When you're a political pollster, the nuts and bolts of your day-to-day work is gathering and presenting data, so there's not usually much fun to be found. But somehow the folks at Public Policy Polling (PPP) manage to have a grand ole time. The left-leaning Raleigh-based firm asks respondents the usual questions about approval ratings and who they'd vote for, but PPP has also done surveys about whether voters think Ted Cruz is the Zodiac Killer, demonstrated that Republicans hate lunch, and regularly asks readers for question suggestions. This week, PPP asked respondents whether they'd support building a wall not just on the Mexican border but also on the Atlantic Ocean, in order to keep Muslim migrants out of the US—and 31 percent of Trump backers liked the idea.
Maybe the pollsters' best prank came this July, when PPP showed that by some measures, Green Party presidential candidate Jill Stein is a less popular than beloved dead gorilla Harambe. Then they spiked the ball for internet points: When Stein memorialized Harambe on Twitter—something she later revealed to be some sort of point about the media—PPP seized the opportunity to tweet its poll result at her.
Among polling wonks, PPP has been controversial for its methodology, which involves a lot of weighting of its raw data. But its results in 2012 were accurate, and poll-focused political website FiveThirtyEight gives PPP a B+ grade.
To find out if trolling voters with poll questions has any actual public benefit, I called up PPP polling director Tom Jensen, who explained what Harambe and giant meteors demonstrate about American democracy. Our conversation has been lightly edited for length and clarity.
VICE: You've generated approval ratings for politicians alongside approval ratings for celebrities like the Duggars. What's the point in doing that?
Tom Jensen: We think that polling can be a combination of what's serious that's going on at any given time, and what's not so serious that's going on at any given time. I don't even remember polling about the Duggars—I'm sure we did, but I can't speak to what the specific motivation for that one was. A lot of the time the sort of funny or less serious questions that we ask really do tell us a lot of things about what's going on in the election.
Is that also why you ask about fake issues?
Something people really found interesting about the poll we put out today was our question about building the wall around the Atlantic Ocean to keep Muslims out from the Middle East. I think the fact that 31 percent of Trump voters support that—which is clearly a ludicrous idea—says something about the unrealistic things that Trump voters might like him to do if he gets elected president.
It seems like we're seeing people who aren't that well informed and then laughing at them. Isn't there some baseline of people like that in every poll?
There may be. But that explains, to some extent, how we ended up at the point where we're at with Trump. I remember exactly this week last year, we had a poll that came out that showed that 66 percent of Trump voters thought that Obama was a Muslim, and 21 percent of Trump voters thought that Obama was actually born in the United States. That kind of finding was something that, I think, kept us from falling into the trap that some people did of assuming that Trump was going to fall apart, because from our perspective, his people all think [those kinds of things]. And he's pretty much the only candidate who's going to appeal to that kind of thinking.
But has demonstrating that they have these beliefs had any actual benefit?
I think it sort of helped explain the staying power he had—from episodes he had where everyone assumed he was going to fall apart, and he was able to hold on. And it's because his supporters have this very unique set of views that only he was able to tap into. A lot of how we figured out what those things were was by asking what you might think of as troll questions, but I think they actually went a long way toward explaining why Trump had the unique appeal that he did.
Would you ask questions that troll Democrats too? For instance, if you asked Hillary Clinton supporters basic questions about guns, they'd probably say some stupid things too, right?
We're a Democratic polling company, so we ask more questions that are sort of geared at Republicans and Trump supporters and that sort of thing. We would not be offended at all if someone on the other side of the aisle were asking questions in that kind of vein. But we ask about the things that we're interested in.
Jill Stein said that stuff like Harambe is a distraction from more important issues. How do you respond to that?
The reason we poll the Harambes, the Deez Nuts, and the giant meteors of the world is kind of twofold from a serious perspective. Obviously it's partially just because they're funny, but they're sort of important. One of the reasons is—and this was especially true when we were polling [joke teen presidential candidate] Deez Nuts this time last year—it showed the extent, early, to which people were going to be really disgusted at the choices of Clinton and Trump. At this time last year, people were so mad about Clinton and Trump that 10 percent of them would go so far as to vote for Deez Nuts.
That's funny, but what's the takeaway?
It foreshadowed that we may have unusual levels of support for third-party candidates this year, which has ended up happening. I also think that when you include the Harambes of the world, and they out poll Jill Stein, what that shows is that the people who are supporting Jill Stein, a lot of them, it not might necessarily be that they have some great belief in Jill Stein. The people who are supporting Gary Johnson, it might not be that they have some great belief in Gary Johnson, they might just support any alternative that they have. So at that point, it's like, Let's calm down a little bit about how we have record numbers of people buying into the Libertarian agenda and all these people excited about the Green Party. If those folks are polling on par with Harambe, it just means people are casting a protest vote more than voting for Johnson or Stein because of a deep belief in what they think.
Is there ever a concern that silly questions in your polls can skew the results?Our number one goal on absolutely every single poll is just to get it right—to get as accurate a picture of what's going on as possible. Nate Silver [of FiveThirtyEight] has actually found this cycle that our polls are leaning one to two points more Republican than the polls as a whole, so even though Trump supporters on the internet hate us, we've actually been giving Trump better polls than most companies have. One thing to be very clear about is we always ask the serious questions at the start and then ask the lighter stuff at the end of the poll, so we make sure we're not messing around with the answers to the more important stuff by asking some of these offbeat questions.
Follow Mike Pearl on Twitter.