      |
|
|
 |
|
« Back to Polling on Trial main page    Transcript for: Polling on TrialTHINK TANK WITH BEN WATTENBERG #1233 Polling on TrialFEED DATE: DECEMBER 9,2004 Mike Traugott and Joan Konner Opening Billboard: Funding for this program is provided by... (Pfizer) At Pfizer, we’re spending over five billion dollars looking for the cures of the future. We have 12,000 scientists and health experts who firmly believe the only thing incurable is our passion. Pfizer, life is our life’s work. Additional funding is provided by the Lynde and Harry Bradley Foundation, the John M. Olin Foundation, the Bernard and Irene Schwartz Foundation, and the Smith Richardson Foundation. Hello, I’m Ben Wattenberg… Survey research has been used to gauge public opinion in American elections since the 1930’s, and since that time its usefulness has been the subject of much debate. But because of close elections in 2000 and 2004, the arguments over polling have intensified. Some trumpet the usefulness of polls to help journalists, voters, and researchers, while others argue that surveys, especially exit polls, are inaccurate and misleading. What is the truth about polling? Do candidates overuse polls to shape their messages? And does this help or hinder our democracy? To find out, Think Tank is joined by… Mike Traugott, Senior Research Scientist with the Center for Political Studies at the University of Michigan and co-author of The Voter’s Guide to Election Polls And… Joan Konner, a veteran, award-winning television documentary producer and Professor and Dean Emerita of the Graduate School of Journalism at Columbia University The Topic Before the House: Polling on Trial, This Week on Think Tank. MR. WATTENBERG: Mike, Joan – welcome to Think Tank. Let’s jump right into it. Mike, you are an expert on polling and you generally have a positive view of that phenomenon. Why? MR. TRAUGOTT: I think Ben, that well-conducted and well-reported polls give citizens an independent voice in how the campaign is going, separate from the – what the strategists or the spin-miesters might be saying about what the public thinks. MR. WATTENBERG: Okay. Joan, you and I were on a commission that CNN appointed together with – it was just three of us – together with the Pulitzer Prizewinning journalist, Jim Risser. And the opening paragraph, which as I recall, you wrote, said this – or part of it said this: “On Election Day, 2000, television news organizations staged a collective drag race on the crowded highway of democracy, recklessly endangering the electoral process, the political life of the country and their own credibility.” You still stand by that? MS. CONNOR: I certainly stand by it and the both of you seconded that motion which is why it’s a collaborative report. MR. WATTENBERG: Right. MS. CONNOR: Even the colorful language that we used. MR. WATTENBERG: That was your language... drag race. I really liked that. That was good stuff. MS. CONNOR: I do think we accomplished something in that report. I noticed a considerable difference... MR. WATTENBERG: That the networks this time... MS. CONNOR: ...in the coverage this time. MR. WATTENBERG: ...were more cautious. MS. CONNOR: The were more cautious and I recall the closing paragraph in which we said that if our recommendations were adopted, it might slow down the process of reporting and that would be a good thing. and I think that happened this time. MR. WATTENBERG: Are the political poll – are the villains in this operation, insofar as there are villains, is it the journalists; is it the polling companies? How do you sort that out? MR. TRAUGOTT: Well, if we’re talking about exit polls now, I think that there’s a very critical management issue about when data and what kind of data are being released. There’s no question that the exit polls are a very important part of the American electoral process that provide... MR. WATTENBERG: Explain for our viewers the difference between an exit poll and a regular poll. Just let’s get the underbrush out of the way. MR. TRAUGOTT: Okay. A pre-election poll is a survey taken of a national cross-section, typically on the telephone; it lasts two or three days. The typical sample size... MR. WATTENBERG: You say it proceeds over the course of two or three days to get the full number of respondents? MR. TRAUGOTT: Yes. MR. WATTENBERG: Right. MR. TRAUGOTT: So they conduct a sample of about a thousand people, plus or minus a little bit. And exit poll is a survey... MR. WATTENBERG: And that gives you a margin of error so-called plus/minus three percent, something like that? MR. TRAUGOTT: For that size sample, yes it would. An exit poll is a survey of people as they leave the polling places, although that technique is being modified now because of other aspects of voting. But it’s a survey done to represent the state and to be able to project the outcome of a particular state race, or it’s done on a national basis. So the exit polls are conducted by the National Election Pool involved a little over 13,500 interviews. MR. WATTENBERG: All the networks pool their money and there’s one exit poll on Election Day, is that right? MR. TRAUGOTT: Yes. This is a cost-saving operation that was incorporated as VRS, Voter Research Service; the VNS, Voter News Service and then when that folded after the 2002 election and the NEP, or National Election Pool. MR. WATTENBERG: And the networks, I mean, you’re a television veteran – they seem to have this macho thing that if you call it first you’re sort – got big muscles and in point of fact I don’t think one person in a thousand know who’s calling first. I mean it’s sort of a silly – silliness. The root of this exit poll problem as we saw it, and you correct me if I’m wrong, was that they released these so-called waves in the course of the day. In other words they will tell their clients - the networks, but also a lot of newspapers, associated press, they’ll say “well, as of noon in terms of the people we’ve interviewed, here are what the results are” and “as of three p.m., here’s what the results are”. And of course you can’t keep a secret in the political world. MS. CONNOR: They call it leaky. MR. WATTENBERG: Leaky, yes. But it gets out immediately and for anyone who’s not tone deaf you can tell immediately what they say. I mean, people at the Kerry headquarters are jubilant or whatever, without saying “here’s what the first wave said”, I mean is this the root of or part of that problem? MS. CONNOR: It was this year. The first report I think was at two o’clock and it not only got out to the news organizations, but it got out to the bloggers and picked up by, you know, various bloggers and it was out there that Kerry was ahead. MR. WATTENBERG: Well you have a whole situation now where there are some people in America who believe that Kerry really won Ohio and that if he won Ohio he would have won the election. It’s sort of the same deal as you had in – although no prominent democrat official – democratic official has said anything like that. To what extent do you think that that early calling of the waves that people thought Kerry was ahead might have contributed to this always present conspiracy mindedness of the American public, which is now intensified by the so-called blogosphere? MR. TRAUGOTT: Yes. I think it’s a combination of two things. It’s some small minority of the population that is prone to these conspiracy theories, but also the speed of information transmission across the web and the internet. And there is some relationship between the speed of information dissemination and the support for these theories, but there’s really no suggestion in any of the election administration officials, for example, that there’s a reason not to believe that the vote is accurate – the vote count is accurate. MS. CONNOR: ... You know, that’s sort of that rational approach to this. The fact is, on exit polling day there was a two-and-a-half-hour blackout, I think, because of machinery. And then there was a hang-up at the eleven o’clock projection. Something went wrong with the... MS. CONNOR: And people hear that and they know that the machinery is not foolproof in the polling; why should it be foolproof in the new machinery for voting? There’s a perfectly good reason to question it. And in fact it does not hurt to have that challenge. I don’t think – I don’t think anything’s going to come of it, but Ralph Nader as a consumer advocate can leave a positive lasting legacy if in fact he finds out if votes were counted. And particularly in Ohio and Florida where you had secretaries of state who were purging legitimate voters from the roles, naturally people are going to be concerned about that. And you know, the rational just push it aside that – and also the absentee ballots were not counted at the point of which Kerry conceded. The numbers I know added up to the fact that he couldn’t win. But why – why dismiss a challenge out of hand? Machinery misfires. MR. WATTENBERG: Let’s make a distinction now, so again that our jurors understand it, between the exit polling, which is what people are very critical of, and public opinion polling generally leading up to the election. The exit polling is done on the day of election; the regular polls go on throughout the election year and indeed throughout the year having nothing to do with election. Just – I mean there is a lot commercial polling going in. MS. CONNOR: But there are these tracking polls and I’m critical of those, too. If you recall when we were reporting, the correspondents themselves were very questioning of putting these polls on and I loved your line. You kept saying “that kind of poll creates news; it doesn’t cover news”. And you were absolutely right. MR. WATTENBERG: And what is a tracking poll, as opposed to...? MS. CONNOR: Keeping up to date about who’s ahead; who’s behind. You know, the horse race. MR. WATTENBERG: In other words, instead of taking all 1,000 respondents at the same time you’ll take 300 one night, 333 the second? MR. TRAUGOTT: For a long period of time, and they’ll usually construct three-day moving averages of the results of these. The tracking polls, I think, generally are very problematical because they feed the worst instinct of journalists... MR. WATTENBERG: Which is to always give you a piece of news. MR. TRAUGOTT: The horserace emphasis on who’s ahead and who’s behind. But I believe that one thing that we observed in this past campaign was actually a reduced number of tracking polls. They didn’t start as soon and they weren’t run as frequently. In the mix of things there seemed to be a shift in the – in the coverage to more emphasis on battleground states – the so-called red and blue states – and so there were more statewide polls that were conducted by Gallup or by the news organizations themselves and essentially I suppose because of an interest in balancing the cost of polling across the campaign the reduced the number of tracking polls. MS. CONNOR: But that polling was reported every day and you don’t now whether it influences the voter or not. that you don’t know. But it does influence reporters and it influences pollsters. And it also, as far as the news organization’s going – are affected – we used to get reporting. We don’t get reporting now. What are people saying? MR. WATTENBERG: Mike, you’ve talked about and written about the so-called indirect effect of polling on both the voters and on the elites in America. Can you explain that thesis? MR. TRAUGOTT: Well, the direct affects have to do with people actually seeing results of polls and responding to them in terms of their preference. But at different... MR. WATTENBERG: I mean give me an example of that. MR. TRAUGOTT: Well that’s what we sometimes talk about “underdog effects”...people see a candidate behind and they think they ought to support that candidate or a bandwagon effect; they see one candidate... MR. WATTENBERG: Yes, I mean there’s sort of two schools of thought. There’s the bandwagon and then there’s the wagon-band which is that people will go for an underdog. MR. TRAUGOTT: Right. And we know that both of these actually occur, although they’re very small. But – but early in the campaign, for example during the primaries and caucuses, people make investment decisions about writing checks to candidates, you know, when they’re trying to secure... MR. WATTENBERG: The so-called “money primary”. MR. TRAUGOTT: Right. Matching – matching funds. And these are investment decisions that they make as much as ideological decisions, so they don’t want to invest in losers and candidates who start to fall behind or who are behind see their funding dry up. We also know that in both the primary... MR. WATTENBERG: And that would be a function of what those early polls were all about, and the early polls would be largely a question of... MR. TRAUGOTT: Name recognition. MR. WATTENBERG: Name recognition. I mean, Hillary Clinton in the early polls is going to do well because everybody knows who she is. I mean, how she would do in a general election or in the later primaries, I, you know, haven’t that I’d even venture a guess, but it really – it had – does not have much to do with anything. MR. TRAUGOTT: No. And then the other major indirect affect is that the standing of the candidates in the polls, both in the primaries and in the general election, has a lot do with their news coverage. There’s all kinds of studies that show that candidates who are ahead get more coverage and get better coverage and if your coverage dries up, for example, in the pre-nomination phase, you can just drop out of sight completely because the schedule is very dense and compact. MS. CONNOR: You know, one of the things that you haven’t talked about – we’re talking about all the good – the good solid citizen polls; people who are – who are even-handed and who’s business is. But there’s also the influence of the pollsters who have an agenda, who work for the different parties and... MR. WATTENBERG: Well the different candidates. MS. CONNOR: ... and have exactly the same impact that you’re talking about on the reporters, on other pollsters and on the public. MR. WATTENBERG: Right. They can just prove “well our candidate really is in the running” or just the obverse of what you were saying before, and that’s one reason that candidates don’t use other people’s polls. They say “well I’ll use my own polls” and their own polls – surprise, surprise – almost invariably show their candidates doing better than other polls. MS. CONNOR: I’ll go out on a limb and say I really think that polls are overused almost to the point of abuse, even though they were good this time; even though they were closer this time... MR. WATTENBERG: They were good this time except for the exit polling. MS. CONNOR: Except for the exit polling. Leading up to the election in general. MR. WATTENBERG: And just to repeat that again, the biggest single problem with the exit polling was the use of those early waves. MS. CONNOR: Early waves. MR. WATTENBERG: That was... What’s their hurry? Why couldn’t they wait another six hours? I mean... MS. CONNOR: Because people want to know. MR. WATTENBERG: Before polling, did candidates – I guess operated mostly on the seat of their pants – have a pretty good sense of what the voters were feeling? In other words, they might not have said “well, Bush never broke 50% in the approval ratings; he’s only at 49%”. That was a big story for awhile. But they would have had a sense, hey; Bush is not as popular as an incumbent president ought to be. I mean, there is sort of a visceral political... MR. TRAUGOTT: Well, we would have been talking about a whole ‘another kind of political system in which we had... MS. CONNOR: I’d like to say that was before my time. (LAUGHING) MR. WATTENBERG: That was before all our time. MR. TRAUGOTT: In which there were strong parties and the state chair or the county chair could contact people down at the precinct level and get intelligence reports outside of election and then on election night could get some sense of turnout and the, you know the division of the vote as well. And essentially it’s that kind of operation that the pre-election polling and the exit polling has replaced. MS. CONNOR: Reporters used to go out and talk to people and talk to state officials and talk to party officials and – and when we’d get stories about how people were feeling about certain issues. And it’s much cheaper for all the networks to get together and the AP to supply the newspapers and have one service giving everybody the same information. MR. WATTENBERG: And you guys... MS. CONNOR: But that’s dangerous, journalistically. MR. WATTENBERG: The scholars particularly are going to be chewing over this – the exit poll data and the regular data as to why people voted as they did for years. I mean, so it just seems to me that waiting the extra four hours or six hours... MR. TRAUGOTT: Well again, going back to this point that we discussed earlier about the two functions of the exit polls – the projections versus the analysis. On the analysis side there is no reason why we couldn’t wait a day or two. MR. WATTENBERG: Right. MR. TRAUGOTT: And in fact from a statistical perspective, if you’re in an estimation situation, you want to estimate how many votes or the popularity of each candidate and be better off to have more than one estimate, more than exit poll than only a single. MR. WATTENBERG: Well, Bill Schneider of CNN, I think when we interviewed him on our little commission, or maybe at some other point, said “one exit poll is worse than none” because then you really – you have no sense of saying well, the Latinos might have voted 44% or 34% or 37%; you’re locked into this one number and it may or may not be right. I mean, after all it’s a sample and you know, a lot of things as we pointed out, there’s a lot of things going on. MS. CONNOR: Single sources – it is a cardinal rule of solid journalistic reporting to have two sources for any fact, and that’s not the case in this system. MR. WATTENBERG: I mean you would think – you would think... MS. CONNOR: Now it does – I mean the fact that they pool allows them to cover the entire country and no partic... no one news organization can afford to do that. And this year the two geniuses of polling, Warren Mitofsky and Joe Lenski were still the two geniuses but they split into two companies so that it had the appearance of being two companies but it wasn’t. I mean they fed into each other as they always have. Even though on paper it looked like it was a double source. MR. WATTENBERG: And most people don’t realize that the various networks, when they’re watching, are using the same data. MR. TRAUGOTT: In addition to the estimation issue, there is the question of how the actual exit poll interview itself is constructed and what kind of news judgments go into deciding which questions to ask that might explain why somebody voted Kerry or voted for Bush. And as you know, there was a question raised about the item on the exit poll survey about what was the most important reason that you voted for the candidate that you did. And they used a short term that could have covered a lot of different issues called “moral values”. MR. WATTENBERG: Moral values. Right. MR. TRAUGOTT: And that is going to frame the – a significant part of the analysis in that 2004 election. MR. WATTENBERG: And that can mean – that can mean a lot. I mean it’s being taken to mean religious values, but it could mean a lot of different things, and again, depending on how you aggregate the various data, there’s moral values; but if you put together terror in Iran that’s much bigger, or if you put together a variety of economic issues that’s much greater than that 22%. So some anonymous person writing a poll putting in moral values shaped our – a lot of the dialogue about our democracy. MS. CONNOR: It was interesting in the analysis that CNN put out a week later in which the exit polls were balanced against the other factors of counted votes, that the issues lined up exactly as had been predicted. That when it came to leadership, clarity, terrorism, moral values it was Bush; when it was the economy and jobs and education and healthcare it was Kerry. And that’s the way it turned out. MR. WATTENBERG: Okay. Let me ask just as exit question first as you... MS. CONNOR: Who did we vote for? MR. WATTENBERG: ...Joan and then Mike. Just a – and again briefly, if you would. If you were the boss, you were the queen, you were the king of running this whole situation, what would be the one or two things you would do to change it for the better? MS. CONNOR: Talking about polls or Election Day generally, I think it ought to be a holiday, 24 hours, so that people can go to the polls. And I definitely think that there should be a one-line law that says any electronic voting machine that does not allow for a recount is illegal. MR. WATTENBERG: In a federal election. MS. CONNOR: Yes, in a federal election. MR. WATTENBERG: Right. Okay. MS. CONNOR: I mean we all have printers. I don’t understand why there can’t be a paper trail. MR. WATTENBERG: Yes. Good point. MR. TRAUGOTT: I would – I would say... MR. WATTENBERG: Mike... MR. TRAUGOTT: ... two things. One about polling. We need to have more and complete fuller disclosure polling methods so that even though the public may not be very well-informed about methodology and many journalists may not be well-informed about methodology, people who are experts can dissect results and compare results and then contribute to the public dialogue about what their meaning is. On the matter of Election Day procedures, since 2000 and the passage of the Help America Vote Act of 2002, it is clear that we are now embarked upon a program moving us towards anywhere voting. We’re not there yet because of a... MR. WATTENBERG: What does that mean, “anywhere voting”? MR. TRAUGOTT: That people will be able to vote from anywhere that’s convenient for them. And that will be... MR. WATTENBERG: You approve of that? MR. TRAUGOTT: I don’t approve of that currently because there’s a variety of very strong and legitimate security concerns. But we participated in elections for many years where we went to – if we lived in the same neighborhood – we went to the same place to vote; we used the same kind of equipment. Our equipment is going to be changing continuously and dynamically into the future. Once you let technology in and evolving technology in, that means within the ability of local jurisdictions to pay, they’re going to be looking for the latest and best technology. So we’re going to be entering a period where people are going to have to - citizens are going to have to be adjusting continuously to new kinds of technology and we’re going to be seeing new kinds of technology introduced by manufacturers. MR. WATTENBERG: Okay. MR. TRAUGOTT: And we’re going to be entering an unstable period, relatively speaking... MR. WATTENBERG: Even more than we are now. MR. TRAUGOTT: Even more than we are and it’s very important to maintain confidence of the American people in the electoral system as we go through this. MR. WATTENBERG: Okay. Thank you very much Mike Traugott, Joan Connor, and thank you. Please remember to send us your comments via email. We think it helps make our program better. For Think tank I’m Ben Wattenberg. Announcer: We at Think Tank depend on your views to make our show better, please send your questions and comments to New River Media, 1219 Connecticut Ave NW, Washington, DC 20036 or email us at thinktank@pbs.org. To learn more about Think Tank, visit us online at pbs.org and please let us know where you watch Think Tank. Funding for this program is provided by... (Pfizer) At Pfizer, we’re spending over five billion dollars looking for the cures of the future. We have 12,000 scientists and health experts who firmly believe the only thing incurable is our passion. Pfizer, life is our life’s work. Additional funding is provided by the Lynde and Harry Bradley Foundation, the John M. Olin Foundation, the Bernard and Irene Schwartz Foundation, and the Smith Richardson Foundation. Back to top  Think Tank is made possible by generous support from the Smith Richardson Foundation, the Bernard and Irene Schwartz Foundation, the Lynde and Harry Bradley Foundation, the John M. Olin Foundation, the Donner Canadian Foundation, the Dodge Jones Foundation, and Pfizer, Inc. ©Copyright Think Tank. All rights reserved. 
  Web development by Bean Creative. |