A Professional Reader of Voters’ Minds Discusses His ‘Blunt Instrument’
Wednesday, October 27, 2010
This is the fourth in a new five-part series called "The New York Vote," a partnership between WNYC and Capital New York. We will be painting a portrait of the New York electorate in 2010, as explained by a diverse cast of political players.
Today a look at how Mickey Carroll, director of the Quinnipiac University Polling Institute, became a believer in polling.
When Mickey Carroll was a reporter at The New York Times many years ago, he argued against the paper’s move into polling. His thinking: why pay for a poll, when intrepid reporters could get the goods just as efficiently?
These days, Carroll, who directs the Quinnipiac University Polling Institute, is a convert. He was wrong, he said, to argue against polls as reporter. First of all, he realized that politicians had increasingly come to rely on polls, and independent, public polling allowed papers and their readers a way to access the same information those politicians were gathering. And second of all, the polls were actually kind of useful in telling a story.
“The polls measure sentiment,” he said. “They measure what’s going on, and they’re the only measure. They’re the only practical—the only legit—measure of public opinion. And it’s nice to know public opinion.”
The challenge, for pollsters—and for the media that dutifully reports on their findings--is that what they “know” isn’t public opinion, but merely a version of it.
Take the polling on gubernatorial candidate Carl Paladino. Quinnipiac made headlines at the end of September when it released a poll showing Paladino was only six points behind Andrew Cuomo in the general election.
“I worried about it,” Carroll admitted. “When I saw the numbers and I talked to the numbers people, I said, ‘This goes completely against what I think.’”
But the team checked it over, and came back with the same results.
The poll, however, had come just after the Republican primary, and almost a third of the people Quinnipiac surveyed—31 percent—said they hadn’t heard enough about Paladino to form an opinion. (By contrast, only 13 percent said they hadn’t heard enough about Cuomo.) Casting further doubt on the results, rival polling outfit Siena Research Institute came out with a poll the very next day showing Cuomo leading Paladino at 57-24 percent.
Not everyone is a convert. The Village Voice’s veteran political reporter Wayne Barrett accused Quinnipiac of deliberate sensationalism. “Quinnipiac knows that the more shocking the finding, the bigger the play,” he wrote.
By the time Quinnipiac released another poll, two weeks later, only 15 percent of those asked said they were lacking information on Paladino, and Cuomo came out well ahead.
If nothing else, the difference between Quinnipiac’s two polls seemed to tell a story about a rational electorate, which reserved judgment on a candidate until learning a little more about him. In Paladino’s case, the information voters had gathered about him over those two weeks, when what Carroll calls his “loony toons” behavior really picked up, turned them off.
“People do pay attention,” Carroll said. “It’s a popular government. We chose our representatives…By and large, people take their vote seriously. They don’t like to throw it away.”
Some people, of course, take their responsibility to democracy more seriously than others. Pollsters generally start out polling registered voters, but as an election nears, they try to winnow down their polling samples to “likely voters.” They ask a series of questions to try to divine a person’s true intentions: How interested are you in this election? Do you care about your candidate? Do you think you’re going to vote?
Ultimately, that group of people looks different from the New York population as a whole.
“The likely voter is probably a little richer, probably a little whiter, probably a little more suburban…probably a little more Republican,” said Carroll. “Poor people don’t vote as much as the middle class or well-to-do people.”
The very young and the very old tend not to vote either.
“The young haven’t gotten there yet and the old, I don’t know--maybe they get fed up, or something,” Carroll said.
As a rule, only people who have voted before count as likely voters, too. So people who’ve recently moved to New York and haven’t voted in an election yet don’t show up. In an election like the 2008 presidential election, where President Obama inspired a significant set of non-voters to get involved in politics, pollsters have a harder time getting accurate results.
“A poll is basically a blunt instrument—yes or no,” Carroll said. “You try to shape it so it’ll reveal attitudes.”
It doesn’t always work. After Paladino swept the Republican primary with his “I’m mad as hell” slogan, Quinnipiac’s polling team tried asking a question they hadn’t before about voters’ feelings towards state government: “Do you feel enthusiastic, satisfied but not enthusiastic, dissatisfied but not angry, or angry?”
“Essentially, it didn’t work,” said Carroll. Plenty of people were angry at state government, but those “angry voters” weren’t particularly inclined toward Paladino. “Because they say they’re angry, doesn’t means they’re angry, and because they don’t say they’re angry, doesn’t mean they’re not,” Carroll said.
These kinds of vagaries might explain why Carroll says he hasn’t gained all that much insight into the New York electorate from his close involvement with polling.
“I was supposed to know about New York voters when I was a reporter!" he said. "I don’t assume I know an awful lot. But I probably know more than people who didn’t devote their lives to covering it.”
When Carroll talks about the New York electorate, he tends to segment the vote along ethnic and geographic lines: “New York is not one great big, homogenized place. Is Carolyn Maloney’s district is the same as Jerry Nadler’s district? No. They’re all Americans, they’re all New Yorkers. But they are different.
“Chuck Schumer makes it his business every year to go to all 62 counties of New York. And he’s going to find that things in Oneida are a little different than things in the Bronx. Every borough is different. On the Supreme Court, you have a member from the Bronx, one member from Queens, one member from Brooklyn, and one member from Manhattan. And each one of them sort of reflects the tone, if you will, of their borough. Look at Judge Sotomayor and Judge Scalia. They’re both very smart judges, but she’s Bronx, he’s Queens.”
In New York, Carroll said, “There’s a larger and ever-growing Hispanic electorate. People talk about the Jewish vote as if it were one bloc…The Jewish vote, to the extent that you can generalize, was a more important vote in New York ten or fifteen years ago. The electorate overall is Democratic – registration shows you that. It’s reasonably liberal. The one rule of thumb in New York elections has always been that the liberal candidate wins.”
(Carroll also offered this more parochial insight into what it's like to poll in some nearby states: “You call Connecticut, you call Pennsylvania, people say, “Oh, yes, yes, can I help you?” You call New Jersey, people say, “Why? What do you want to know that for? Where’d you get my number?”)
Sometimes, the polls attempt to get at more subtle questions about public figures than whether they’re electable. Quinnipiac did one just after Michael Bloomberg became mayor. They wanted to find out if people related to Bloomberg, and so they asked, “Would you like to have Thanksgiving dinner with Mayor Bloomberg?” (They did not.)
Pollsters often describe what they do as taking as snapshot in time. And they do ask voters what they would do if the election "were held today.” But, “Everybody looks on them as predictions, and essentially they are,” Carroll said.
Sometimes the polls are more than just predictors. By setting voter expectations, they actually become factors in the election--affecting the outcome rather than merely guessing at it. And that can be problematic.
“Do polls affect elections? They do, and that’s not good,” he said. “The bad part is that they make it look a horse race. In some cases, they dry up interest. If it shows that someone is going to win in a walk, his people may decide, ‘Oh, I’ll stay home. Why bother?’”
The key, according to Carroll, is not to take polls at face value, and then decide what it is, exactly, that the poll is saying about voters.
"People should look at polls with respect, but with skepticism," Carroll said. "They should look at the questions and then make your judgments based on them. Look at the track record. Is this a poll that has been tested, has been reasonable, seems to not have an ax to grind? And then, look at ‘em and see and if the poll’s right, you say, ‘Wow, that’s pretty good.’"