Republican presidential nominee Donald Trump debated Democratic nominee Hillary Clinton on Monday night on Long Island. (Left photo by Paul J. Richards/AFP via Getty Images; right photo by David Goldman/AP)

Is the 2016 presidential election a personality contest? That’s what many observers apparently think — and it’s how many interpreted Monday night’s debate between the two nominees.

“From the beginning of the 2016 presidential race, Donald Trump has faced a nagging question: Does he have the right temperament to be commander-in-chief?” wrote Linda Feldmann of the Christian Science Monitor. At USA Today, political scientist Ross K. Baker wrote, “Clinton will probably get a modest bump in the polls for her solid and well-informed performance, but the debate was a personality test more than it was an SAT.”

Can Republican nominee Donald Trump prove to voters that he has the temperament to be president? Can Democratic nominee Hillary Clinton persuade the public that she’s more trustworthy than many believe?

Of course, the “personality contest” narrative isn’t new. Ronald Reagan’s charisma and Bill Clinton’s ability to connect with voters are, in the folklore, key to their electoral success. Because Trump and Clinton’s difficulties seem particularly acute, many assume that voters’ sense of their personalities will decide who reaches the White House.

But that theory is often wrong. In a new paper, I show that over the last 60 years, presidential candidates’ personal attributes have actually become less important to voters and less correlated with election outcomes.

So, what really matters? Policy, group affiliations and ideology.

Voters don’t care as much as they used to about candidates’ personalities

Consider the graph below. There we plot the percentage of American National Election Studies respondents who, when asked what they liked and disliked about the major party candidates, mentioned personal attributes. These include comments about the candidates’ integrity, reliability, competence and leadership abilities, along with appearance and such demographic information as age, race, social class and former occupation.


In 1952, for instance, 79 percent mentioned Dwight Eisenhower’s or Adlai Stevenson’s leadership abilities, intelligence and so forth. By contrast, that number had fallen to 61 percent in 2012 — an election in which Mitt Romney’s “empathy deficit” was supposed to be a big voter concern.

Other research has been finding the same thing: What voters say they like and dislike about the candidates now are their positions on issues. And so as the parties have polarized on everything from policing to abortion to health care, the candidates’ policy positions have mattered more and their personalities have mattered less.

In fact, I find that young Americans are now substantially less likely than older people to focus on candidate personality. Having grown up in a much more polarized political environment, young voters have come to focus more on policies than candidate character. If this generational pattern continues, personality will probably keep getting less and less important in elections.

“Which candidate would you want to have a beer with?”

Another way to examine this trend is to look at how favorably voters have evaluated the candidate’s personal qualities since the Eisenhower era — or, to use the common line, which candidate a voter would like to have a beer with. To do so, I created a score based on how many positive and negative comments ANES respondents made about each candidate’s personal attributes. When there’s a higher ratio of positive to negative comments, there’s a higher score.

I plot these scores in the graph below, with separate lines for the winning and losing candidate. Both lines trend downward. That tells you that voters like the candidates less and less over the years.


Here’s the flip side of this: Having the more appealing personal qualities is no longer a good predictor of who wins. In the 10 elections between 1952 and 1988, the candidate with the personality advantage won seven times. Since then, however, the candidate with the higher score has won only twice: Bush in 2004, Obama in 2012. Bill Clinton survived abysmal ratings on integrity to win in the 1990s; the 2000 election was fittingly a tie on personality, and in 2008, McCain enjoyed a wide lead on competence over Obama.

Or to be more direct: Presidential elections are rarely won or lost these days on the basis of candidate personality.

In our increasingly polarized politics, people have come to hold more black-and-white views of the candidates — and judge personal character through the lens of political bias. Marc Hetherington, Meri Long and Thomas Rudolph find more and more polarization in ratings of recent candidates’ honesty, competence and leadership in recent elections, with voters assessing those traits based on party.

I found the same pattern in the open-ended comments. Between 1952 and 2000 the average correlation between the personal ratings of the two major candidates was -0.26; in the past three elections the correlation has averaged -0.40. What this means is that, increasingly, the more one likes one candidate, the less one likes the other. Gone are the days when people commonly said they liked both candidates’ personal qualities.

Of course, voters still care about candidates’ personalities. It’s always better to be seen in a positive light than a negative light. But a disadvantage on candidate personality can now be overridden by other factors, such as policy and group affiliations.

Martin Wattenberg is professor of political science at the University of California at Irvine and author of “Is Voting for Young People?” (Routledge, 2011).