1
The Strategic Landscape How to Find It, How to Read It, What It Reveals
To begin at the end: when a fresh methodological approach is applied to the evolution of American politics over the past quarter-century by way of continuing and consistent survey items, it becomes evident that:
⢠There was an old order to American politics in which the political landscape was organized principally by economic values, though by the time these data tap into this order in the 1980s, cultural values had begun to show substantial nascent influence as well. In that world, there was one political party, the Republicans, that was united by economic conservatism but remained heterogeneous on culture, while there was another political party, the Democrats, that, with noteworthy exceptions, was united by cultural liberalism but remained heterogeneous on economics.
⢠A quarter-century later, there is a new order to American politics in which the political landscape is organized simultaneously by both economic and cultural values. In this world, economics has actually gained aligning power, but culture has gained more rapidly, to the point where the landscape is now organized symmetrically across the full range of both ideological continuums. The modal Republican is thus a joint economic and cultural conservative, just as the modal Democrat is now a joint economic and cultural liberal.
⢠This is, by extension, a world in which political values play a greater role in aligning the vote than they once did. Yet evolution toward this world has been uneven and mottled in both a minor and a major way. In the minor way, the two elections of the 1990s in which Ross Perot was an independent candidate prove to be the transitional period, though not in the manner usually adduced. For Perot did not so much elicit, or even crystallize, this transition as he provided a parking place for voters who were breaking loose from the old order and seeking their proper place in the new.
⢠In the larger mottling influence, one that is still very much with us, different social groups responded to national changeâand shaped itâin distinctive ways. Few remained impervious to national trends, though there were groups whose members resisted and even groups whose members actually moved away. Most, however, integrated national trends with the social backgrounds and political values that characterized the group, beginning and ending at different places on the ideological landscape and sometimes changing intergroup relationships as well.
⢠Implicitly, as a result, some established ways of thinking about American politicsâways that are theoretically revered by social scientists and compulsively applied by campaign strategistsâsuffered a serious loss of plausibility. The methodological and conceptual innovations used here are in this sense an effort to create a framework that might substitute for these approaches, while bringing theorists and practitioners, kicking and screaming, at least somewhat closer together.
The Search for a Political Landscape
This project necessarily begins with a methodologically distinctive approach. Most critically, this revised methodology does not make the unnecessaryâeven arbitraryâassumptions that most research on public preferences and voting behavior reflexively does, and indeed must. Instead, two great attitudinal domains, involving public preferences in the realms of economics and culture, are allowed to produce vote distributions stochastically. Individuals are still conceived as having an intrinsically unobservable position in this two-dimensional attitude space. Yet the latter can nevertheless be expressed formally in percentiles relative to the general population, and because the statistical laws that determine relations between attitude position and item response can be estimated from observable data, it is possible to compute a probability distribution of those attitudes for every individual.
So, rather than simply assigning scores plus an error term to these individuals, this approach works from attitudinal distributions that are then cumulated, for a particular group or for the nation as a whole, to yield a political landscape. Even at the purely methodological level, this approach works to unite social scientists and campaign strategists.1 The former insist on seeing statistical normality everywhere. The latter just try to adapt to whatever they find on the ground. A map of the ideological terrain escapes the imposed preferences of the former while forcing systematic description on the latter. This is done not merely in the absence of conventional assumptions but in a way that credibly demonstrates how concepts that are not directly observableâthe latent valuesâcan explain an outcome that everyone recognizes, namely the vote.
The project likewise begins with two conceptually distinctive approaches. In the first, it treats voting behavior not as a bivariate outcome, Democratic or Republican, but always as a tripartite resultâDemocratic voting, Republican voting, or Non votingâand sometimes, when there is a credible third candidate, as a quadripartite result. This means that the same things do not have to be going on, as simple reciprocals, among Democratic and Republican voters. Analytic results are not simply imposed through this assumption, as they so often are in conventional vote analysis. In consequence, the social coalitions dominating the Republican and Democratic parties are not only free to live in different places on the ideological landscape; they can even be constructed from different policy realms, mixed differentially.
In the second conceptually distinctive approach, this way of proceeding pays much-expanded attention to the distribution of political values among partisan votersâto their actual density within the voting public; that is, to how many people actually stand whereârather than just to relationships to this vote. It is a commonplace in teaching political analysis to emphasize that the nature of an underlying distribution can be more important, depending on the question, than any statistical relationship within that distribution. Yet most research (and here, not excepting research on voting behavior) then pays little further attention to this possibility.
A focus on campaign strategy, however, brings actual distributions insistently to the fore, sometimes to theoretically provocative effect. For example, voting relationships, much loved by social scientists, are commonly interpreted to counsel a move toward the ideological center. Yet density maps frequently suggest that this is foolish and sometimes even counsel a move away. As often as not, it is these distributions that shape the strategic thinking of real campaigns, so that this becomes yet another incarnation of the enduring differences between social scientists and campaign strategists.
In any case, the end result is a picture of the strategic environment for electoral politics in the United States over the last twenty-five years. This is a picture of American ideological terrain unlike standard two-dimensional (not to mention one-dimensional!) portraits. It is a picture that brings into focus the strategic choices facing both major political parties, in a world that is nevertheless constantly evolving. As above, there is a fully national aspect to this picture, where what was once overwhelmingly a politics of economic preferences has evolved into a politics in which both economics and culture are normally in play, across the full ideological spectrum and in a more or less symmetric fashion.
Nevertheless, there are multiple and various group aspects to this national story as well, since social groups do not simply recapitulate national trends within their individual memberships. At a minimum, a composite national picture leaves room for (and effectively requires) multiple responses, as diverse groups respond to national developments in ways that reflect the social backgrounds and political values of group members, rather than of some abstract national median. Beyond that, social groups differ in the extent to which group membership itselfâsocial identityâhas the power to shape the role of political values. At the extremes, some groups have lost this shaping power, while others have remained more or less impervious to external developments. And most, of course, are a mix of group-based and nationwide influences.
A different way to summarize what has happened to the American political landscape over the last twenty-five years is occasioned by a shift of focus from voting relationships to density maps, both within social groups and for the nation as a whole. Two developments stand out. In the first, Republican and Democratic voting blocs within most social groups have actually pulled farther apart in ideological terms. From one side, this separation was a development fueled especially by the national rise of cultural values as a voting influence. From the other side, this separation has been exaggerated in a partially autonomous fashion by the political parties themselves, most especially by changes inside the Republican Party.
In the second, closely related, development, the dominant coalitions inside these parties have in turn converted the older pattern under which these coalitions were previously constructedâwith Republicans anchored by economic preferences and Democrats anchored by cultural preferencesâinto the modern pattern of joint ideological polarization. Now, the Republican coalition is anchored by individuals who are both economic and cultural conservatives, just as the Democratic coalition is anchored by individuals who are both economic and cultural liberals. The Democratic coalition does, however, contain a secondary concentration of supporters that has no counterpart on the Republican side and that presents special advantages and special challenges for strategic management.
Along the way, a popular theoretical approach to parties and coalitionsâone that emphasizes median voters and partisan convergenceâfares very badly. The real world has been insistently unkind to this otherwise-elegant framework, and we have no desire to kick any theory that is already down. Yet a focus on voting relationships to economic or cultural values individually, on voting relationships to economic and cultural values jointly, but most especially on the density maps associated with both relationships, does, in passing, offer two reasons for this less-than-sterling empirical performance.
The first merely emphasizes countervailing influences. When mapped onto an ideological landscape for electoral politics, Republican and Democratic voting blocs are disproportionately headquartered in terrain that is more or less inherently hostile to moves toward the ideological center. And those living there tend to be advantaged in expressing their hostility. A second countervailing effect is more abstractly provocative. Stratified by political values, by group memberships, and by partisan voting blocs, the American political landscape can beâit is not always, but it always can beâeffectively empty at the center. Regardless, it is more often voters at the ideological extremes, not voters in the middle, who can be convinced to abandon their usual partisan choice, albeit normally in favor of some third alternative rather than of the other major party.
At a less grand level, there are numerous smaller but still-consequential aspects of the American political landscape that heave into view in the process of using these methodological approaches to uncover those larger stories. Thus there are social groups where group identity is nearly everything and political values nearly nothing. Black Americans are the outstanding example. Conversely, there are social groups where political values are nearly everything and group identity nearly nothing. Men and women do differ in their partisan propensities, yet men and women holding identical values vote in effectively identical ways. There are groups that have not surrendered the clear dominance of economic values in their voting behaviorâhigh school dropouts are the best exampleâjust as there are groups that have enshrined cultural values to the near-exclusion of economicsâof which postgraduates are the archetypal case.
The presence of identical items across a quarter-centuryâthis is a great virtue of the Pew Values surveys, which contribute the central measures hereâallows the further pursuit of change in these relationships. Hispanic Americans provide the greatest single instance, in which a social group that was largely impervious to the national link between political values and voting behavior moved into (its own version of) full conformity with national relationships, even as these latter were themselves changing impressively. Yet other principles of social division have also reordered their group-related behavior in major ways: the augmented Republicanism of the Evangelical Protestants, at every point from the far right to the far left; the augmented Democracy of the non-Christians, likewise at every point; and the collapse of old differences between Catholics and Mainstream Protestants in the ideological middle; all these provide obvious and insistent examples. Finally, some social groups remain remarkable for the opposite reason. Thus social class as measured by family income has managed to foster strong relationships to political values while keeping intragroup relationships remarkably invariant.
The Organization of the Book
In order to unpack this comprehensive story, Chapter 2 sets up the analysis, introducing the theory and the method by which opinion data will be handled throughout. Chapter 3 isolates the contribution of various demographic categoriesâsocial class, race and ethnicity, religious background, and domestic rolesâto the economic and cultural values associated with them. Chapter 4 introduces the social groups that provide a structure to these political values, in ways that can potentially be linked to voting behavior. Chapter 5 then returns to method, setting up the analytic devices by which these links can be teased out and displayed.
Chapter 6 applies these devices to the year with the earliest and richest relevant data, in order to elaborate a benchmark survey of national and group relationships among social backgrounds, political values, and voting behavior. Chapter 7 traces the evolution of the national picture from 1984 through 2008, concentrating on the opening and closing surveys in this series but stopping along the way to capture important intervening developments. Chapter 8 traces the group-based side of this same evolution, seeking the role of social groups in placing individual members within national trends. And the Conclusion returns to the grand strategic questionsâeven dilemmasâthat maps of this ideological landscape inescapably pose.
We have tried to make these elements fit together seamlessly. To that end, theoretical concepts that would already have been recognizable at the turn of the twentieth centuryâsocial backgrounds, political values, and voting behaviorâhave been reinterpreted through methodological approaches that are only just coming into fashion at the turn of the twenty-first. The idea is to keep two things at the forefront of the analysis: (1) who people areâthat is, their social backgrounds, presumably as these reflect common experiences and encourage common interests; and (2) what people wantâthat is, their political values, as these are captured by two great underlying dimensions of public policy and reflected in the policy conflicts that come to define the substance of politics. Together, social backgrounds and political values create the ideological landscape upon which voting behavior occurs and through which the resulting voter distributions come to constrain el...