Thursday, June 18, 2009

Half of Americans Troubled? Not so much.

WorldNetDaily yesterday touted the results of a survey it had conducted, in an article entitled Shocker! Most Americans know of Obama eligibility questions. The surprisingly chatty and seemingly not-so-disinterested survey conductor, Phil Wenzel, had this to say about the results:

"Our polling shows that the questions surrounding Barack Obama's eligibility to serve as president clearly strike a nerve across America, probably because it is a problem that everybody understands. Every American citizen has a birth certificate, and once in a while we all have to produce them to get a drivers license or gain entrance to school. Everyone understands the simple rules – if you don't produce it, you don't get in. And while Obama did get in to the White House, nearly half the country's adults – 49 percent – are troubled by this issue and still want him to produce his official long-form birth certificate."

I bolded the last sentence because Wenzel is misstating the results of his own survey. 49% of respondents said they were troubled by the issue OR want him to produce all records, including a long-form birth certificate. "OR," not "AND." In fact, only 7.8% of respondents actually said they were "troubled."

Beyond Mr. Wenzel's misrepresentations of his own data, a cursory look at the survey itself shows that it is horribly constructed in more than one way. And it's the second question, "What is your view of those questions about Obama's eligibility to be President?"

First, as the results from the first question show, 49% of respondents said either that they weren't aware of those questions, or where unsure. So you have almost half of the survey-takers admitting a lack of familiarity with the issue. To then turn around and ask them all about their view of questions they're unfamiliar with is bizarre; it automatically means that half of your respondents just gave you an uninformed opinion.

Now if Wenzel had broken down the results for the 51% who said they WERE familiar with the questions, then that might've produced some interesting data. But he didn't. So we're stuck the possibility that a lot of people told their pollster "No, I can't say I'm familiar with those questions. But sure, I think he should release his records."

Second, Wenzel's choice of response categories is inexplicably bad for a professional pollster. For the question "What is your view of those questions about Obama's eligibility to be President?", respondents were given 7 choices:

1. I am not concerned.
2. Questions not valid.
3. Obama has met requirements.
4. Obama has answered all questions.
5. Requirements outdated - should be ignored.
6. I am troubled by these questions.
7. Obama should release all records, incl. long-form birth certificate.

WND and Wenzel trumpet the finding of the number of "troubled" Americans, but if ascertaining that was the purpose of the survey, these options are horrible. You have one response of 'Those questions trouble me,' alongside four virtually synonymous responses that are all essentially 'Those questions don't trouble me' (each one of which garners more respondents than #6). Even #5, that the Constitution's requirements should be ignored, is basically a confession of not being troubled.

And then there's option #7, which was chosen by 41.5% of respondents. If the point was to measure how troubled people are by the questions, this is a downright terrible inclusion. It doesn't fall anywhere on a spectrum of 'troubled'; respondents could be troubled and want to see the records as a result, or they might not be troubled and want to see them merely out of curiosity. Heck, respondents could have chosen #7 on the grounds that they think the questions are total bunk and that Obama should release the records in the hopes of making the conspiracy nuts shut up. The option is simply too vague and unrelated to the other options to be constructive here.

Additionally, given that half of respondents didn't know what the eligibility questions were to start with, it's impossible to say how many of the 41.5% chose this option out of sheer ignorance.

If Wenzel really wanted to include that option, he needed at least two questions, not one. The first question would ask something to the effect of "Are you concerned about whether Obama is Constitutionally eligible to be President," with options ranging from "Very much so" to "Not at all." Then there would be a second question asking "Should Obama be required to provide more proof that he is eligible?" There could even be a third, "Would you like to see Obama release additional records?"

As the old saying goes, there are lies, damned lies, and statistics. And WorldNetDaily, it seems, isn't afraid to use all three.

No comments:

Post a Comment