As Abraham Lincoln famously said, "Nearly all men can stand adversity, but if you want to test a man's character give him power." Because the presidency is a uniquely personal and powerful office, there is ample evidence that character matters enormously in terms of governing, as Fred Greenstein (2009) has so aptly noted in The Presidential Difference. Furthermore, American presidential elections are inherently personal contests, as unlike in parliamentary systems, voters are able to cast a vote directly for the nation's chief executive. Recognizing how factors like personal integrity, competence, reliability and leadership skills have made a difference in past presidencies, American voters naturally take such factors into account when they cast their ballots.
Journalists are typically the most aware of the personal factor in presidential elections. It is hard not to ignore candidates' personal attributes after spending day after day riding on the same planes and buses with them, listening to their speeches, and asking questions whenever given the opportunity. In short, they are repeatedly exposed to the person as well as the message, thereby giving them the chance to evaluate how personal characteristics might affect governing. As Peter Hamby (2013) wrote in his review of Halperin and Heilemann's (2013) best-selling journalistic account of the 2012 campaign, "Candidates matter. Voters tell pollsters that they make their choices based on issues such as education, health care, taxes and the economy--and they do. But they also care about temperament, empathy, strength, reason, trust and the human side of these strange and wily people who think they're up to the task of running the country." Or, as veteran campaign reporter Jack Germond (2002, 259) sagely put it in his memoir, "The fatal flaw in the sorting of candidates by issues is that it is almost impossible to anticipate which issues will confront a president during his four years in office. No one was thinking about Saddam Hussein when George Bush was elected or about Monica Lewinsky when Bill Clinton was elected." In sum, voters are faced with the task of hiring someone to take on a job replete with unforeseeable challenges; a president's character strengths and weaknesses are bound to affect how they approach the decisions and nondecisions they make.
Scholars of the American presidency are not only aware of how much difference character makes, but also how presidents often try to prime the public to think about them in personal rather than policy terms. The idea here is that policy decisions will inevitably have opponents as well as proponents but that if a president can be portrayed with a positive personal image there will be little down side. Druckman and Jacobs (2015) have recently shown how presidents have used polls to devise methods to focus on personality features, such as strong leadership, competence, and so on, as opposed to controversial policies. For example, they find that Richard Nixon and Ronald Reagan regularly conducted polls to assess the public's assessments of their trustworthiness as well as perceptions of their competence and strength of leadership.
But how much do voters really focus on personality matters at election time, and has the importance of personal attributes increased or decreased in recent elections? Even though the classic analysis in The American Voter (Campbell et al. I960) identified candidate personality as one of the three major factors in voting behavior, scholarly analyses of the role of candidate attributes have been relatively rare in comparison to the other major factors of party identification and issues. (Some exceptions are Funk 1999, Kinder 1986, and Miller, Wattenberg, and Malanchuk 1986). This article attempts to counter this imbalance a bit by examining the role of candidate attributes in presidential elections between 1952 and 2012. During this sixty-year period it will be shown that the electorate's focus on candidate attributes has declined substantially. Although candidate personality is still important in American voting behavior, it has become significantly less so in recent years--in particular during the elections of 2008 and 2012.
The Declining Mentions of Candidate Attributes to Open-Ended Questions
The analyses in this article are based on the set of open-ended questions that have been asked in the American National Election Studies (ANES) in every presidential election from 1952 to 2012. Respondents have been asked the following two questions with respect to each candidate: "Is there anything in particular about [candidate's name] that might make you want to vote for him?" followed by "Is there anything about [candidate's name] that might make you want to vote against him?" Interviewers have transcribed exactly what people have said in response to these questions, thus allowing respondents to express whatever is on their minds in their own terms. The responses to these open-ended questions over the last sixty years provide one of the richest data sources available on the factors determining voting decisions and the popularity of presidential candidates.
The ANES has carefully coded all of the responses for the studies conducted between 1952 and 2004 using an elaborate coding system. Although such coding has not yet been done for the 2008 and 2012 studies, ANES has released an Excel spreadsheet containing all the verbatim responses to these questions, thereby enabling scholars to code the responses according to their own framework. For this article, I had my research assistant Sierra Powell use the standard ANES coding scheme to specifically code just the responses that referred to a candidate's personal attributes. (1)
There is good reason to expect that survey respondents should be less likely to mention candidate personal attributes over time as the electorate's focus has turned from performance to policy in recent years (Wattenberg and Powell 2015). Previous research on what people say about candidate attributes has found that much of what respondents say is performance relevant (Miller, Wattenberg, and Malanchuk 1986). For example, when someone says that a candidate does not have enough experience, it often means that they do not think he will be capable of doing a good job. Or, if someone says that a candidate is not a good leader, the implicit assumption is that he will not be able to rally people around him to accomplish his stated goals. Thus, personal attributes are often part and parcel of performance evaluations. However, if elections are becoming less about who should govern and more about how we should be governed, then there is less cause for voters to focus on their personal capabilities and background. For example, in 1952 Eisenhower said he would go to Korea but never said what he would do there or how he might bring the Korean War to a successful conclusion. Therefore, the focus of the electorate was naturally on his personal military experience and leadership skills. In contrast, by 2004 and 2008 the debate centered more on what should specifically be done regarding the war in Iraq. With the focus now on what the candidates are likely to do rather than how capable they are, there is less reason to expect people to be commenting on the candidate's personal attributes.
Figure 1 confirms that fewer people now comment on presidential candidate attributes than in the past. Between 1952 and 1980, the typical survey found that 80% of respondents said something about candidate attributes. After Ronald Reagan came into office, however, there was a clear drop in the salience of...