As Election Day 2004 approaches, there will undoubtedly be a lot of talk concerning which major party candidate is most personally appealing. A commonly held view is that presidential elections are largely personality contests, and that the candidate with the best-liked personality wins. If such a view is correct, then presidential elections are fairly similar to elections for class president in high school, albeit on a much larger scale. But is this really the case? Based on a careful analysis of national survey data from the last 11 presidential elections, this article concludes that such a view is unfounded.
Despite commonplace assertions from pundits that the most personally appealing candidate is bound to be victorious, the American National Election Studies data show this was not the case in a number of recent presidential election contests. Furthermore, personal popularity is shown to be a complex phenomenon, with a candidate who does well on one dimension, such as integrity, frequently not doing so well on another dimension, such as competence. In short, presidential elections are not simply personality contests.
While elections are not just about who has the most pleasing personality, individuals do matter in presidential elections. The presidency is a distinctly personal office, and which individuals are nominated by the major parties certainly influences the outcome. It has often been said that American presidential elections are candidate centered, and indeed they are. Television, direct mail, and now the Internet have freed candidates from reliance on political parties, thereby allowing campaigns to be run independently of party affiliation. Presidential nominees are self-starters who assemble their own personal campaign organizations, raise most of the funds needed to run in the primaries through these organizations, and set their own individual issue agendas. Presidential candidates of the 21st century stress their own specific issue stances, and citizens naturally view the major issues primarily through the perspective of the competing candidates.
Personalities surely matter somewhat, but what has become the core of candidate-centered politics in presidential campaigns has been the issues. In the early days of voting behavior research, presidential election issues typically centered on relatively stable images of the parties, such as which party would be for more government aid to provide jobs or involvement in foreign conflicts. In recent elections, however, the public has come to focus far more on the short-term issues defined by the candidates during the campaign. Because these short-term issues have come to be seen as shaping the party as well, presidential candidates and their parties are now evaluated far more similarly than in the early days of candidate-centered politics, when candidates staked out positions independent of their parties. Whereas once a candidate's chances of winning hinged greatly on the popularity of his party and the issues it stands for, now it is more the other way around in American presidential elections. This article reviews a wealth of historical data on presidential candidate popularity from 1952 to 2000 that illuminates these points.
Presidential Candidate Popularity, 1952-2000
Since 1952, the American National Election Studies have conducted a nationally representative survey of the U.S. population every four years concerning the presidential election. Respondents are asked a variety of questions in the fall prior to the election and then reinterviewed as soon as possible following the election. The data from these studies have long provided an invaluable resource for the study of U.S. presidential elections. In particular, these studies have consistently asked people to say what they liked and disliked about the major presidential candidates. The exact wording has been as follows:
Is there anything in particular about--that might make you want to
vote for him? …