Sign In | Create an Account | Welcome, . My Account | Logout | Subscribe | Submit News | Staff Contacts | All Access E-Edition | Home RSS

Attack on the pollsters

June 29, 2009 - Jim Anderson

A recent New York Times/CBS poll on health care is under attack from the right.

In the poll, 57 percent of the respondents said they were willing to pay higher taxes so that all Americans can have health insurance. Thirty-seven percent were not willing, and six percent had no opinion.

A number of critics, including Bill O’Reilly, are complaining that the poll over-sampled supporters of Barack Obama. When asked who they voted for in the 2008 election, 48 percent of the respondents said Obama, and 25 percent answered McCain. The rest, about 19 percent, did not vote.

The popular vote in the election last November was 53 percent for Obama and 46 percent for McCain.

“Apparently The New York Times skewed the polling by offering the questions to mostly Obama voters,” O’Reilly charged.

“This kind of dishonesty is not uncommon in the media,” he continued. “The Times says its poll is ‘scientific.’ Sure it is. Scientifically stacking the deck.”

O’Reilly says that as soon as the pollsters learned that most of the respondents were Obama people, they should have thrown the results out.

There’s just one major problem with that “reasoning.”

It’s not unusual, at all, for more people to claim they voted for a sitting president than actually did.

“In the 1930s, George Gallup found that Franklin Delano Roosevelt was more popular in post-election polls than he was on Election Day,” writes Christopher Beam at

“The same was true after the 2000 election, in which George W. Bush lost the popular vote. By 2004, polls showed Bush having won in a landslide,” Beam writes.

Pollsters say the main explanation is that people who didn’t vote often say that they did.

“These people tend to say they picked the winning candidate,” Beam notes.

In the Times/CBS poll, for instance, about 80 percent of respondents said they had voted. Turnout was actually about 61 percent.

Of course, no poll is perfect, and there’s a risk in reading too much into the results of any survey. There is room, certainly, to criticize the methodology of many polls, and the reporting that accompanies it.

But the claim that the Times/CBS poll is flat-out dishonest because the deck was stacked with Obama voters seems largely without merit. In fact, slinging that charge of dishonesty while ignoring the history of such polling is irresponsible in its own right.

In the future, Obama’s numbers could drop. In a 2006 Times poll, Beam notes, more people said they voted for John Kerry in 2004 than voted for Bush.

If McCain’s numbers should one day pass Obama’s, will O’Reilly again screech “stacked deck?”

Only, perhaps, if something in the results fails to please him.


Article Comments

No comments posted for this article.

Post a Comment

You must first login before you can comment.

*Your email address:
Remember my email address.


I am looking for:
News, Blogs & Events Web