“Push polls.”

There is a commonality among people who don’t know anything about polling and fancy that they know a thing or two about the political process: they figure they can spot a “push poll” at thirty yards. They’re wrong.

A “push poll” is when a campaign or its surrogates telephone a significant portion of the electorate in order to alter their opinion, rather than to measure it. If a candidate discovers that he’s down by 2% among centrist voters in a district likely to have 4,000 votes cast on election day, she might telephone a couple of hundred centrists and, pretending to simply be asking a question, make a statement that would influence the vote of the receipients of that call. Wikipedia has a helpful definition of a push poll.

A recent use of push polls was by George W. Bush’s campaign in the South Carolina primary in 2000, when they telephoned voters asking, among other things, if they would be more or less likely to support John McCain’s candidacy if they knew that he’d fathered an illegitimate black child. Of course, he’d done no such thing—the sole purpose of the call was to make voters think that he had.

There has been some sturm und drang in the past few days over an alleged push poll being used against U.S. Senate candidate Jim Webb, with finger-pointing at rival Harris Miller. By a lucky coincidence, a good friend received that very survey call last Friday.

The survey asked about Webb’s negatives — “would you be more or less likely to vote for him if you knew he worked in the Regan administration?” — but it likewise asked about Miller’s negatives. My friend, a follower of politics and a one-time campaign treasurer, perceived no difference in the strength or tone of the negatives that would indicate that either Webb or Miller was being slandered. The entire call featured a battery of questions, lasting well over twenty minutes.

Polls that ask about candidates’ negatives are common. Frankly, I think any candidate who runs a major survey who doesn’t poll about his negatives and his opponent’s negatives is wasting a valuable opportunity. If I were a candidate, I’d want to know if knowing that I’d once sued Charlottesville over civil liberties violations would make voters more or less likely to support me. I’d also want to know if knowing that my opponent had once been arrested for drunk driving (or whatever) would make voters more or less likely to support her. That’s not a push poll. That’s just good sense.

Push polls are rare. They’re generally used immediately before the election, and they’re used by desperate campaigns. Spotting one is like seeing a UFO; sure, people think they see them, but they’re probably wrong. Smart people err on the side of sanity and claim only that they witnessed something unusual, and leave it at that.

Published by Waldo Jaquith

Waldo Jaquith (JAKE-with) is an open government technologist who lives near Char­lottes­­ville, VA, USA. more »

27 replies on ““Push polls.””

  1. is there a legal requirement for campaigns to disclose their polling procedure? By this, I mean are campaigns required to make public the text of the polling material? I can understand that the results should remain private, but the methodology seems like it should be in the public domain, just as we put financial contributions in the open. The best cure for push-polling and other political antics is sunshine.

    The current scenario reeks of confusion and allegation, which is confusing enough (and opens the door to plausible deniability) to allow push-polling to occur without any risk to the perpetrator.

  2. Waldo–

    Did your friend say who the call was from? Did they identify themselves? Did he or she ask? My understanding of the situation (admittedly limited) is that the issue is less whether or not it was a push poll and more that the folks running it (affirmatively) incorrectly identified themselves.

    Am I incorrect?

  3. Push polls are rare

    Huh? Somebody should let Cooper & Seacrest know. That’s just about all they do.

    Your definition of push polling is extremely restrictive, and I’d argue inaccurate. Anyone who’s ever worked for C&S knows that push polling (as defined by Wikipedia) is a very, very large chunk of their business. I never saw a poll there that was as bad as what Rove allegedly did in SC, but they’re push polls nonetheless. The questions are asinine and strategically designed to influence opinion… and anyone who would attempt to compare a C&S poll with a legitimate opinion poll is either hopelessly partisan or an idiot.

    Bottom line is that while the extreme examples are very rare, a great deal of the slightly more subtle variety goes on — on both sides — all the time, in and out of election cycles.

  4. Did your friend say who the call was from? Did they identify themselves? Did he or she ask?

    The information was volunteered by the caller, but he doesn’t remember how it was identified.

    My understanding of the situation (admittedly limited) is that the issue is less whether or not it was a push poll and more that the folks running it (affirmatively) incorrectly identified themselves.

    I just haven’t seen any good evidence that’s so but, yes, that is the accusation by some. I think the problem is that a survey that long doesn’t lend itself to easy recollection of the specifics, combined with the secondhand nature of the information. Envision this conversation:

    PERSON 1: I got a call — some kind of a survey about the two Senate candidates.
    PERSON 2: Really? Who was it from?
    PERSON 1: Well, they said something about Virginia Democrats.
    PERSON 2: Ah, so it was from the Virginia Democrats. But they say they’re not running any such poll. They must be lying.

    It’s similar to the conversations that likely yield the belief that it’s a push poll:

    PERSON 1: I got a call — some kind of a survey about the two Senate candidates. They said terrible things about Jim Webb. And he’s such a nice man.
    PERSON 2: Really? What did they say?
    PERSON 1: That he’s in favor of don’t ask don’t tell, keeps going back and forth between being a Democrat and a Republican, and he worked for Ronald Regan.
    PERSON 2: Wow, that sounds like a push poll if I’ve ever heard one.

    Somebody should let Cooper & Seacrest know. That’s just about all they do.

    Uh. Right. Cue the black helicopters.

  5. I’m coming around to the viewoint that this was just an incompetently administered poll about candidate negatives. I am still troubled by allegations that the callers misidentified themselves and that the Webb negatives seemed to be more prominently featured. But this goes to bumbling execution rather than malicious intent.

    I doubt this was a push poll, but Harris Miller’s campaign deserves a smack on the nose with a rolled-up newspaper for managing the poll and the backlash so poorly. Miller’s answer to Ben Tribbett was just plain bad politics.

  6. I, myself, claim to have been pushpolled a couple of years ago, but mostly because it was fun to do so. Don’t take away all the joy in “participating” in the political process, Waldo!

  7. Either you know little to nothing about C&S, or you’re in serious denial. Ask anybody who’s ever worked there.

  8. Okay, I just got off the phone with an acquaintance who worked for Cooper & Secrest. She said they don’t do push-polling.

    Now, what? I don’t know anyone else who’s ever worked there.

  9. I think push polling happens more than you suggest, but less than is perceived. Sometimes, a push poll will often ask whether you are undecided and then on the basis of your answer either continue to ask these questions, or simply hang up.

  10. True story. I received a call one night from a supporter. “I just got a poll about you,” she said. Then she began to recount the questions. “They said really terrible things about you,” the supporter recounted. “I think it was one of those push polls.”

    But as she described the questions, I realized that it was MY poll. I always, always poll on my own negatives, having experienced the “joy” of going through some tough campaigns. But I also poll on my opponent’s negatives, to the extent I know them. My supporter, God love her, completely overlooked the stuff about my opponent’s negatives and zeroed in on the bad stuff the poll said about me.

  11. Waldo and friends – On this matter I have a little history. In the ’93 and ’95 election cycles Cooper – Secrest push polls were a big deal in House of Delegates races. I will strive to be careful to be precise in what I say. Mr. Secrest once suggested legal action over how I charaterized his firm’s practices.

    I recognize the legitimate (in my view) use of testing negatives about your opponent IF the hypothesis is true and you are willing to go public with the charge. Back in those days the Cooper – Secrest polls failed both tests. Editorial pages all over the state, including the Daily Progress, denounced it as “sleaze polling.” Check the archives.

    What seemed to be the pattern was that an over-sample was used – something like 500 calls in a House district -when only 300 were needed. And invariably close friends, campaign workers, fellow professional workers, neighbors and family members were included in the sample. These “polls” were done in midsummer and seemed to be conducted whereever there was a competitive race.

    As for the the hypothetical questions themselves: “Would you still vote for Bobby Orrock if you knew he voted to ease regulations on day care centers to benefit his monther-in-law who runs one?” But Bobby’s mother-in-law was a retired bank teller and, in her words, “the only kids I keep are my grandchildren when their parents go to church on Wednesday night.” Obscure, I know, but a little history may be informative.

    Burned by “sleaze polling,” Republicans held a news conference in front of the Cooper – Seacrest call center in Seminole Square. At-that-time Del. Peter Way, at-that-time Sen. Ed Robb and current Del. Riley Ingram from Hopwell were participants. Riley, as is his wont, got a little excited and entered the business to express his displeasure with the employess. I was not inside but I believe he made references to the UVa honor system and called into question the callers’ integrity.

    The response by Cooper – Secret was to call the cops. Charlottesville police served on Del. Ingram a summons for distrubing the peace or some similar misdemeanor. Riley returned to Charlottesville to face the charges in District Court. I can not recall the specific disposition but it was some combination of a plea bargain, appology, etc.

    Subsequently, House Republicans tried repeatedly to pass legislation requiring the identification of the paying sponsors of polls. The rationale’ was that if polls deliver a message just like a radio ad, they should be subject to the same displosure. Peter Way tried repeatedly. Dickie Cranwell and House Democrats fought furiously the proposals.

    After Peter’s retirement, Chris Jones took up the reform challenge and it has become law.

    I have no idea what is going on in Democratic circles these days. My phone in western Henrico has yet to ring, surprise, surprise. But it sounds like deja vu all over again.

  12. I recognize the legitimate (in my view) use of testing negatives about your opponent IF the hypothesis is true and you are willing to go public with the charge.

    That sounds like a perfectly reasonable standard to apply.

    I was too young in the ’93 & ’95 elections to be aware of that kerfuffle (which I don’t doubt happened), but it certainly sounds to me as Cooper & Seacrest learned their lesson since then. I’d like to hope that, had they not, they’d be out of business by now.

    I lost my faith in Cooper & Seacrest after Alan Seacrest’s comments about Rep. Jim Moran in 2004. I don’t care what Jim Moran said—there’s no excuse for Alan Seacrest making that public. That doesn’t mean that their polling is bad, of course, just that I’m not sure I’d entrust any secrets (like internal polling data) to the company.

  13. One other thing about push-polls: they are not done by legitimate, above-board polling organizations. Legit polling organizations won’t touch them.

    Which brings me to a second point about push-polls, related to your comment that Bush’s 2000 campaign did a push-poll asking voters “if they would be more or less likely to support John McCain’s candidacy if they knew that he’d fathered an illegitimate black child”.

    As far as I’ve been able to tell — and I’ve researched this story before — there’s a lot less meat to this story than many assume to be the case. While everybody just assumes it happened as alleged, the evidence just isn’t there. There are allegations that somebody–but we don’t know who–did a push poll, calling some people–though, we’ve got no evidence of that–and telling them something about McCain, on behalf of the Bush campaign–except there’s no evidence to tie them together. As Rich Lowry noted in a column on the subject…

    McCain’s charge was based on the testimony of one 14-year-old boy. The Bush campaign released the script of the advocacy calls it was making, and the script said only, “Don’t be misled by McCain’s negative tactics.” Asked by the Los Angeles Times to provide voters who had received the smear calls, the McCain campaign unearthed only six. According to the Times, of the voters it could reach, “three described questions that, while negative, appear to have been part of a legitimate poll. Another said she heard no negative information at all.”

    So, while “everybody knows” it happened, there doesn’t seem to be any beginning to the chain of accusations. Just a lot of people swearing it happened, because other people swear it happened.

    In all likelihood, if it did happen — and it’s hard to believe there would be a push-poll of such limited scope that it would leave no witnesses — it would have been some small group of unaffiliated cranks acting on their own, rather than a campaign activity.

  14. I’m not surprised, Harry. Should’ve warned you that most of the employees are very liberal, and don’t consider the scripts they use to be push polling. My mistake. FYI, typical questions go very much like this:

    “Do you approve of President Bush’s spying on American citizens?”

    “Does Jerry Kilgore’s effeminate speech affect which way you’ll vote?”

    “If the Republicans block Gov. Kaine’s transportation initiative, how do you think it will affect traffic in your area?”

    And so on, and so on, and so on.

    But nah, that’s not push polling.

  15. I’m not sure you’re playing fair, I.Publius. First, you say, “ask anybody who’s ever worked there.” Okay, so I asked somebody who worked there. Then, you say that you’re “not surprised” that the employee would deny push-polling because “most of the employees are very liberal”.

    I don’t like this game.

  16. Jon said, As far as I’ve been able to tell — and I’ve researched this story before — there’s a lot less meat to this story than many assume to be the case. While everybody just assumes it happened as alleged, the evidence just isn’t there.

    It is a shame that we don’t have transparency on polling methodology. I appreciate you shedding some light on what appears to be an urban legend, but it seems to me that we would have less problems if polling scripts were in the public domain. Such a reform might raise the ethical standards for the practice.

  17. In the three-and-a-half years I worked for Cooper & Secrest, I never gave a survey which I’d characterize as a “push poll.”

    Surveyers were told to refuse, if asked by the respondant, to identify who had written the survey, which makes sense; people answer really differently depending on who they think they’re talking to — the idea is that if they think they’re speaking to some nebulous, objective, nonpartisan entity, they’re more likely to be honest.

    Of course, there is no nebulous objective nonpartisan polling company– specific political parties are the ones who do the polls, because they’re the ones who care if they’re winning or losing an election; they care enough to hire a non-partisan survey company like Cooper & Secrest to find out.

    C&S employees were also frequently told to give a pseudonym as the name of the polling company — for instance, instead of identifying themselves as calling from C&S, saying “Hi this is _______ from Southeast Research Associates” which at first seems sort of duplicitous, but which also makes sense when you realize those being polled in the southwest are more likely to be candid when they think they’re talking to someone in their demographic then when they realize they’re talking to somebody in a different time-zone. Also by using the company’s name less helps to cut down on sticky lawsuit-type situations from annoyed respondents… the company is calls literally tens of thousands of numbers a day, and i assure that a fair number of those called have both a short temper and very fuzzy understanding of the laws regarding unsolicted telephone calls. I figure it’s just easier for them not to be bogged down in dismissing hundreds of legally unfounded lawsuits all the time. We were also told never to throw a survey or a sample point (list of phone numbers) in the trash– apparently they’ve had locals actually sort through their trash before and it’s a legal liability for them. I’m also under contract to not tell you stuff like this, although at this point it’s far too late for them to fire me.

    Now, I have no idea if C&S actually belongs to an association of researchers in the southwest– i’m guessing not. But I can tell you that we sometimes called areas with local laws about unsolicited phone calls, and in those situations we were told specifically that we should identify ourselves as calling from “Cooper and Secrest Research Associates.” Other times we were requred to identify the political party who had conducted the survey, if asked… but only after the survey was completed. I seriously doubt that cooper and secrest has any interest in either breaking the law or being politically immoral– they just want to get some surveys done, and different polling areas have different laws about how you’re allowed to do that.

    On a typical political survey (i’d guess C&S does 98% political surveys, maybe 2% or less industry-based economic surveys), there are a dozen or so questions which help to identify the respondant politically, and then a few yes-or-no, good-or-bad type questions about specific cantidates, followed by another dozen demographic questions. by and large, the surveys I worked on made a clumsy and incoherent attempt to be objective, and some were more effective than others. about half the surveys would have an equal number of good/bad points about each cantidate, and about half would be weighted towards a specific cantidate. Again, you have to consider the motivations of the survey’s authors — nobody wants to accidentally negatively push-poll themselves. they want to make themselves look good, even while they pretend to be objective. the more intelligent, articulate, and informed survey respondants (which, i can assure you, are a rare beast indeed) were usually able to tell who had written the survey and we both would do sort of a verbal nudge-and-wink on the topic of “i’m not allowed to disclose that information as it might affect your answers.” very, very rarely did i list a pro- or con- statement about a cantidate that the respondant had not already heard discussed to death in the local news.

    that said, these surveys were also incredibly poorly written. to start with, they were all 10th-generation faxed copies that were nearly illegible… in addition to which they were clumsily worded and had generous spelling and grammatical errors. some of them were particuarly bad, in the sense that they were so long (8 to 10 pages) that nobody would ever finish them, or that they were just filled with empty poltical blather or cartoonishly self-interested slogans, making no attempt to present themselves as something the average voter could identify with and take seriously.

  18. they care enough to hire a non-partisan survey company like Cooper & Secrest

    Heh heh. Non-partisan. James made a funny. Heck, even those who won’t admit that they’re push pollsters will at least refer to C&S as the “Democratic polling firm” that it is. Talk about denial. You sound as if you worked for Zogby or Gallup or something.

    Three and a half years?!!??! Good lord, man, how did you do it? I thought five months was living hell. And I was embarrassed to have worked there that long, seeing as they usually hired any deadbeat who knocked on the door and didn’t have liquor on his breath. Heck, most of the people I knew there could barely read.

  19. Perhaps Waldo could apply to Cooper & Secrest, work a week or two, and report back firsthand with his observations. This is a busy election year and I’m sure his observations would make for a fun read!

  20. Three and a half years?!!??! Good lord, man, how did you do it?

    I should point out that this was actually about twelve months of non-consecutive work, spread out over three years. If you’re a college-age student who spends his summers in Charlottesville, but doesn’t live there year-round (because i was attending college elsewhere), it’s incredibly hard to get a job. Nobody wants to hire you for just the summer; they all want a UVa student who can keep working in the fall. Many of my friends from high school had the same dilemma when they came back to town for the summer. Your choices are basically= Cooper& Secrest, or moving furniture in 90+degree heat. C&S pays better, and lets you sit down in an air-conditioned room all day. Although sometimes I wonder if I wouldn’t have been better off moving furniture.

  21. I took a poll last year that I thought might be a push poll. I told them I was undecided between Kaine and Kilgore. They asked a series of questions that gave me the impression whoever designed the poll thought Kaine was better. They didn’t seem to be particularly well designed to persuade *me* that Kaine was better, but they had that implicit assumption.

    Then they asked which tax plan I liked better, and they explained the difference between the tax plans a couple of different ways and asked what I thought now. Each explanation made Kaine’s tax plan sound a whole lot better than Kilgore’s.

    So I thought maybe they were trying to persuade me. I explained that to the woman doing the polling, who said she was very new on the job and she didn’t understand what push-polling was, so I explained it to her etc.

    But when I thought about it, it made just as much sense for them to try to find out which explanation worked better so they’d know which to use on other people. Ideally they’d use them in the reverse order on half the people, and get a sense which one did the most good, and whether it was worth using the other one too afterward.

    It was late spring, far from the end of the campaign. Maybe they wanted to persuade me right then on the phone but it made perfect sense they wanted *more* to find out what worked so they could keep doing it .

    Push polling makes sense after you already know what works and it’s one more way to get the message out. If you’re still finding out what people want to hear, why waste your time getting your bad message out to a bunch of people one at a time?

  22. sure thing. don’t even get me started about the eccentric co-workers and insane respondees… we’d be here all night.

Comments are closed.