Harry Enten is a senior political writer and analyst at FiveThirtyEight. Prior to working for FiveThirtyEight, he covered the 2012 election cycle with The Guardian. He graduated in 2011 from Dartmouth College with a B.A. in Government, where he started Margin of Error, an elections statistics blog.
The Politic: How did you get interested in political forecasting?
Harry Enten: In a couple of ways. One, my father took me into the voting booth when I was four for the 1992 presidential election and he took me to vote pretty much every year after that –’93 with the NYC Mayoral Election, ’94 for Governor of New York, ’96 for the presidential election, and so on. Two, I’ve always had a fascination with trying to predict things. When I was very young, I was interested in baseball and football, and I still am. Eventually what happened was the early interest with elections sort of merged with the early interest in prediction, and it turned out that I was pretty decent at it, or at least better than other people were at it. And since I was pretty good at it, and people seemed to listen to me or at least hear me out, I decided late in high school that I was really interested in that sort of thing. Then I went up to Dartmouth, majored in Government and concentrated on the statistical aspect of it. I used statistics at any chance I got, whether in a Government class or Women’s and Gender Studies class, projecting why pro same-sex marriage ballot measures seemed to be doing better. When I was in college, most states had a ban on it. Any chance I had, I would try to use statistics to explain a political event or try to forecast one going forward.
The Politic: You worked at The Guardian before starting at FiveThirtyEight. What prior experiences led you to your work at The Guardian and now FiveThirtyEight?
Harry Enten: When I was in college, I did an internship with the NBC News Political Unit for eight weeks, and then a year after that, I interned with Pollster.com. Half of the internship was with the National Journal and the other half was with Huffington Post. And then, I got out of college, stayed on my couch for six months, and knew what I wanted to do. I just didn’t know where I could do it. Eventually, I kept up my blog, and The Guardian came knocking, and I did some part-time work with them as the 2012 Republican Primary developed. After that, I worked full-time with them as we entered the fall campaign. I’d been at The Guardian for a year and a half when FiveThirtyEight moved from the New York Times and expanded by signing a contract with ESPN. Nate came knocking, we mutually agreed it would be nice for me to work there, and I signed up with them in November 2013.
The Politic: FiveThirtyEight is arguably the most well known organization in America dealing with this sort of data-driven research and analysis. However, it is definitely not the only place. What makes working at FiveThirtyEight unique or different from other places?
Harry Enten: One of the things that’s different is that we tend to have good support within the organization. We have a great graphics team, so we’re able to display what we’re saying statistically in a fashion that is easily digestible by the reader. But another thing that’s key is that this is a statistical website where we can get nerdy. We don’t have to be afraid that it might be too complex for the reader. A mainstream medium like the Guardian is fantastic in terms of getting broad readership. But the statistical analysis – at least the explanation of it – will be on a different level. Being at FiveThirtyEight allows me to expound and explain what I’m doing statistically. We also have very smart people in the room– statistically smart people. Instead of there being one or two of those, there’s seven or eight or nine of them. If I want to ask a question or bounce an idea off of someone, it’s very rare that I won’t be able to have the opportunity to do that.
The Politic: That’s fantastic. So how do you go about writing statistical pieces? What’s the process behind the brainstorming, research, and publishing?
Harry Enten: There are two ways you go about doing it. One is you see what’s in the news. And then, you’re just interested in it, and you find an angle to answer a particular question that’s being posed or implied by people’s research. You’re going to look at the polls and the raw data, and see if there is anything that supports or opposes the conventional wisdom. The other way is that you can take a broad topic that’s just generally interesting, like the 2016 campaign right now. I’ll spend a few minutes looking at a polling database and see if there’s anything that catches my eye. One of the things that’s interesting to me is looking at Chris Christie, who has become very unpopular with Republican voters. Does that matter? It turns out it’s never happened that somebody who was as unpopular as Christie is at this point in the primary process has gone on to win [the nomination]. That only happened because I had a few minutes that I could go around and see something. I investigated it further and was able to come up with a finding that made sense but could also make its own news.
The Politic: With regards to data making its own news…you see a lot of mainstream news media that use data to substantiate some perspective rather than letting the data do the talking. How do you try to remain impartial while hunting for data that can still make a point?
Harry Enten: Any good scientist who has a method generally has a hypothesis when they’re going to look at the data. But at FiveThirtyEight, if we see the data does not prove our hypothesis but actually opposes it, then we try to make that known as well. For example, with Christie, if it turned out that someone as unpopular as him actually had gone on to win the nomination, we’d publish that. And we’ve done that sort of piece before – at least Nate has. In the 2012 presidential election, Romney’s favorability [ratings] early on were not very good, but we saw that historically, early favorability ratings don’t mean much. If you have good data, you’ll write the story if it supports or opposes what you thought. If the data neither confirms nor denies your data, then you’re at a crossroads. If you’re writing for a scientific journal or academic institution, that would be published no matter what. But as a news outlet, we always want to make news. We don’t manipulate stories, but we want to add something new to what’s already being said. We want to publish something that will help people better understand something, and if we can’t, then we probably won’t.
The Politic: With sites such as FiveThirtyEight and startups such as CrowdPac, there’s more access to political data for the public. Do you think that this data can have an effect on political campaigns and public opinion, or do you think that traditional forms of campaigning and mainstream news media will prevail?
Harry Enten: I don’t think most voters are reading our website. I think most voters would make up their minds the traditional way –through TV, especially in the primaries. We can help the press better understand why voters are making the decisions they’re making. For example, why is Donald Trump doing so well? Well we can tell you he’s not doing any better among moderates than he’s doing among conservatives, or he isn’t doing any better among Tea Party supporters than non-Tea Party supporters. Everybody thought Trump was getting all these really conservative voters, but that’s in fact not what’s happening. We’re able to say that there’s nothing really special about him, at least nothing we can measure with the traditional metrics. Back to the question: no, I think we just help people who are interested be better informed. Campaigns obviously have their own metrics they use, and they probably have better data then we do. But based upon the public available data that we have, we do a pretty good job of informing people who are interested in being informed?
The Politic: You mentioned Donald Trump, one of the many candidates in this 2016 presidential race, which I want to get your take on. How do you compare Hillary Clinton’s current campaign to her campaign eight years ago?
Harry Enten: I think they’re very different. One, when she entered the [2008] primary, she was polling in the 30s nationally and in New Hampshire about the same. I think in Iowa, the first poll had her trailing John Edwards. This year, the polling looks very different. She’s polling mostly in the 50s and low 60s, depending on which polls you look at nationally. In New Hampshire, she’s polling in the 40s – that’s her closest state – and in Iowa she’s polling in the high 40s to mid 50s. Her support is just substantially greater than it was eight years ago. Also, what makes this year different is that there is no real alternative. Barack Obama was a credible alternative in the 2008 campaign. Even John Edwards was a credible alternative. Bill Richardson and Joe Biden were credible alternatives. These were people who held office or previously ran for the presidency and were legitimate candidates. This year, who do we have who’s against her? We have Bernie Sanders, who is for all intents and purposes an Independent with Socialist leanings and is just not going to be supported by the party establishment. We have Martin O’Malley, who probably wouldn’t have been able to win re-election in Maryland. We have Lincoln Chafee, who has never won office as a Democrat. We have Jim Webb, who won one term and is far too centrist for the current Democratic Party. So, Clinton is in a much better position now, not only because of where she is in the polls, but also because the field is significantly weaker. She has much more support from the party. Every single one of the current Congressmen or governors that has endorsed a candidate has endorsed Clinton except for one.
The Politic: You recently wrote about John Kasich, and he’s been picking up some steam recently. What potential do you think he has?
Harry Enten: Kasich has made a point in the past of making his positions seem more moderate, and voters to seem to be picking up on that. Kasich has risen in some polls, but I think the average of the last five polls this morning was 2.8 percent. In New Hampshire, he’s risen more so. But keep in mind, he’s been up on the air there, so of course he’d rise a little bit. To me, Kasich is probably the best shot of anyone not named Bush, Rubio, or Walker. But it’s certainly not higher than ten percent, maybe the high single digits.
The Politic: In the national contest, what do you see as the most electable GOP ticket and the most electable Democratic ticket?
Harry Enten: On the Republican side, I think one of the four people – Bush, Rubio, Walker, or Kasich – would be the most electable. These are people who have won in swing states. Walker, who’s seen as more conservative, has won Wisconsin and could do fairly well in the Midwest. I think the Republican Party will elect somebody who can win in the fall. In terms of vice presidential candidates, it depends on who the actual presidential candidate is. You probably want somebody who crosses off a check box – maybe somebody who represents a certain swing state, or maybe somebody who represents a certain constituency. But vice presidents tend to be overrated in terms of their helpfulness. What you’re really interested is not so much helping out your ticket as much as not hurting your ticket. If you want a swing state person, you could choose Portman, the Senator from Ohio. You could choose a Latina in Susana Martinez. Perhaps Brian Sandoval from Nevada, but he’s pro-choice, so he is probably off of the ticket. Essentially what you’re looking for is to not make the mistake McCain did, where you choose a vice presidential candidate who at the end of the day arguably hurt the ticket.
On the Democratic side, I would say Hillary. It’s too far away to predict a vice presidential candidate. It will depend on how she’s doing in the polls at that point, and whether she needs more Latino support or African-American support or what state she’s doing well in. She’ll be the nominee though – I have no doubt about that.
The Politic: I watched the Ted Talk you gave about a model that can predict which party will take the White House in a presidential election. Using that model, do you have an idea of what party might have the edge in 2016?
Harry Enten: For 2016, if we’re looking straight at the economy, I think you would put the Republican candidate as having a better shot to win than the Democratic candidate. Of course, the margin of error on this is fairly wide, and we don’t know where the economy will be a year from now and how it will develop. But based upon what we see – based on upon purely the economy – the Republican candidate is likely to be the favorite at this point. Not a tremendous favorite, but a likely favorite.
One of the things I’d stress is that we’re dealing with a small sample size, going back to elections where we had some of this [economic data]. And more than that, we’re dealing with a small sample size in terms of elections that didn’t include an incumbent president, where the largest errors for this sort of model occur. The best bet at this point is to predict a relatively close race.
The NBC/WSJ poll that came out last night had the generic Republican candidate leading the generic Democratic candidate by two [percent]. That’s been mostly within the margin of error the last few times they’ve asked that question. But I would certainly look at that over these Hillary Clinton vs. whoever polls. Those tend to be highly reliant on name recognition and how each person is doing in the primary. If you look back to 2000, for instance, at this point George W. Bush had a large lead over Al Gore in the horse race polls. But the generic ballot tended to show the generic Democrat and Republican candidates were fairly even. Obviously, if the economy rises or falls, we can see results significantly different from the generic ballot, but at this point, I think the best bet is a close election.
The Politic: I’ve got one more question. How would you advise someone who’s entering college and interested in forecasting and analytics to pursue their interest?
Harry Enten: One, I would suggest obviously taking a statistics course. Two, I would suggest doing independent research projects, in which you write a lot. A test grade doesn’t bring you closer to your goals and your future. Three, I would suggest starting your own blog. Few will probably read it at first, but it will help you hone your writing and you can feel free to send those writings to someone like myself. You can tweet them out. You can tweet them at reporters. Four, make sure what you’re writing is of interest to people. Don’t write something on the 1952 election, something that’s not interesting at this point. Five, I would definitely make sure that you get internships. And the way you do that partially is by distinguishing yourself – pointing to a blog or pointing to past experience that you have. And then finally, you have to get an eye and ear for the news, and learn what people want to report on and find out about. If you can do all of those types of things, then you could potentially really have something.