(This article was originally published by Politico.com on November 19, 2007)
As recent reports about “push polling” on Republican presidential candidate Mitt Romney’s religious affiliation make their way through the newspapers and the blogosphere, I thought it might be helpful to provide a brief pollster’s guide to what constitutes a push poll and what doesn’t.
Push polls are not polls, in that they don’t really seek people’s opinions. In fact, the purveyors of push polls not only don’t really care how the respondents answer the questions they ask, they usually don’t even record them. Push pollers care only about conveying negative information about a candidate under the guise of conducting a public opinion poll.
Push polls are generally very short — no longer than three or four questions, usually lasting less than one minute on the phone.
Push polls contact thousands of voters with their negative message. They don’t “sample” public opinion; they try to change it.
Push polls generally occur very close to Election Day, to make it more difficult to track down the initiator of the push polls.
Public opinion polls sometimes ask message questions that convey positive or negative information about candidates, even about their own candidate. Campaigns then use this information to determine which messages have the most traction with voters.
Testing negatives about candidates on a public opinion poll doesn’t make that instrument a push poll.
Public opinion polls can be as long as 20 or 25 minutes, or as short as five or six minutes. And during such polls, voters are asked demographic information, such as their age, education or partisan affiliation. That’s not the case with push polls.
Public opinion polls scientifically sample voters in a specific constituency, such as a state, county or congressional district. These polls usually include between 300 and 800 respondents.
Public opinion polls that test campaign messages are usually fielded days or weeks prior to the main media crush in a campaign, when candidates are going back and forth with TV ads and mailings.
Campaigns need to complete their research with sufficient time to digest the results and take action on what they’ve learned.
So if you think you’ve been push polled, ask yourself these three easy questions:
• Was I on the phone for more than three or four minutes?
• Did the caller ask me my age or party affiliation?
• Did the caller ask me more than five or six questions?
If the answer to these questions is yes, then in all likelihood you have not been push polled. Just polled.
Get over it.