Professional organizations affiliated with polling all condemn the practice of "push polling," which is actually not polling at all, but a negative campaigning tactic that is conducted under the guise of polling. However, many voters and reporters misunderstand the term and condemn polls that the polling industry considers legitimate.
Earlier this month, for example, a report by PoliticsNH.com suggested that retired Army Gen. Wesley Clark, "may already be the victim of campaign tactics intended to damage a political career still being written." The report noted that at least three people had received calls that "appeared to be 'push polling,' a method of asking how a potential voter might react about a candidate, in this case Clark, had they known some negative, possibly untrue fact, about the candidate."
In fact, the poll appears not to be a push poll, but rather one that included a method of strategic analysis that campaign pollsters routinely use and that is not condemned by the industry. The confusion arises in part from the terminology. Push questions are considered legitimate; push polls are not.
Background and Definitions
For 30 years or more, campaign pollsters have asked questions that attempt to discover how voters will react to negative information about a candidate's opponent. The word "push" emerged as a description of what the pollsters were trying to do -- to see which information would "push" voters away from a candidate.
The article on PoliticsNH.com provides an example that quotes a respondent, Sue Mayer, who was interviewed about Clark in the so-called push poll. She told the reporter that after indicating she was favorable to a Clark candidacy, the interviewer asked her if she would still support Clark if she had known 1) that he had voted for Republicans before, and 2) that his integrity had been questioned by retired Gen. Hugh Shelton, Clark's former commanding officer.
Her description makes it clear that the question's intent was to determine whether Clark's candidacy would be hurt if more voters were aware of the information provided. Mayer also indicated that she was asked several other questions about the primary, including a horse race question and a question probing the types of candidates she would find most attractive. Given this information, I would conclude that the "push" questions about Clark were probably part of a larger legitimate public opinion poll, and not a "push poll."
The term "push poll," as currently defined, emerged sometime in the late 1980s or early 1990s. My sense is the term was previously applied to campaign polls that included a substantial number of push questions, and thus initially the term was not considered pejorative. But at some point in the past 10 to 15 years, perhaps even earlier, some political operatives decided it would be clever to call people and pretend the call was part of a poll, and then use the occasion to pass along negative information about a political candidate. Sometimes the negative information would be accurate, other times a distortion of the truth, or even an outright lie. Because the tactics included what appeared to be push questions, the overall activity was labeled a "push poll," though the intent was not to measure public opinion, but to sway it.
Given the intent of push polls, it is clear that they differ from legitimate public opinion polls in at least three ways:
- Legitimate polls interview a representative sample of respondents to measure opinion, while push polls talk with large numbers of respondents to influence their opinion.
- Legitimate polls are based on relatively small samples of respondents, usually a few hundred to 1,000 or so, while push polls usually involve calling thousands of potential voters.
- Legitimate polls with push questions will include many other questions as part of the effort to measure opinion. Push polls may include only a couple of objective questions upfront to deceive the respondent about the nature of the call, while the overwhelming focus of the push poll is to give negative information about a candidate or issue.
There are many other differences between legitimate polls and push polls, all arising out of the fact that the push poll is not really a poll, but a method of negative advertising. The National Council on Public Polls defines a push poll, in a 1995 release, as:
. . . a telemarketing technique in which telephone calls are used to canvass vast numbers of potential voters, feeding them false and damaging 'information' about a candidate under the guise of taking a poll to see how this 'information' affects voter preferences. In fact, the intent is to 'push' the voters away from one candidate and toward the opposing candidate. This is clearly political telemarketing, using innuendo and, in many cases, clearly false information to influence voters; there is no intent to conduct research. [emphasis added]