Is that a 'push poll' on the phone, or message test?

by Mark Silva

In the wake of weekend reports of "push polling'' in New Hampshire attacking Republican Mitt Romney and promoting rival John McCain -- which both candidates have formally protested in formal complaints with the state -- the American Association for Public Opinion Research has issued a statement of guidelines calling the practice "unethical political telemarketing.''

"There are guidelines to help determine whether or not a poll is a valid survey,'' the association says. "The reminder from AAPOR comes after a round of stories described telephone calls to Iowa and New Hampshire voters as "push polls" because they included harsh information about a candidate.

"AAPOR defines a push poll as unethical political telemarketing, calls disguised as research that are designed to persuade large numbers of voters -- not to measure opinion.''

"Negative or disturbing information about a candidate does not automatically make a survey a push poll," AAPOR President Nancy Mathiowetz,'' said in an issued statement. "Message testing, when campaigns test the effectiveness of possible messages about opponents and even themselves, is very different; and it is a legitimate form of surveying.

"What's changed in the recent election cycles is that the practice of message testing, once largely invisible to the public, is now receiving a lot of scrutiny,'' she said. "Add to that a crowded, highly competitive field in the early primary and caucus states and there are going to be plenty of surveys to critique.''

AAPOR offers guidelines spell out the difference between "push polls" and message testing. For example, in message-testing surveys, the call will contain more than a few questions; the organization or call center making the calls will be identified; and the survey will include questions about the respondents' demographic characteristics. Message testing is usually based on a random sample of voters, and the number of calls will fall within the range of legitimate surveys, typically between 400 and 1,500 interviews.

"Push polls" usually ask one or only a few questions about a single candidate or a single issue; the questions are uniformly strongly negative (or sometimes uniformly positive descriptions of the candidate or issue; the organization or call center conducting the calls is not identified, or a phony name is used; and the calls are placed to large numbers of people.

""It can be tough for any one respondent to tell the difference since we identify 'push polls,' in part, by the number of people that are contacted within a relatively short time frame," Mathiowetz said. "In addition, it would be surprising to see a 'push poll' at this point in the election cycle since they are intended to persuade voters near the time of the election.

"The speed at which we are learning about these calls -- and the number of stories on the subject -- do raise interesting questions for campaigns, she said. "Campaigns have traditionally been able to conduct message testing in relative privacy. Now that's changed."