Renewing with the Turner Tradition of Electoral Analysis

The Canadian Review or American Studies, Volume IO, No. 3, Winter 1979 Renewingwiththe TurnerTradition of Electoral Analysis J.A. Laponce Joel S1lby,Allan Bogue and William Flanigan, eds. The History of American Electoral Behavior. Princeton. New Jersey: Pnnceton University Press, 1978. 384 + ix pp. The use of aggregate statistics to analyze the behavior of individual electors belongs to a relatively old tradition dating back to the tum of the century and associated in the United States with the pioneer work of a few historians, notably Frederick Jackson Turner and his students- and associated in France with the name of the social geographer Andre Siegfried whose Tableau politique de la France de l'Ouest (I913) still awaits translation into English. Both Turner and Siegfried used the simple technique of comparative cartography to explain differences in party support by differences in social and cultural characteristics noted to occur within the same geographical boundaries-typically, the technique consisted in comparing the geographical patterns of party support to the geographical patterns of selected census characteristics. That tradition of cartographic aggregate analysis has, to this day, remained very much alive in France where Goguel, Dogan, Lancelot and Leroy-Ladurie, among others, have applied and refined the Siegfried approach in their search for correlation and causation. In the United States, on the contrary, the Turner technique fell into disuse after the advent of survey research, even before the oft-quoted methodological criticisms levied by W.S. Robinson in 1950against the use of areal aggregates as units of analysis when individuals were the unit of interest (a "sin" that has become known as the "ecological fallacy"). In the 195O'sand 60's the field of election studies thus became dominated, in the United States, by one discipline, political science, and by one research tool, the personal interview of randomly selected individuals. Not only was comparative cartography rejected as inadequate but, more generally, the rejection affected all forms of aggregate analysis of voting behavior, cartographic or not. Such rejection was backed by serious methodological 336 L. A. Laponce arguments. It may also have been particularly welcomed by a society that favored the individual over the group, the present over the past, and that may have felt, at the time of its sudden international dominance, emancipated from the constraints of geography. The 1970's marked a partial comeback of aggregate analysis to the field of American electoral behavior. This return is related to the increase in the number of students of American elections who became dissatisfied with the "a-historical" nature of the data offered them by survey research (the first large scale "academic" study of an electorate based on personal interviews was conducted by Lazarsfeld in 1940); it appears related also to the increasing number of "quantifiers" among American historians, quantifiers who found excellent time series in the data provided by census and election statistics covering nearly two centuries. It is likely also that European scholars - particularly Stein Rokkan - were instrumental in "reconverting" some of their American colleagues to the use of aggregate areal data. The rapprochement between political scientists and historians wishing to apply quantitative analysis to the study of pre-survey elections was powerfully helped by the Inter-University Consortium for Political and Social Research (ICPR) based at the University of Michigan. The ICPR-a joint organization of many American and a few Canadian departments of Political Science, which has for its main function the cleaning, storing and disseminating of machine readable data (typically but not exclusively election survey data)-decided, in the late 1960's, to create an historical archive. and built an extensive machine-readable file of census and election data. That file, available since the early l970's, has been widely used by both historians and political scientists. Among other institutions that have oriented historians to the use of quantitative analysis one should also mention the Committee for the Collection of the Basic Quantitative Data of American Political History set up by the American Historical Association, a committee that worked in close collaboration with the ICPR; one should mention, finally, the History Advisory Committee of the Mathematical Social Science Board, a committee that has been sponsoring, since 1969, interdisciplinary conferences that brought together...