Reading Between The Polls: What Voters Should Watch
SCOTT SIMON, HOST:
As we just heard from Ari, early polling can do much to shape political campaigns, but voters who are just trying to follow the debate, polls and surveys can seem contradictory and confusing. To help us see through some of the fog of polling, we're joined now by Michael Dimock. He's the associate director for research at the Pew Research Center in Washington D.C. Thanks for being with us.
MICHAEL DIMOCK: Thank you.
SIMON: And J. Ann Selzer, president of Selzer and Company, a public opinion and research firm in Iowa. Ms. Selzer, thank you very much for being with us.
J. ANN SELZER: Happy to be here.
SIMON: And let me ask you each, compared with a decade ago, how reliable are polls today?
DIMOCK: We're finding that the polls are still doing a pretty good job. Right now, you may see a lot of this variation, a lot of spread between what one poll tells you and another, but by the time you get to the weekend or so before the election, most of the polls end up really coming together and giving a fairly clear picture about not just who the public supports, but why they're supporting one candidate or another.
SELZER: It's also fair to note that there are a lot more polling organizations that are active now, than there were even a decade ago. And they all bring with them sort of nuances in their methodology. Some are talking about the general population, that is every adult age 18 and over. Some are talking about registered voters, that's anybody who's ever bothered to sign up to vote. Some are narrowing the field to likely voters, people who say they'll definitely vote in November. They're really talking about different populations and you get different answers.
SIMON: It occurs to me, because you're of course in Iowa, Ann Selzer, the Iowa poll that's been run for years by the Des Moines Register organization has been considered highly accurate because it just has so many people responding, whereas there are national polls that reflect what, just a couple of thousand people.
DIMOCK: A lot of the polls that you see will have even fewer than a thousand people that they're reporting on, and anytime that's the case, there's a certain amount of error. Even if the poll is done extremely well, statistics will say that there's a lot of room for error around those polls. And when you look at what may look like contrasting polls, they may actually be close enough that the difference is within that margin of error.
SELZER: But I will chime in and say that a poll of 2,000 people is not statistically that much more robust than a poll of a thousand people; that once you get a thousand, you really have a pretty strong and reliable read on that universe.
SIMON: Could you explain that to us? Because I just find that incomprehensible. I mean, three times that many people went to my high school.
(SOUNDBITE OF LAUGHTER)
DIMOCK: And they are perfectly representative of the American public.
SELZER: Well, that's the trick, which is that's a cluster of people, and they may or may not be representative. So the whole science of polling is: how do you choose who you're inviting to participate in your poll, and if you do that properly, whether you talk to a thousand or 3,000, there's not going to be that much variation.
SIMON: What would you urge our informed and discriminating listeners to pay attention to the next time we report poll results?
SELZER: Well, I think this early in the campaign, and really until we get to probably October, you need to keep in mind that any measure of the horse race is really a reflection of the mood. Sort of how are people feeling about this campaign right now? There's no poll in May or June that is going to predict what's going to happen, but the advise I would give is to choose a couple of polls and follow them, because for all we're saying, gee, there are nuances, there are differences, there are different populations that people are taking a look at, all of that is controlled for if you follow the same poll over time.
SIMON: I'm sorry if this gets too personal, but do you tell friends and family members if they get a call from a pollster they should participate?
DIMOCK: Of course you should participate, and not only because it helps me personally, but the experience that I've had - I listen to a lot of interviews being conducted - and the reality of it is, most people actually like it once they get over the first 20 seconds of skepticism, that being asked your opinions about the important issues of the day is actually quite fun, and often the polls you find are challenging people to actually think concretely about things that they've probably read about and heard about, but haven't necessarily come to grips with how they feel about it.
SELZER: There's also the piece that is your part for democracy, that is, how often besides our vote do we have a chance for our voice to be heard. Through a poll, though you're aggregated with other people, you do have a chance, and it's in a format that you know that the candidates who are running for office are paying attention to. So there's really a payoff there. So not only do I encourage friends and family to do it, I also thank them when they report to me that they've done it.
SIMON: Well, you guys can call us anytime. Michael Dimock, associate director for research at Pew Research Center, and J. Ann Selzer, who is the president of her own company out there in Iowa, Selzer and Company. Thanks very much.
DIMOCK: Thank you.
SELZER: Thank you, my pleasure.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.