Paul Goren has written a pathbreaking book about voter intelligence. In answering the paramount question—Are American voters fools?—he sides with the optimists. Philip Converse, the leading scholar articulating the more pessimistic view, in The American Voter (1960), recently returned to the “problem” surrounding voter competence. Goren, in his first book, takes this “voter competence” phrase for his title (p. 3). He argues that “[n]early everyone acquires and subsequently uses policy principles to guide candidate choice,” so going directly against the notion that the typical American voter does not know much (p. 7). Instead, the voter makes choices from core policy principles: limited government, traditional morality, and military strength. In this review, I discuss such policy principles and how they differ from liberal–conservative ideology and issue attitudes. Then, I consider the three faces of policy voting, as well as the conditioning role of political sophistication. Next, I evaluate measures of the core variables, finally assessing how they relate to presidential vote choice.

In the early chapters, the author attempts to justify his focus on the three policy “cleavages,” as he calls them, regarding the role of government, morality, and the military. With Chapter 2, a helpful dissection of their historical evolution is offered up. Aware of possible criticism coming from the omission of the racial cleavage, he contends that it became less important over the election period under study (1988–2008). This exclusion perhaps needs more justification, in light of the apparent racial cost suffered by Barack Obama in his 2008 victory, a cost calculated in a number of published papers. But, be that as it may, the potential importance of the three cleavages he does include still merit evaluation.

Goren identifies “the three faces of policy voting”: liberal–conservative attitudes, issue preferences, and policy principles (Chapter 4). Policy principles are more abstract than single issues, and in some ways represent a cluster of issues. However, policy principles are less abstract than liberal–conservative attitudes, which demand a degree of constraint that few voters possess. For example, work I have done shows that in the 2004 American National Election Study, 24% of the sample could not place themselves on a simple liberal–conservative scale, and of those who did place themselves, about one-third put themselves in the neutral “middle-of-the-road” category. Thus, only slightly more than half the respondents (51%) selected into an ideological label. This is the sort of evidence often put forward to support the notion that the mass of voters are not engaged in policy debates.

The limited identification with liberal or conservative labels squares with the finding, subscribed to by Goren, that the American electorate, on average, has little political knowledge. Only among the politically sophisticated do knowledge levels become more impressive. This leads the author to theorize that, in fact, the first two faces of policy voting—liberal–conservative ideology and issue attitudes—are conditional on political sophistication; however, policy principles are not. In general, the requirements for policy voting are demanding. Goren, in a slight modification of the criteria for issue voting presented in The American Voter, labels these attitudinal requirements “Availability” (e.g., awareness of the issue), “Centrality” (e.g., holds an opinion on the issue), and “Position-Matching” (e.g., sees one party as closer on the issue). He claims that it is mostly the politically sophisticated that can meet these standards, estimating “that liberal–conservative voting eludes the remaining 60–70%” of the public on election day (p. 78). Moving on to the less demanding notion of simple issue voting, he still concludes that little of that occurs in the mass public: “most potential voters do not hold enduring evaluations on most issues” (p. 81).

To find policy voting in the general electorate, Goren says we must look at its third face—policy principles voting, which taps into attitudes well in place before the campaign begins. These policy principles represent more general attitudes. Take the principle of limited government, held to constrain lesser relevant attitudes, for example, minimun wage, health care, and tax cuts. Put another way, limited government is a latent varible reflecting lesser observed elements in the policy space. To assess these concepts empirically, the measurement model of confirmatory factor analysis (CFA) is used, where the selected survey items are viewed as responses from a latent evaluative disposition. The data analyzed are from the American National Election Study, 1988–2008, and multiple items are employed to measure each policy principle. For example, limited government uses responses to questions concerning how much federal spending there should be on programs such as welfare, food stamps, the poor, the homeless, social security, and child care.

The CFA measurement model shows that the items load well on the latent factors representing the policy principles, with correlations of around .6 or so, indicating that these attitudes are well structured. Further, and tellingly, panel analysis demonstrates that, over time, these policy principles are almost as stable as party identification. In other words, they appear to be real attitudes. Last but not least, these real attitudes are not obtained at the sacrifice of large portions of, presumably, the politically uninvolved. As Goren summarizes, “with one exception, levels of non-response are low for the items used,” and they are not related to political sophistication (p. 118). In other words, most American citizens seem to have these policy principles in their heads, available to them for making political assessments.

The key question, then, “is whether policy principles shape the presidential vote,” and in particular, whether that shaping is “to a comparable degree for politically sophisticated and unsophisticated individuals” (p. 195). To answer, Goren tests his principled-voter theory against three rival theories: the partisan voter, the retrospective voter, and the ideological voter. (He omits other theories of voting, perhaps in the interests of parsimony. The foremost omission is sociological theory, à la Paul Lazarsfeld, but others could be mentioned, such as communications or leadership theories). Estimation begins with a baseline logistic regression model (for each of the six elections) that predicts presidential vote from party identification, retrospective evaluations, liberal–conservative identification, and political sophistication. He then introduces the three policy principles variables, followed finally by their introduction as interacted with political sophistication. The coefficients for the policy principle variables are generally statistically significant and often substantively strong (e.g., in 1988, the probability of voting for Michael Dukakis is only .08 if you want to limit government, but is .61 if you want to expand government).

Thus, Goren establishes the influence of policy principles on presidential vote choice. Further, the policy principle effects are not conditioned by political sophistication (unlike the liberal–conservative effects, which are so conditioned). That is, pretty much everybody votes on the basis of his or her policy principles, and equally so regardless of degree of political sophistication. American voters are not fools. Instead, they are competent, using political principles they hold dear in order to decide whom to favor. Bear in mind that if the standard of competence were liberal–conservative identification, or appreciation of the immediate issues at stake, very few voters would pass the test; only the elite few could meet those standards. However, principled voting does not depend on a high degree of political sophistication. It is easy to grasp, and almost everyone does. The masses manage to act on political principle to the same extent as the elites do, as least with regard to the vote.

What more should be said? Goren has shown the importance of political principles for determining vote choice. But he is not the only political scientist to raise concerns over the impact of principles, or values, or predispositions, as they are sometimes called. Stanley Feldman, Jon Hurwitz and Mark Peffley, and Warren E. Miller and J. Merrill Shanks, for example, have all written about their role for the vote, and in ways different from that of Goren. Hence, a task ahead is to sort out what values—and what value measures—should be included in the future specifications of the fundamental equation explaining American voter choice.