We are awash in political polls that gauge support for the presidential candidates of both parties. Depending on how a given poll was conducted, you'll get wildly different stories on who’s ahead, among whom, and by how much. But do these polls matter?
On top of all the traditional challenges, the mechanics of the Iowa Democratic caucus makes polling a difficult thing to get right.
On a Saturday night in Des Moines, a few hundred people turn out for a performance of "Caucus! The Musical". (You really can't escape the caucuses this time of year.)
In the theater lobby, Dan Chase describes what he's thinking about the Democratic candidates:
“I've been a Bernie fan for the longest time but I'm starting to waffle a bit,” said Chase.
His wife is an ardent Martin O'Malley supporter, and now he's leaning that way too.
"I'm actually in the precinct captaining position for Sanders,” Chase admitted. “But I'm going to have to see who I pull the trigger for when we get there on Monday night.”
“Don't tell the Sanders campaign!" he added with a chuckle.
And that’s the trick in polling for a caucus; a lot can change once the participants are in the room. You might be swayed by a neighbor's impassioned argument in support of another candidate. Or your preferred candidate might not get enough support to stay in, and you’re forced to move on to your second choice. It's a fluid process.
Ann Selzer is a pollster, but not just any pollster. Selzer conducts the “Iowa Poll” for the Des Moines Register, it’s one of the most solid barometers of what will happen in the Iowa caucuses.
This month, NPR sent a cautionary note to the entire public radio system: beware of polls — they are not created equal. But among them, the veracity of Selzer's work is trusted.
Selzer polls likely caucus goers for four consecutive days – she started Tuesday and finished on Friday. And her poll will show how each candidate’s momentum changed in these final days before the Caucuses.
“You want to see it? I have charts!” Selzer says.
She pages through a manila folder and looks for the poll results from a year that might be of particular interest to Vermonters.
“Here's 2004," she says, pulling a chart out of a sheaf of papers. "Here's John Edwards, John Kerry and Howard Dean, all within two points of each other on the first day,” Selzer says.
She recalls that days before the caucuses that year, Howard Dean was widely expected to win. But that's not what her poll showed.
"All we can measure is their intent when they walk in the door because things can change once they're in the room." — Ann Selzer, Iowan pollster
“So here he is day two, he's now down below Kerry,” Selzer explains, gesturing to the data points. “Oh, here he goes, oops, day three. Day four he comes in behind Gephardt. He comes in fourth place. We referred to this as the 'Register Graph of Doom for Howard Dean.'”
Dean finished in third place in Iowa, not fourth as Selzer's poll suggested, but well out of first place, which was the prevailing expectation.
In 2012, her poll was one of the few indications that Republican Rick Santorum would go on to win Iowa that year.
But even one of Iowa's most respected pollsters hedges on what she can predict.
"All we can measure is their intent when they walk in the door,” Selzer explains. “Because things can change once they're in the room."
Which brings us back to Dan Chase, the Sanders supporter who's starting to waffle.
"I think if O'Malley shows any viability, and it's a matter of one or two people stepping into that camp [and] taking one delegate into his camp, I'd be very seriously considering it," Chase says. “ I think that'd be a real possibility."
Selzer's poll earlier this month showed Sanders at 40 percent and Hillary Clinton at 42 percent. But it's the results she'll release Saturday in the Des Moines Register that will be a better indication of what could happen on Monday night in Iowa.
VPR ’s coverage of the 2016 presidential campaign is made possible in part by the VPR Journalism Fund.
