The determinants of the voters' decision in direct democratic votes have received growing attention over the last 15 years. The reasons for this are mainly twofold. First, there has been a sharp increase in the number of popular votes worldwide. In Europe, the successive waves of referenda on the European Union, up to the recent rejection of the EU constitution in France and in the Netherlands, have put the issue of direct democratic choices on the top of both the scientific and political agendas. Second, we have witnessed important theoretical developments in the study of opinion formation and political behavior (e.g., the contributions of social and political psychology with respect to the role of cognitive strategies, or the contributions of communication studies with respect to media and campaign effects).
As the country with the most far-reaching experience with direct legislation and referendum campaigns, Switzerland constitutes a unique “laboratory” for the analysis of direct democratic choices. Exploiting this “comparative advantage,” Hanspeter Kriesi delivers a pathbreaking study. At the theoretical level, he uses the long-lasting dispute between the elitist and the pluralist conceptions of direct democracy as a starting point, and argues for a third, “realistic” conception that sheds light on the conditions under which “ordinary citizens are able to arrive at an enlightened political judgment and to make a reasonable choice” (p. 8). These conditions depend on the political elites and on citizens themselves.
First, the realistic conception is premised on the belief that—by delivering political messages, forming coalitions, and mobilizing support during a referendum campaign—the elites play a leading role in the individual process of opinion formation. Second, the author builds on the dual-process theories and distinguishes between two paths of individual opinion formation: a heuristic path, based on shortcuts and simplifying strategies, and a systematic path, based on the arguments provided by the elite. While the former has given rise to an extensive literature, in the United States and abroad, empirical tests of the impact of argument-based reasoning on the voters' decision are scarce. Kriesi's study fills a gap in that respect.
Methodologically, the author also innovates by relying extensively on hierarchical (two-level) models, which enables him to evaluate how contextual, project-related variables and individual variables interact, and jointly influence the voters' choice. Empirically, Kriesi analyzes an impressive data set of postelection surveys covering nearly 20 years (1981–99) of direct legislation in several policy domains (148 popular votes altogether). This survey data provides systematic information about respondents' sociodemographic background and political attitudes. At the contextual level, the author takes into account both the intrinsic characteristics of the proposals submitted to the voters (e.g., their institutional form, familiarity, and salience) and the characteristics of the related referendum campaign (e.g., the intensity and direction of the campaign, based on newspaper ads, and the level and type of conflict among the elite, based on parties' voting recommendations).
In the first part of the book, Kriesi introduces the reader to these context-related variables and convincingly shows that they strongly structure voters' participation, their level of political information on the issues at stake, and their voting decisions. Thus, the intensity of the referendum campaign and the familiarity of the proposal submitted to the voters significantly increase participation, in interaction with individual factors. Similarly, intense campaigns and familiar projects lead to higher levels of political awareness among individuals. Higher awareness, in turn, increases the likelihood of taking part in the vote, and at the same time reduces the risk of unreasonable decisions. Finally, and in sharp contrast with the view that one could “buy” a popular vote, Kriesi's study demonstrates that once the type of coalition among the party elite is controlled for, campaign spending no longer plays a role for the outcome of a popular vote. Overall then, the analysis confirms that while direct democratic votes necessarily introduce some element of uncertainty for the party elite, the latter still holds fair control over (the quality of) the individual decision.
The core—and major contribution—of the book stems from the analysis of the heuristic and systematic paths of opinion formation (Chapters 6 to 8). The resort to two-level models appears as a successful strategy and leads to several interesting—albeit in some cases fairly complex—findings. Among the three different heuristics included in the analysis (status quo, trust in government, and partisan cues), the last two are at first sight decisive. For example, when the elite is divided, reliance on parties' voting recommendations appears as a widespread strategy, and this even among unaware voters. However, the effects of the partisan and trust heuristics turn out to be partly spurious: They disappear when systematic strategies are controlled for. Indeed, according to Kriesi's main—and most intriguing—finding, argument-based, systematic strategies are of overriding importance, and clearly outweigh heuristic cues. From this and other findings, Kriesi derives normative implications, ending up with a nuanced, but overall optimistic, conclusion regarding the quality of direct democratic procedures.
While this conclusion is certainly in line with the results, the empirical tests on which it is based raise some questions. To be sure, Kriesi's technical work is of first-rate quality and the author has made the best possible use of the data. But as he himself acknowledges, the data are far from perfect. One problem is that the questions and response categories included in the postelection surveys have varied over time, which affects the measurement of some central variables. Thus, political awareness is mostly based on two crude measures of factual knowledge. Similarly, the measure of the trust heuristic as it is formulated in the survey presumably raises a problem of endogeneity. Finally, and perhaps more importantly, the indicator for systematic reasoning is based on the respondents' agreement with some arguments put forward by the yes and no camps during the referendum campaign. These arguments are always asked at the end of the interview, that is, well after the question on the voting choice. Consequently, one cannot be sure whether a given argument really served as a voting motive, or whether respondents simply used the argument to justify their vote a posteriori—perhaps even without having heard about it before. If the latter holds, then the direction of causality is reversed: It does not go from the argument to the vote, as it is assumed by Kriesi, but from the vote to the argument. This would obviously render the test of the impact of argument-based strategies problematic.
Even with this qualification in mind, I warmly recommend this thorough and highly enlightening book, which provides several convincing answers to the ever more crucial issue of direct democratic choices, and offers an important and stimulating contribution to the ongoing debate between advocates and opponents of direct democracy.
