far off the 37%-37%-26% balance reflected by the Roper Exit Poll.

Also, and this is important, the NYT/CBS Poll admits that 20% of its respondents did not vote in the 2004 Election, and another 6% say they voted but don’t know whom they voted for!

So, at this point we can safely say that the NYT/CBS Poll is not, in fact, trustworthy, nor should they be believed in their claims. The poll oversamples blacks, Democrats, people under 30, and people who did not vote in the last election and who admit they are not paying close attention to what is going on in important events. In other words, just the sort of people who get their news and opinion from the New York Times or CBS News, instead of checking out the facts.

No comments: