Throughout the mid-1980s, empirical state-level initiative studies were sufficiently rare that David Magleby's now classic book (Direct Legislation: Voting on Ballot Propositions in the United States, 1984) addressed topics as varied as the status of direct democracy laws, the impact of various cues on voting choices, and the readability of ballot pamphlets. Research on the initiative process subsequently burgeoned, and direct democracy has become one of the most widely considered aspects of state politics. Quantitative studies have focused especially on the substantive impact of the initiative process on public policy, as well as on the extent to which campaign contributions influence electoral outcomes.

Yet as Daniel Smith and Caroline Tolbert emphasize, another aspect of the initiative process has received considerably less attention: the extent to which ballot-measure decisions educate citizens about the workings of democracy and interest them in democratic processes. This absence is notable because extant arguments about the virtues of direct democracy commonly stress its educative role. While a few recent articles have addressed selected topics in this area, there has been no book-length study. In short, Smith and Tolbert have identified an important niche to fill.

After a brief overview, the authors turn to an explication of the Progressive Era claims about the educative impact of direct democracy, or lack thereof. They mine a rich array of sources, including the flowering of works on ballot measures published in the then-young political science academic journals. Many Progressive Era scholars believed that adopting the initiative would bolster election turnout, enhance civic engagement, increase political knowledge, and promote a sense of governmental responsiveness. Some also believed that adoption of the initiative would curb the power of party bosses. Yet skeptical voices also were heard, with some commentators questioning the educative value of direct democracy, given such factors as the complexity of the initiative process, continued role of special interests, and potential for political parties to hijack ballot measures.

Chapters 2–4 of Educated by Initiative constitute its analytical heart. In these chapters, Smith and Tolbert examine quantitative evidence bearing on key claims of Progressive Era (and modern) initiative proponents. The authors compare citizens in the roughly half of American states that permit ballot initiatives with citizens in the other states. Consistent with much recent work on direct democracy, they consider the frequency of initiative use, rather than simply the presence of the direct democracy mechanism, recognizing that some states (e.g., Oregon) use the initiative much more frequently than others (e.g., Illinois). They draw on aggregate data, such as turnout rates, and individual data from surveys such as the National Election Studies.

In general, Smith and Tolbert find support for the claims made by initiative proponents regarding the effects of direct democracy on electoral turnout, information about and interest in politics, and civic engagement. For example, they find that, controlling for other variables, states with more initiatives on the ballot have higher turnout, especially in midterm elections. Additionally, states with more initiatives rank more highly on Robert Putnam's social capital index, a notable measure of civic engagement. As the authors indicate in their concluding chapter, initiative advocates will “likely be heartened” by these findings.

However, subsequent chapters find less support for a couple of other Progressive Era claims, notably the idea that the presence of the initiative undermines partisan politics. In fact, Smith and Tolbert amass evidence that initiatives have been used for partisan purposes in such states as California and Colorado. Furthermore, in contrast to Magleby, they find that party identification exercises a large impact on voting for a wide range of ballot measures.

A notable feature of the book is the clarity of presentation. Smith and Tolbert indicate in their preface that they intend to reach practitioners as well as academics; I believe their strong organization of material and stylistic choices should help them to do so. For example, they avoid cluttering the main text with regression tables, instead highlighting key statistical findings in short boxes while allowing the true connoisseurs of statistical analysis to turn to the appendices.

Although the overall project is commendable, I have two related concerns. First, the authors' data boxes and summary statements, if not the fine detail of the text, appear in places to overstate the evidence for initiatives' beneficial effects. Consider the results of nine regression analyses (contained in Tables A.5 to A.6) assessing the impact of the initiative on political knowledge, political interest, and political discussion in 1996, 1998, and 2000, based on NES surveys. While eight of the nine coefficients for the initiative variable have the “right sign,” only one is statistically significant at the .05 level, and only four are statistically significant at the .10 level. By contrast, the claim about the positive impact of initiatives on voting obtains more consistent statistical support.

Second, the authors could give more consideration to how initiative states might differ from noninitiative states in unmeasured ways. For example, noninitiative states are disproportionately clustered in the group of states with a “traditional” political culture in Daniel Elazar's widely used classification scheme; most of these states are in the South. Previous literature has shown that states with a traditional political culture are less responsive to differences in public opinion (Robert S. Erikson, Gerald C. Wright, and John P. McIver, Statehouse Democracy, 1993, Chapter 7). In this regard, it is notable that Smith and Tolbert include a southern state control variable for many of the regression analyses related to whether people vote, but not for the analyses of the other potential educative effects of direct democracy.

Thus, more refinements in the data analysis may be needed. Nevertheless, Smith and Tolbert have pursued a valuable area of research and have provided a solid foundation for further work. They have helped to specify and examine noninstrumental effects of direct democracy that may ultimately prove as important as any instrumental ones. For this we owe them a debt.