Progress in political analysis is anchored in both theory and observation, and Russell J. Dalton and Christopher J. Anderson's edited volume provides a particularly vivid example that is worthy of both consideration and emulation. The focus of their collective effort is on 1) institutional and compositional variation across modern democracies, 2) the consequences that arise for the incentives and behavior of individual citizens, and 3) the implications for politics in the aggregate—the level at which politics takes on significance and meaning.
The authors are thus engaged in a frontal assault on what Heinz Eulau would have called the micro-macro problem in political analysis (Politics, Self, and Society, 1986). This problem is based on the recognition that a single level of analysis—either micro or macro—necessarily misses the point of democratic politics. The real action occurs at the intersection between individuals and their surroundings, and relevant analyses must thus address the interdependence between the two. Hence, the authors join a chorus of influential voices who have produced a series of distinguished analyses—V. O. Key in Southern Politics in State and Nation (1949); Philip Converse in “Of Time and Partisan Stability” (Comparative Political Studies 2 [July 1969]: 139–71); Warren Miller in “One-Party Politics and the Voter” (American Political Science Review 50 [September 1956]: 707–25); and Adam Przeworski and Henry Teune in The Logic of Comparative Social Inquiry (1971).
The wonderful element of big problems is that they never cease to pose new questions and, hence, new and revealing analyses. The authors of Citizens, Context, and Choice take advantage of a singularly important data-collection effort—the Comparative Study of Electoral Systems—that involves collaboration among individuals and research teams carrying out national election studies across a wide array of the world's democratic political systems. These cooperating research teams adopted a common core of shared survey questions that was subsequently asked of all respondents in every country study. Needless to say, the implementation of such a research plan involved a heroic effort by scholarly entrepreneurs in a great many countries, and their efforts have produced a game changer, not only in the study of comparative politics but also in political analysis more generally.
These efforts generate no foregone conclusions. A comparative understanding of citizens in the contexts of particular democratic systems produces analytic results that sometimes confirm the importance of context for individual behavior and sometimes refute it. For example, the chapter by André Blais and Thomas Gschwend addresses strategic party defection and provides little support for the common expectation that there should be higher levels of strategic defection in more disproportional systems or, alternatively, that defection should be higher in systems with more party options. G. Bingham Powell's chapter, in contrast, confirms expectations regarding the trade-offs involved in more or less polarized party systems: Greater polarization encourages individual citizen engagement, while at the same time making it more difficult to achieve congruence between citizens and policymakers.
Other equally important results demonstrate the mutually contingent effects of political system characteristics and individual attributes and predispositions. At numerous points along the way, the authors of the various chapters show that the effects of context and setting vary across individual characteristics, just as the effects of individual characteristics vary across contexts. In these ways, the authors teach us a great deal about the mechanisms of citizen engagement, as well as the ways in which effects arising due to citizen traits are contingent on particular political and institutional settings.
Indeed, the authors show that indirect and contingent effects arising due to context and setting are more common than simple direct effects. In their summary, the editors suggest that “the direct effects that are most easily observed in aggregate statistical analyses are only a small part of contextual influences. What is more, they do not provide tests of the causal mechanisms that we often assume undergird these effects. As a result,” they conclude, “past aggregate analyses of direct effects in all likelihood have produced a biased and limited assessment of the importance of context. This is because the indirect and contingent effects of context appear to be more important in explaining individual level behavior than do direct contextual effects” (p. 246). In other words, aggregate analyses are typically a misleading substitute for a thoroughgoing multilevel analysis that imbeds individuals within the settings where they encounter politics.
We should not be surprised! In responding to the challenges posed by ecological fallacies more than 50 years ago, Leo Goodman (“Ecological Regressions and Behavior of Individuals,” American Sociological Review 18 [December 1953]: 663–64; “Some Alternatives to Ecological Correlation,” American Journal of Sociology 64 [May 1959]: 610–625) warned us that simple aggregate analyses carry the potential to be highly misleading when individual behavior varies systematically across settings. Electoral politics research took this lesson to heart, and aggregate analyses suddenly became quite rare in the study of elections, public opinion, and voting. Indeed, the advent of national election surveys created a new model with a focus on individual-level explanations for citizen engagement and behavior.
This watershed moment had the unfortunate consequence of failing to recognize the full extent of Goodman's argument. That is, ecological fallacies occur only if individual behavior is context specific, and hence context-free individual-level analyses fail to address the larger problem. Rather, the danger of ecological fallacies was replaced by the danger of individual-level fallacies. As the authors of this volume show, when citizen responses occur at the intersection between individual predisposition and collective circumstance, neither the individual characteristic nor the environmental setting provides an adequate account, and either considered alone might well generate misleading results.
No easy solution exists for these problems. Indeed, it becomes necessary to embrace the futility of both individual and environmental determinism in constructing accounts of democratic politics. When individual behavior varies systematically as a consequence of environmental contingencies, the analyst must set aside the possibility of an individual-level explanation that is context free. Similarly, when the response to environmental conditions varies across individual predispositions, a simple aggregate account will not provide adequate explanatory power.
The collaboration that produced this volume is an important part of an effort that carries the potential to remedy this problem. At the same time, there is nothing about the authors' effort that contradicts either psychological or economic models of political behavior. Indeed, this volume demonstrates important context-dependent effects arising as a consequence of both strategic behavior and psychologically based responses.
In conclusion, not only do theory and observation walk hand in hand in explaining democratic politics, but progress is often impeded by observational logjams as well. As long as students of democratic politics focused on aggregates, we failed to understand the crucial roles of individual citizens. When an observational revolution produced a strong tendency to focus on individuals, we typically failed to address the contextual contingencies operating on political engagement. The authors of this volume, along with the entrepreneurs who gave birth to the Comparative Study of Electoral Systems, have successfully removed yet another logjam inhibiting our understanding of democratic politics.
