ST. PETERSBURG — Candidates often decry them as useless, but politics at the state and national levels just wouldn’t be the same without polls.
At the hyper-local level, though, scientific surveys of voter sentiment typically are nonexistent.
Two guys are trying to change that, at least in Pinellas County. After predicting with some accuracy how local voters would behave in the 2012 presidential election, The Pier referendum and the city’s mayoral race, it looks like local politicos have started paying attention.
St. Pete Polls President Matt Florell said he got the idea to start polling on local issues and elections a couple years ago, in part as a way to pique voter interest.
“My interest in this was started when I looked into how low the voter turnout was in the city’s 2011 election,” Florell said.
So he used software he developed for business telephone systems — his day job — to call voters selected from publicly available registration records. It made sense for a couple of reasons.
“We wanted not only to get people out to vote, but also to kind of gauge people’s opinions on topics of the day, like the Pier,” he said.
He ran the first poll in November 2011. It asked voters about a range of issues, including emergency medical services, the Tampa Bay Rays and The Pier. He partnered soon after with Leonard Schmiege, founder of the voting rights nonprofit group ShadowVote. Schmiege had run unsuccessfully for St. Petersburg City Council’s District 8 seat in 2009 and had hired Florell to set up robocalls for him ahead of Election Day. He said he got interested in polling because he’s skeptical of electronic voting systems.
“Polling is the ultimate double-check on elections,” he said. “Mistakes happen. Any time you have polling on top of it that’s published you have a fantastic double-check.”
Now he mostly handles sales for the firm’s commissioned polls.
Florell said more than half of the polling they do now is sponsored by candidates and other private entities, and results often aren’t released. But that amount changes from month-to-month, and their independent work doesn’t turn a profit. Writer and political consultant Peter Schorsch, founder of the local and state politics blog SaintPetersblog, commissions a fair amount of polling.
“What St. Pete Polls has achieved is to make sure their data is influencing their community,” Schorsch said in an email. “I can’t count how many times I’ve heard a local candidate or official talk about what the local polling shows. That’s quite a feat.”
At a time when politics is heavily driven by data, he said, they’re not only gathering crucial numbers in a way no one else is, but doing it in the right way. “They knew not only what questions to ask, but how and when,” Schorsch said.
A page on the St. Pete Polls website details their methodology, which involves amassing a list of registered voters who provide landline phone numbers. The list must be big enough to provide a representative sample of a given city, county or congressional district. After the polling is done, the results are weighted using an algorithm “to ensure a proportional and more accurate total percentage.”
There’s still lingering skepticism about automated telephone polls, though, especially those that only call landlines.
“It’s not a very rigorous method of polling,” said Robert Santos, president of the American Association for Public Opinion Research and chief methodologist at The Urban Institute, a Washington, D.C.-based public policy think tank. “What you end up with is a self-selected poll.”
That’s because those who are politically engaged would be more likely to participate, but less-informed voters might just hang up, he said, adding that calling landlines might cause results to have “a disproportionately high number of retirees.
“In these types of surveys, they work until they don’t work,” Santos said.
Due to recent FCC rules changes, polling agencies may no longer call cellphones, which has put a damper on their work in places like Tampa, where many people have a cellphone but no landline. In retiree-heavy Pinellas County it’s still possible to get a representative sample, though, Florell said.
And so far, they’ve had some luck in their findings. Schorsch said that they accurately predicted the outcome of the Gainesville mayor’s race in April, as well as the results for the top-of-the-ballot questions in the 2012 general election.
Despite being called “junk science” by Mayor Bill Foster during his unsuccessful re-election bid, they successfully predicted Mayor-Elect Rick Kriseman would win, though they underestimated by how much. His margin of victory — 11.8 percent — was almost double the six percent they predicted, which Florell attributed to a much higher number of Democrats turning out than in the August primary, which was the basis of their November projection.
“Their projects are, like any pollster, hit or miss,” Schorsch said. “But they hit more often than they miss.”