If one were to judge this book by its cover, then David RePass’s Listening to the American Voter delivers what is promised. For the author, “listening to the American voter” means relying on survey respondents’ answers to open-ended questions. This book relies entirely on the 15 ANES presidential-year surveys conducted between 1960 and 2016, and the recorded answers to two sets of questions that have always been part of those surveys: questions about what respondents say they like and dislike about the major candidates running in the presidential election, and a separate set of questions about the most important problems facing the country, plus the follow-up question of which party is better able to handle the most important problem.

The strength of this book, and the reasons you should buy it, come from chapter 3 (“Measuring Issues”), responses to the most important problem questions; and chapter 4 (“Images of the Candidates”), responses to the candidate likes and dislikes questions. Older readers will remember—and may have even voted in—all these elections. For them (well, us), these chapters will read like a stroll down memory lane. But for younger readers interested in American politics who have no direct memory of many of those campaigns, these chapters capture in a very few pages the guts of each of those elections—why they seemed important at the time, and how political pundits look back and think about them now.

Chapter 3 reports data on the “issue publics” that existed in each of these elections, operationally defined as being mentioned as an important problem facing the country by at least 7% of the public, and as the most important problem by at least 3%. We do not have to accept RePass’s assertion that these publics include all the voters with a real attitude toward this issue to find these data interesting. It is not the responses from any particular election but rather the entire historical overview provided by these data that I find fascinating. We see, for example, that the public was very concerned about the threat from communism until the Soviet Union collapsed in 1989, that civil rights and the urban riots were major issues during the 1960s, and that health care became a major issue in 1992 and has remained one ever since. Of course. But we also see some very surprising results, such as education being a major concern of the public from 1988 onward; did the press ever talk about education as a major issue in any of those elections? However, gay rights, gun control, and school prayer have often been major topics in the media covering American elections, but they have never made the list of the most important problems facing the country. Neither have abortion or taxes, issues of particular interest to special interest groups, but not to the general public.

To the extent that American presidential elections are candidate-centered events, a list of what survey respondents said they liked and disliked about the major presidential candidates, as reported in chapter 4, is a list of the most important factors in each of these campaigns. Looking at all these elections together, candidate policy positions was the broad category of responses most frequently mentioned, but personality (character, leadership, sincerity), past performance in office, candidate-group linkages, and straight-out positive or negative affect were also frequently cited. As was the case with the most important problems, what does not (often) make the list is as interesting as what does. Abortion, ideology, and taxes were rarely mentioned as reasons to vote for or against a presidential candidate. To quote the author on a point he makes several times throughout the book, “Analysts cannot assume that the prominence of an issue in the political elite’s dialog will have an impact on voters’ decisions” (p. 105).

Together, these two chapters comprise almost half the book, and alone they make it an interesting read. Early on the author describes the book as presenting a method for accurately measuring the mandate associated with a presidential election, and I can see the argument that responses to open-ended questions such as those considered here come pretty close to that. Indeed, I wish RePass had returned to this point throughout the book. Unfortunately, however, the book tries to do much more than simply report responses to these open-ended questions, and here, I believe, it comes up short.

My concerns are essentially twofold. RePass claims that the book is examining “what was in the forefront of voters’ minds as they went into the voting booth” (p. 2). But that is not what these data are. What we have in these open-ended responses is what was in the forefront of the respondents’ minds when the friendly pollster happened to knock on their front door. Anyone familiar with John Zaller’s argument (The Nature and Origins of Mass Opinion, 1992) that survey responses are a random sample from a list of considerations that come to mind when a question is asked, or with Lodge and colleagues’ (Milton Lodge, Kathleen McGraw, and Patrick Stroh, American Political Science Review, 1989) impression-driven model of candidate evaluation that argues that memory has little to do with actual information exposure and the determinates of the vote choice, would question this assertion. Open-ended questions provide some information that close-ended questions do not, but they also include their own set of unique problems that should not be ignored.

My second major concern has to do with the multiple regression model that, RePass claims, almost perfectly explains the vote decision across these various election years. This model includes only four predictors: party identification (the familiar close-ended question), attitudes toward the Democratic and Republican candidates (the number of reported likes minus the number of reported dislikes for each candidate), and beliefs about which party is better able to handle what each respondent believed was the most important problem facing the country. I firmly believe that party identification is learned early in life and is therefore generally exogenous to any vote decision, but I have serious endogeneity reservations about the other three “predictors.” The open-ended questions about the major candidates are always asked in the preelection survey, but by then large majorities of the respondents to the ANES surveys have already decided how they are going to vote. Psychologists argue that people typically have very little insight into the causes of their own behavior. What we learn from these responses are some combination of the actual reasons that determined the vote decision, and plausible justifications for a decision that was made for some other reason. Responses to the most important problem questions are even more problematic. After 1968 they have almost always been asked in the postelection survey—that is, after the election has concluded. I cannot believe that many respondents after the election reported that they voted for candidate X, but they believe party Y is best able to handle the most important problem facing the country.

In conclusion, if you are looking for cutting-edge research on voter decision making, I would look elsewhere. But if you want to know what American voters were thinking about over the past six decades of American presidential elections, I strongly recommend Listening to the American Voter.