Today, Monday June 29,2009, the Mitchell Daily Republic has a Bill O'Reilly column titled "Do not be deceived by shoddy polls". It ostensibly concerns a NYT poll on health insurance. The message here about O'Reilly is do not be deceived by shoddy analysis of a shoddy columnist.
I will be a bit generous and assume that O'Reilly cannot be so ignorant he unknowingly deceives, and suggest he knows and chooses to ignore some obvious facts about the data he presents.
In short, O'Reilly alledges that the NYT and CBS and others intentionally deceive when they report that there poll shows 57% of the respondents say they are willing to pay extra taxes so that all Americans will have health insurance. He notes the data below from the poll results.
37% say no, the will not pay more taxes to provide universal health care
6% claim they have no opinion. (That totals 100% of respondents apparently).
O'Reilly claims this must be a carefully selected sample because when asked whom they voted for in the 2008 presidential election, the respondents:
25% said they voted for McCain
19% said they did not vote (that totals only 92%, so not sure what was left out)
Then O'Reilly notes that the actual election results were more like
47% for McCain.
He then leaps to the conclusion that the poll sample is hugely skewed to Obama voters and that directly relates to the 57% support for increased taxes to pay for health insurance.
Without having access to all the information, two things immediately jump out. A sample of voters in which 19% did not vote may represent the general population. A "sample" of all the voters does not include the people who did not vote.
O'Reilly is comparing apples and oranges. To further muddy his "analysis", is the generally accepted fact that some part of the population will claim they voted for the winner in a previously held election whether or not they actually did.
In short, it does not take a sinister eastern liberal elite plot to have a poll turn up with support for increased taxes to pay for universal health care. It does however take either a columnist who fails to understand polls or understands them and intentionally distorts results to generate his claim that a poll is "shoddy".
In reality, Bill O'Reilly is a shoddy columnist who never lets facts stand in the way of his favorite myths in support of Republican policies or positions of the very rich executives in the insurance industry.
** Stay tuned for more on the attempts to generate FUD..fear, uncertainy, and doubt among voters who would actually benefit if the private health insurance industry disappeared--- Doug Wiken
Look at Rasmussen Reports. They show that only 32% of people are willing to pay the higher cost.
This falls in line with other polls that show marginal support for higher taxes for health care.
As far as the sample, if it was a telephone poll they should have used RDD (Random Digit Dial) or another method of scientific sampling. The 48%/25% does not seem to support a scientifically arrived at sample. Also, without knowing the sample size we wouldn't know the Recommended Allowance for Sampling Error.
Either way, 57% seems rather high compared to other polls by ACTUAL polling firms and WOULD support Bill's assertion that they did a shoddy job. I don't know of ANY polling firm that would release a poll with those sample numbers as they are.
[[[ Dakota Today Note: You make a point or two here, but they are irrelevant in relation to what O'Reilly did. I did look at the Rasmussen site which has both a good reputation for predicting actual presidential election results, but also a reputation otherwise as being a reliable part of the vast GOP Noise Machine. ]]]]]
Posted by: CaGuy | Jul 06, 2009 at 01:53 AM