Barry Kay on the accuracy of polling and the presidential elections

Story Transcript

Polls and the Presidential election Pt. 1

PAUL JAY, SENIOR EDITOR: Public opinion polls—do they measure public opinion or make it? Barry Kay is a professor of political science at Wilfred Laurier University in Waterloo, Ontario. He specializes in electoral systems, public opinion polls, and lectures regularly on American politics. He heads up LISPOP, the Laurier Institute for the Study of Public Opinion and Policy. Thank you, Barry.


JAY: Polling itself is unpredictable. Polling affects elections. It’s sort of like this thing “you can’t drop a stone in the same river once [sic]”—as soon as you poll it, you’ve changed it. What should we really make of polling in this election?

KAY: Well, polling is there, and there’s sort of an obsession on the part of the media with polling, and I don’t think it can be easily ignored. Are there flaws? Absolutely, and it’s a matter of sort of knowing how to judge and what the limits are. And, frankly, I feel much more comfortable when I’ve got five or six polls in the sack, looking at them, rather than just one individually. But people should be mindful of the fact that the polling industry has changed. Getting information by telephone is so much more difficult than it used to be with cell phones now. Technology is changing, even with the traditional telephone type of poll that we’re probably only seeing that maybe one in five people that are contacted will actually agree to participate in the survey, and that frequently is after several callbacks. There are changes in that industry that are going to suggest that online polling will be the wave of the future. We’re not quite there yet.

JAY: And in the last, say, five or ten years, how accurate has polling been in terms of determining election results?

KAY: In general pretty good.

JAY: Especially talking American elections.

KAY: Yeah. Same, but Canada’s not so very different, either, unless public opinion shifts at the last moment, and that’s happened a little bit the odd time. But, generally, polling within two, three points, the margin of error, been quite good.

JAY: The recent polling on Obama and McCain has shown that Zogby, in the most exaggerated way, big shifts, and having Obama up as much as five or ten points, and now about even, which shows, according to Zogby poll, a drop. There’s a sort of a general feeling that the Obama campaign’s kind of stalled. Is this of significance? What’s polling [inaudible]?

KAY: It’s early. When I say it’s early, this campaign’s a perpetual—it’s been going on for the better part of two years, but it’s early in the sense that there’s still two months until the election day itself. Yeah, I think “stall” is the right word. Obama has not dominated the news cycle lately. I think that will start to change, certainly with the conventions that are now taking place. But I think for people that are really concerned about these things, the first presidential debate isn’t until September 26. I think for people—and there are a number of people that have heard about Obama, may be aware of his celebrity, but aren’t really familiar with him. I think the opportunity for Obama to sort of sell himself is going to take place. And for people that aren’t really keen on—I’m a political junkie, but for people that aren’t, I’m not even sure this week is going to be so important. I think it’s in the debates that some people who aren’t perhaps all that attentive to politics will begin to tune in.

JAY: Now, the role of the media in public opinion, obviously, is enormous. Right now, the television news, particularly, seems to be like a kind of echo chamber. The Republican campaign says this; the media reports on what they said. The Democratic campaign does the same. The media very rarely actually tries to take their own take on whether or not what’s being said is true or not. There’s very little deconstruction of these campaigns.

KAY: Well, a lot of that’s in the eye of the beholder, anyway. There are different constituencies. Fox News is obviously going to play to a very different kind of clientèle and a different constituency, and what they’re interested in is selling advertising and trying to hype their constituency. But there are other kinds of constituencies as well. So you’re right: there is an echo-chamber aspect to it, particularly among sort of the mainstream media. But you get, depending on the particular avenue or channel for media consumption, there are different stories out there to hear.

JAY: And how effective are the millions and millions of dollars that are spent on television advertising? They must be, or they wouldn’t keep spending so much money.

KAY: Lots of money spent very unevenly, and it’s very important in a close election. If this were an election where one candidate had a ten-, twelve-point lead, it probably wouldn’t be so critical. What’s really interesting, though, about all the money—and there are going to be not just millions but hundreds of millions of dollars spent during this campaign, but it’s going to be spent very unevenly. State like California, the biggest state in the country, there was not one dollar spent in the presidential election of 2000 or 2004 (I emphasize the presidential election; I’m not talking about congressional) because California is not a close state, and that would be true of Texas and a lot of other states. But there are some states—Ohio, Missouri, Michigan, Pennsylvania, Iowa, Wisconsin—those states, enormous amounts of money, because those states seem to be the pivot in what at the moment looks like a fairly close presidential election.

JAY: Is polling feeding a kind of superficiality of the elections? What I mean by that is everything is—the TV campaign is completely on the surface; there’s no real issues talked about in any way that challenges the facts of the campaigns. The polling reinforces how people are kind of instantly responding to these campaigns. The television media echoes the advertising campaign. I mean, we’re practically doing polling on whether you like Coke or Pepsi.

KAY: Polling focuses on the horse-race aspect to this, and it’s a matter of what’s happening now. You asked about the stall just a minute ago with regard to Obama. He had a lead of perhaps five or six points a few weeks ago. Now it’s down to one, two, three, depending. I mean, that’s based on averages. There’s polls that show even outliers to that. But that indeed I don’t think—you know, just sort of inherent in your question was the notion of should we be concerned that he’s down a couple of points from two weeks ago. I think if this was November 2 or 3, yes, that would be very important; but I think, given that we’re still in the month of August, it isn’t such a critical thing. I think a lot of things are going to happen between now and election day. But you’re right: the media focuses on polls because of the fact that it entertains and amuses their audience. (05:27) . . .

JAY: . . . (06:10) Something like 40 percent of Americans don’t vote. To a large extent—.

KAY: Even more.

JAY: Even more. To a large extent, I’m assuming that this is mostly poor Americans, although not entirely.

KAY: It’s disinterested Americans, and disinterested Americans tend to be poor Americans. But it’s not just the poor. But, yeah, you’re right: the tendencies are less educated, poorer people, less likely to feel connected with the community, less likely to think their vote makes a difference, less likely to vote.

JAY: Do they show up in polling results when we’re seeing these polls?

KAY: No. And, actually, there’s an element of artistry with regard to how you sort of factor out those people. People who aren’t interested in participating politically normally aren’t interested in participating in polls either. So some of it is sort of self-correcting. But you’re right. And among different pollsters, sometimes we see very much outlying numbers. I don’t want to pick on any one particular polling house, but some numbers don’t seem consistent with the others. And a lot of it relates to how they decide to try to factor in the notion of where the real electorate is. But in general the people that don’t vote don’t participate in polls. The problem is that some people who do vote don’t participate in polls either.

JAY: But on the whole, the polling still winds up [inaudible] to what’s happening in elections.

KAY: When you average-in the numbers, generally, at the national level, it’s not to say there can’t be outlying results or that you can’t have situations where public opinion changes at the very last minute. We’ve seen a little bit of that in the Canadian election, for example, of 2004. That happened on sort of the last night. There have been a few American elections that’s happened. But, in general, polls on average are pretty good.

JAY: Okay. Let’s move on. In the next segment of our interview, let’s talk about will the United States elect an African-American president.


Please note that TRNN transcripts are typed from a recording of the program; The Real News Network cannot guarantee their complete accuracy.

Creative Commons License

Republish our articles for free, online or in print, under a Creative Commons license.

Dr. Barry Kay is a Professor of Political Science at Wilfrid Laurier University. His research focuses on the topics of elections and public opinion. He is a past member of the Canadian National Election Study team, and recent publications pertain to electoral systems, public opinion polling, and the impact of single-issue interest groups. He has developed a model for projecting parliamentary seat distributions from popular vote or opinion polls, which is updated regularly and can be found at He is also a political analyst with Global Television, for their national election coverage.