Dept of Lousy Math and Lousier Constitutional Law
June 22, 2009 5:49 pm ET by Jamison Foser
Newsbusters' Noel Sheppard complains about last week's CBS/New York Times poll:
Realizing that Barack Obama's healthcare initiative has hit some roadblocks in Congress, the good folks at CBS News and the New York Times figured they'd help it along by creating a new poll on the subject that WAY oversampled people who voted for Obama.
...
As can be plainly seen on page 7 of the poll's data, only 73 percent of respondents divulged who they voted for last November. 48 percent said Obama, 25 percent McCain.
What this means is this poll surveyed 66 percent Obama supporters versus 34 percent McCain.
Uh ... no. What this means is that 48 percent of respondents say they voted for Obama, and 25 percent say they voted for McCain, and 27 percent either say they didn't vote, say they voted for someone else, or refuse to say for whom they voted. You can't just wish away those 27 percent and pretend that the poll "surveyed 66 percent Obama supporters versus 34 percent McCain."
And while we're on the topic, it's a pretty widely-known fact of polling that questions that ask who respondents voted for in the last election tend to overstate the vote for the winner, so Sheppard's conclusion that the poll "WAY oversampled" Obama voters isn't really supported by the evidence he provides.
And, as Eric Boehlert noted earlier, "the Times sampling in terms of party affiliation was in line with years' worth of previous polls." Not to mention the fact that the poll found that, by an 11-point plurality (50 to 39), Republicans favor a "government administered health insurance plan like Medicare that would compete with private health insurance plans." So the Times poll could have consisted only of Republicans, and it still would have shown strong support for a public plan.
Sheppard's conclusion would seem to apply better to Sheppard himself than to the New York Times (were it not for the unconstitutionality and general stupidity of applying it to anyone):
Honestly, stuff like this should be illegal and any news organization found doing it should be significantly fined.
In any industry you could name, such deception of the public would meet with very serious consequences.
Why are so-called news outlets allowed to get away with such obvious deceit with total impunity?
Actually, that pretty nicely sums up the conservative media critics' view of journalism: They think it should be illegal for news organizations to do things they don't like (even when their unhappiness is based on a complete lack of understanding of polling and basic math) and the journalists involved should be fined.
In other words, conservative media critics like Sheppard don't believe in independent media. They don't believe in freedom of the press. So why on earth should any journalist ever take anything they say seriously?
Another conclusion is that respondents misreported how they voted. By my caluclations, McCain voters do not appear to be underrepresented, but non-voters were seemingly underrepresented by the same amount as Obama voters were seemingly overrepresented. It's possible that about half of those polled who didn't vote said they voted for Obama.
Eric notes that this is a known phenomenon, people saying they voted for the winner when they didn't. I bet that in June 2001 more people said they voted Bush. Now they probably wouldn't. It would be interesting to see a tracking poll of how people say they voted.
Of course it doesn't really matter. If someone wants to believe that CBS and the Times rigged the poll. then he's goint to believe it, even if he can't explain why they published the so-called evidence of their fraud.
As Eric notes, half of all Republicans want a public option. The wingnuts can't win this debate, only deflect.
They're really saying who they WISHED the voted for, so the difference between exits polling done at the time and the same pole conducted a year later might be a decent measure of how he's doing! At least, relative to the other option. This could be somewhat more pragmatic than approval rating, becasue even a low approval rating doesn't matter unless there's a strong opposition candidate - a BETTER CHOICE. So even if Obama's approval was at 40%, if people still think he's better than McCain would have been he's still doing fine. (And I still think that McCain (with Palin) could have beaten Hillary Clinton, for example.)
Since they sued to be allowed to lie...
http://www.cuttingedge.org/news_updates/na784.htm