How much credence should you give a political poll?

Posted to: Elections News
Login or register to post comments

What you need to know about a poll

The American Association for Public Opinion Research says public opinion researchers should release the following information about their polls, at a minimum:

-- Who sponsored the survey, and who conducted it.

-- Exact wording of questions, including text of instructions or explanation

-- A definition of the population under study and a description of the sampling frame used

-- A description of the method by which the respondents were selected

-- Sample sizes, screening procedures and response rates

-- A discussion of the precision of the findings, including estimates of sampling error, and a description of any weighting or estimating

-- Which results are based on parts of the sample rather that on the total sample

-- Method, location and dates of data collection

Some polling terminology

Sample -- The group, or subset, of some larger group that is selected to participate in a poll.

Random sampling -- Any method that selects poll respondents arbitrarily and gives each possible respondent a fair chance of being selected. When used with probability sampling design, random sampling provides the basis for calculating a poll’s margin of sampling error.

Sample frame The list from which a sample is chosen that contains all of the elements in the population.

Margin of error -- A statistical measure of variation, or uncertainty, associated with any finding of a poll.

SLOP -- Self-selected opinion poll. These polls are frequently organized by broadcast media to assemble opinions by listeners or viewers. One significant problem is that there is no control over respondent selection because anyone can call in an opinion.

Approval ratings -- Closed-end poll questions used to measure the extent to which the public approves of the manner in which the president (or some other elected official) is handling current domestic and foreign policy issues facing the nation.

Source: American Association for Public Opinion Research

By Patrick Wilson
The Virginian-Pilot
© October 23, 2008

With so many polls delivered daily, it's easy to believe the ones you agree with and dismiss the ones you don't.

If it were only that simple.

People should study all the polls, not just the ones they like, said Richard Kulka, president of the American Association for Public Opinion Research and a group vice president at a survey firm, Abt Associates in Raleigh, N.C.

"If one shows your candidate leading by 15 points and the other shows your candidate leading more like 5, you shouldn't necessarily assume that the large one is correct," he said. "The extremes are probably not going to be correct most of the time."

Experts also warn about online polls. Polls that ask Web site users to vote with a click of the mouse aren't regarded by experts as good measures of public opinion.

For example, MSNBC says on its Web site that amid the Monica Lewinsky scandal in the 1990s, 73 percent of people voting in an online survey said President Clinton should leave office. But an NBC News-Wall Street Journal poll at the same time found that only 34 percent of people supported that view.

Online surveys are not scientific, but they can affect public opinion, said Joshua Behr, an associate professor of political science and geography at Old Dominion University in Norfolk.

"It's influential," said Behr, who said he doesn't like online polls. "People have interest in it and it's generating conversation."

Polls on PilotOnline.com have this disclaimer: "This survey is an unscientific sampling of visitors to this Web site."

Behr also warned poll readers about "disaffection" - when supporters of a losing candidate may refuse to even participate in a poll.

When one candidate is leading another by up to 10 percentage points in a race, Behr said, the supporters of the losing candidate may be "disaffected" enough that they don't want to participate in a poll.

"They're less likely to register their preference or even participate in the poll," Behr said.

And that, he said, could tip polls even further in the favor of the frontrunner without truly reflecting how people will vote.

Patrick Wilson, (757) 446-2957, patrick.wilson@pilotonline.com

COMMENTS ADVISORY: Users are solely responsible for opinions they post here; comments do not reflect the views of The Virginian-Pilot or its websites. Users must follow agreed-upon rules: Be civil, be clean, be on topic; don't attack private individuals, other users or classes of people. Read the full rules here.
- Comments are automatically checked for inappropriate language, but readers might find some comments offensive or inaccurate. If you believe a comment violates our rules, click the report violation link below it.

WHAT IS THE POWELL DOCTRINE?

Submitted by RJS on Fri, 10/24/2008 at 7:29 pm.

Gen. Powell was instrumental in minimizing American casualties and the bill for the American tax payer. The Powell Doctrine does not just require the support of the American people. Kill the enemy quick and get out, that's what we did in Desert Storm. Bush Jr. refused to listen to his own dad and Powell, who were able to get the support of our allies, make them pay the bill and avoid overstretching our military.

Sen. McCain himself said he would like Warren Buffett as his Treasury Chief. One of the few things in 2008, that I agree with him on. Buffett is of the best choices, and no comparison with McCain's own economic advisors who told me to stop whining. Buffett supports paying more taxes himself and giving a break to Veterans and the middle class this time around, while Limbaugh, Hannity, and McCain’s lobbyists are fighting their tax hike very hard by trying to call me a socialist.

Peter-

Submitted by Craig Paskewic on Thu, 10/23/2008 at 3:58 pm.

Concur polling is big business and big money. In my mind, their time and money tends to be wasted.
What the pollsters do not reflect is how they obtained their results, nor do they justify their findings. Ergo, we have to ask the leading questions.
If these polls were as accurate as they depict themselves to be, the results would be identical, or, nearly identical.

Polls

Submitted by kensgarbagecan on Thu, 10/23/2008 at 2:41 pm.

The media has been so biased during this election that I don't put much weight in the polls.

Another thing you should always ask about a political poll

Submitted by cFrederick on Thu, 10/23/2008 at 12:11 pm.

The reason for disagreement among otherwise similar polls is the way different researchers determine who is actually likely to vote. There are surveys that talk to anyone. But most savvy pollsters have ways of predicting who they think will actually show up on election day, and who won't. Obviously, those who are likely to show up are given greater credence because they're the people who'll determine the outcome of the public election.

The standard is to prefer who have actually voted before. But this misrepresents the possible outcome of elections like this one, where there are likely to be lots of new voters. Others have more complex criteria for sorting people out. There's nothing fishy or unethical about this. There's just no magic 100% completely sure way of doing it. So as someone suggested earlier, it's probably a good idea to triangulate different reliable surveys.

Everyone try to remember

Submitted by Confused on Thu, 10/23/2008 at 11:13 am.

First of all, the "race" is between McCain and Obama not Palin and Obama.
Second, Colin Powell was instrumental in convincing everyone we needed to invade Iraq and there were "weapons of mass distruction".
Third, Warren Buffett is one of the crooks from Wall Street payouts - he definately "understands" the economy and knows how to work it to his benefit while the rest of us loose our shirts.
And lastly, the Polls mean nothing, just as our votes will ultimately mean nothing on November 4th. The final decision on who will become our next President will be made on December 15th when the Electoral Delegates cast their votes. A candidate can win the popular vote and still lose the election.

Polls Count

Submitted by Peter_Langlands on Thu, 10/23/2008 at 10:08 am.

Polling is big business. Plenty of money is spent and earned on polling. The process is statistically sophisticated and points to probable results with certain assumptions. The key is to look at several advertised poles. Right now the polling numbers all say Senator Obama will be our next president.