Public Affairs Forum
January 20, 2016
An interview with Christopher Wlezien, Hogg Professor of Government, University of Texas at Austin

Wlezien: Hi, I'm Chris Wlezien. I'm Hogg Professor of Government at the University of Texas at Austin.

(1) What decides elections?

Wlezien:Certain things matter more than others, and those are the things that I as a political scientist and other political scientists focus on. We call them internal, fundamental variables. The best known one is one's partisan inclination: Republicans tend to vote Republican, Democrats tend to vote Democrat pretty reliably year after year. What we're interested in is what these people who aren't Democrats or Republicans do. We call these floating voters. They can move back between candidates of different parties year after year. Those people are affected by what we call external fundamentals, things that are outside of them. These include things like the economy and a lot of it is just referendum-based judgments. Is the economy doing well, do we want to go four more years? But the choice also matters too. It's not just about your own internal fundamentals, and also not just about the sitting president and how he or she is doing, but it's also about the choice being offered on Election Day, that matters as well.

(2) What economic variables matter to voters?

Wlezien: Economics matters on Election Day, and it's really something larger. It's not so much about individual circumstances but about the collective. And it makes sense because individual circumstances—while related to the national economy, related to the state economy—are affected by a lot of very, very unique, specific things: health conditions, a bad boss, a disagreement with a coworker, all of which can affect one's individual circumstances. We tend not to hold presidents responsible for those kinds of things. We hold them responsible for the overall average trend—are things getting better for all of us on average, or worse? Are they getting a lot better, or just a little bit better? Those are the kinds of things that matter on Election Day. The slope of the national economy is what we should be paying attention to.

(3) What do trial-heat polls tell us about elections months away?

Wlezien: Well, survey organizations ask trial-heat polls all the time. These are polls that pit pairs of candidates, a Democratic candidate with a Republican candidate. They ask many different Democratic candidates and many different Republican candidates. What does a poll in January of an election year tell us about what's going to happen in November? Based on our research in the past from 1952 to 2012, it tells us absolutely nothing. You may as well just flip a coin. In fact, some people would say you're better off flipping a coin.

(4) How are voters assessing candidates between January and November?

Wlezien: The way I think about it, some Democrats may flirt with Republican candidates, and Republicans will flirt with Democratic candidates, but when it comes time to actually date, let alone marry, that's a different question altogether. Over time, the Democrats will learn that's really a Republican, and the Republicans will learn that's really a Democrat, and then they will end up voting for the candidate of their own party.

That's part of the story. The other part of the story is these external fundamentals, things like the economy and the performance of the president. At the beginning of the campaign season, the voters start taking into account the performance of the president which is oftentimes reflected in approval ratings. Those are really factored in during the first three or four months of the election year. After that, the economy evolves, these floating voters in the middle are taking into account what's happening with the economy—is it growing, is it growing a lot? And that becomes factored in. And then, at the time of the conventions, we see the candidates unveil their positions and platforms, and the choice between the Democratic candidate and the Republican candidate gets factored in. Then pretty much we're done. In the last 16 election years, we knew the winners in the polls in every one of them two weeks after the convention. So what we had in the fall was just a bunch of fine tuning.

(5) Do social media have a meaningful impact on voter decisions?

Wlezien: Social media and the Internet have allowed voters, seemingly younger voters and middle-age voters, a greater opportunity to control their information environments. One thing we know about that is that limits their exposure to information which would cross cut...information that would conflict with their prior beliefs, and so we find information that tends to be more reinforcing.

(6) How do polling methods influence poll results?

Wlezien: So we've gone from person-to-person interviewing to phone-based interviewing and now phone-based interviewing to Internet-based interviewing. There's a whole variety of standards being employed in the polls by phone and a whole variety of standards being employed by the surveys conducted by Internet, and so we have a lot of difference out there. This leads to a distribution of poll results which still may be centered on the truth but may lead us to see, from day-to-day movement as different organizations report poll results, a lot of bouncing around. Voter preferences may not have changed but the survey results have. My preference would be to look at the average of these polls. Better yet, if we are taking it really seriously, is try to adjust for the systematic differences in these polls.