Accueil Revues Revue Numéro Article

Revue d'Histoire des Sciences Humaines

2002/1 (no 6)

ALERTES EMAIL - REVUE Revue d'Histoire des Sciences Humaines

Votre alerte a bien été prise en compte.

Vous recevrez un email à chaque nouvelle parution d'un numéro de cette revue.


Article précédent Pages 65 - 94 Article suivant

World War II marked a turning point in the history of American sociology. Whereas World War I had been won thanks to American resources, World War II, it was argued, would be won on the psychological front. National morale – of American combat soldiers, of those left behind to man the home front, of the enemy and of occupied nations – were seen to be the key to victory. National morale was also sociology’s ticket of entry into the lucrative world of military support and federal contracts. Whereas psychology and economics had proven their applied value during World War I, sociology had yet to establish its policy relevance. From the perspective of the discipline, one of the lasting results of the wartime experience was a shift in the hierarchy of methodologies. Statistics moved from a relatively marginal, contested tool to the authoritative method against which all other methods and approaches had to define themselves [1][1]  Platt, 1996, 37 ; Wilner, 1985.. This shift has been documented through the changing content of textbooks, journal articles and other quantitative indices [2][2]  Platt, 1996, chap. 2.. It is also evident in a change in the exemplars – or classic texts – used to represent and define the field. This paper uses one such exemplar – The American Soldier – to explore the broader set of intellectual and institutional changes associated with the elevation of statistics to this new status. The interest of The American Soldier is that it documents the historical convergence of survey research, quantification and a new type of theorizing involving the definition of concepts in terms of measurement procedures and the elaboration of empirically testable hypotheses.


The basic theoretical premise of the paper is that formal statistical tools can be associated with a wide variety of different ontological models, types of theorizing, types of claims and forms of institutions [3][3] The problematic builds on Ian Hacking’s call for historians.... The specificity of the American sociological tradition lies, not in the authority of scientific, quantitative approaches per se, but in the particular style of reasoning that came to be associated with it. This style has been variously characterized as instrumental positivism [4][4]  Bryant, 1985, chap. 5. and abstracted empiricism [5][5]  Mills, 1959, chap. 3.. It involved, among other features : an individualistic or atomistic model of society, a variable approach to social explanation and a concern with induction, verification and incrementalism. While an analysis of the production and reception of The American Soldier cannot fully document these developments, it does provide insights into the content and context for this reorientation.


Three related processes came together around the production and reception of The American Soldier. These include the association of survey research with statistical analysis, such that after the war the two appeared as almost synonymous, the reorientation of sociology away from political and institutional analysis to the study of values, norms and social integration and finally the associated bid for government patronage and the claim that applied research could contribute to basic research. These developments were associated with a number of more general developments including the creation of the « behavioral sciences » combining cultural anthropology, social psychology and sociology, the establishment of a new type of interdisciplinary, university based research center and the use of government and in particular military contracts to fund basic research [6][6] For a review of US military support for the social....


In focusing on the reception of the book amongst sociologists I in no way wish to minimize the importance of other wartime projects or the consequences of these experiences for other disciplines. In many ways, psychology as a discipline was much more successful at capitalizing on the postwar opportunities than sociology. Similarly, social psychology flourished in the interdisciplinary institutional context of the 1950s and early sixties [7][7]  Sewell, 1989.. Given the choice of disciplinary bases, more social psychologists chose psychology over sociology. At the same time, survey research and the new social psychology more generally did not transform psychology as a discipline. Instead they were absorbed into a broader turn to experimentation that extended well beyond this new speciality [8][8] For discussions of the development of social psychology.... In contrast, The American Soldier was at the heart of a much more far-reaching transformation of sociology as an academic discipline, one which, I would argue, redefined the dominant methodological, theoretical and epistemological positions of the field.


This article begins with a brief description of the volumes making up The American Soldier and the place it occupies in the historiography on American sociology. It then traces the changing status of quantification in order to better establish the nature of the turning point associated with wartime research. The discussion identifies three distinct approaches to quantification, each of which contributed to the mainstream methodological tradition of the postwar period. These include the work of Franklin P. Giddings and his students, the work of Paul F. Lazarsfeld and the Bureau for Applied Social Research and the programmatic writings of George Lundberg. Having laid out this context, the analysis returns to The American Soldier and the research that produced the data informing the publication. The discussion explores the different models of sociological practice embodied in different volumes of this corpus. The first two volumes exemplified a new form of applied social research in the service of bureaucratic patrons (what Mills described as liberal practicality). A central issue amongst sociologists in this period involved the ability of applied research to contribute to pure or basic social science [9][9] For a more general discussion of sociologists formulations.... In the case of The American Soldier this criticism focused on the poverty of conceptual and theoretical reflection.


Volumes three and four can be read as a methodological response to these criticisms, while Continuities involved an attempt to valorize the corpus by bringing out the scientific, as opposed to policy, contributions of the work. This clear attempt to redefine the content of the project raises the question of why it was undertaken. The paper thus turns to the situation of sociology as a discipline in the 1940s. More specifically, it considers the impact of debates over the creation of the National Science Foundation on both the perceived relation of applied and pure science and on the scientific status of the social sciences. An examination of organizational changes in the discipline link this intellectual reorientation to changes in the center of elite sociology and the creation of interdisciplinary research centers in the behavioral sciences and area studies at leading universities. The paper concludes with a reflection on the consequences of The American Soldier and wartime research more generally for the reorientation of sociological theory in the late 1940s and 1950s.


From the perspective of the history of quantification, my paper argues for a broad approach to the history of statistical methods, one which focuses not only on formal techniques and tools, but also on the epistemological and institutional contexts in which these tools were applied. As such, the approach follows the path set by L. Daston and I. Hacking in their path-breaking work in the historical epistemology of statistics [10][10]  Daston, 1993 ; Hacking, 1992.. This work emphasizes the historicity of categories such as science and objectivity. It also underlines the need to situate formal tools and techniques in the broader set of scientific goals, epistemological principles and practices in which they were developed. This paper contributes to their project by examining the broader conditions supporting the marriage of large-scale survey research and quantification and their adoption by postwar sociology as the authoritative method.

I - The Book(s)


In 1949 Princeton University Press announced the upcoming publication of The American Soldier. The two volumes, they said, were « the record of one of the largest social science investigations ever made ». The book reported the findings of the Research Branch of the Information and Education Division of the War Department of the United States Army during World War II. The Research Branch had been established to provide the Army Command with information about the attitudes of soldiers. More specifically it was concerned with problems of morale and the relation of morale to military performance. According to the Press, the book represented « one of the most elaborate applications ever made of the new methods of objective study, which are revolutionizing social science research and taking the study of man out of the realm of guesswork and conjecture » [11][11] Cited by Lerner in Merton, Lazarsfeld, 1950, 217..


As this publicity indicates, The American Soldier was heralded as an example of a new approach to social science, one which combined large-scale attitude surveys with quantitative analysis to produce policy-relevant, scientific knowledge. The approach was presented as novel, scientific and practical. Early reviews of the two volume set re-enforced this public image. N.J. Demerath, writing in Social Forces, opened his review by announcing : « Here is a book. Not since Thomas and Znaniecki’s Polish Peasant has there been a socio-psychological work of such scope, imaginativeness, technical rigor and important results » [12][12]  Demerath, 1949, 87.. And John W. Riley Jr. described the book for the American Sociological Review as « undoubtedly one of the most significant publications in the social sciences during the last twenty years » [13][13]  Riley, 1949, 557.. Even critics accepted the association of The American Soldier with a new scientific approach to sociology. Two of the more cited critical reviews were entitled « The Science of Inhuman Relations » by Robert S. Lynd [14][14]  Lynd, 1949. and « "The American Soldier" as Science : can sociology fulfill its ambitions ? » by Nathan Glazer [15][15]  Glazer, 1949.. Nor did later sociologists do anything to dispel this association. In later years, the book figured both as a marker of the wartime contribution to sociology and as an exemplar of the new approach [16][16]  Mullins, for example, cites The American Soldier as....


Historians of sociology agree with reviewers’ assessments that WWII marked a turning point in the adoption of statistics as the authoritative method in sociology and the model for its practice against which other forms came to define themselves. They also agree that the war established a strong association between statistics and survey research. A brief regard back in time underlines the novelty of this association. During the interwar period surveys were just as likely to have been associated with the case study method as with statistics. In the 1930s the terms of the debate and practices began to change. The Depression shifted the focus of social problems from the local to the national level and survey research followed [17][17]  Converse, 1987, 35-38.. National level surveys, which until then had been cultivated outside of academia in electoral polling and market research, made their first inroads into the university. After the war this trend became the mainstream. Debates over scientism were replaced by a general consensus that sociology should strive to be scientific, survey research was institutionalized as an academic practice in university based research centers across the country and statistics took its place as the dominant (albeit still minoritarian) method [18][18]  Platt (1996, 99) specifically notes the curious association....


The symbolic importance of The American Soldier and the centrality of the methodological developments with which it is associated all recommend it as an object for inquiry into the intensified quantification of American sociology in the immediate postwar period. Histories of specific tools, such as attitude scaling, controlled experiments, and modeling, all attest to the contribution (real or symbolic) of The American Soldier to technical developments in sociology. Reviews of primary group and reference group theory similarly attest to the substantive contribution of the volumes (the book is particularly well known for introducing the concept of relative deprivation which spawned a small industry within sociology). My paper focuses on another, more general aspect of the book in particular and of quantification in general, namely the theoretical models with which it came to be associated. The questions that interest me are : « what was the new approach that came to be associated with The American Soldier ? In what ways did it differ from earlier forms of scientism and quantification and what clues or insights does the book provide into the conditions supporting this intellectual turn ? ».


To explore these questions I propose to start with a slightly artificial, though hopefully revealing, historical problem. The label The American Soldier has been used to refer to two overlapping corpuses. In 1949, Princeton University Press published the first two volumes of a series entitled « Studies in Social Psychology in World War II ». These two volumes, rather than the entire corpus, were the object of the publicity campaign and reviews cited above. Later in the year Princeton published a third volume under the title Experiments on Mass Communication. In 1950 a fourth volume in the series appeared under the heading of Measurement and Prediction. Then, in 1950 Robert K. Merton and Paul F. Lazarsfeld, both young but already well-established sociologists at Columbia University, organized a symposium on the implications of the first two volumes for theory development in sociology and social psychology. The symposium led to the publication of an edited collection : Continuities in Social Research : Studies in the Scope and Method of « The American Soldier ». Of the six articles in the volume, two became classic studies : the article by Robert I. Merton and Alice S. Kitt on reference group theory and the paper by Patricia L. Kendall and Paul F. Lazarsfeld on survey analysis. Whereas volumes 1 and 2 were limited to a straightforward analysis of Research Branch data, the subsequent three volumes introduced new techniques, methods and theories. In doing so they reframed the accomplishments of the original research unit.


As these remarks indicate, The American Soldier was the product of four years of research and four more years of elaboration and re-elaboration. The question I wish to begin with is : « How to account for this progression from research to the first two volumes to the methodological texts and finally to Continuities ? ». In raising the question, I do not wish to suggest that each step necessarily built on the preceeding one – all four volumes were part of the initial publication conception – but rather that they reflect different institutional and intellectual goals (and audiences) in the campaign to lend scientific status to survey research in particular and to sociology in general. Continuities, in contrast, was designed to respond to early critics of the initial two volumes. In discussing these texts I will focus on the contrast between The American Soldier, Measurement and Prediction and Continuities as different moments or elements in the quantification of sociology as a discipline (rather than in the development of technical tools) [19][19] The third volume was produced by Carl I. Hovland at....

II - The specificity of American sociology


The history of empirical social research is a relatively new topic for historians. In 1960 Paul F. Lazarsfeld and Robert K. Merton initiated a research program at Columbia University on the history of social research with the explicit aim of improving the status of applied social research [20][20]  Oberschall, 1978.. While their seminar may not have achieved this goal, it did train a number of scholars who helped to put the topic on the map, notably Anthony Oberschall and Terry Clark in the United States and Bernard Lécuyer in France. Subsequent studies have uncovered numerous previously ignored chapters in the history of sociological practice [21][21] Key book length works on American sociological practice,.... They have also been the occasion for theoretical discussions on how to account for the development of an academic discipline [22][22] For an overview of this discussion, cf. the symposium....


Whereas previously the University of Chicago had been identified as the intellectual center of American sociology, more recent studies have highlighted the crucial role of Columbia University in general, and Franklin P. Giddings in particular, in the quantification of sociology during the first decades of the century. Camic and Xie [23][23]  Camic, Xie, 1994. argue that local interdisciplinary conditions, notably the importance of statistics as a criteria of scientificity in the neighboring disciplines of economics and psychology, account for the role of Columbia University in this development. Turner [24][24]  Turner, Turner, 1994, 42-46. identifies Giddings as the founder of a methodological tradition that has defined mainstream sociology until today [25][25] In making this argument Turner distinguishes between.... His argument rests on the importance of teacher-student networks. In the mid-thirties, the quantitative empirical tradition (that Turner associates with Giddings) was challenged, modified and in other ways re-enforced by two alternate programs. The first involved Paul F. Lazarsfeld’s project for applied social research. The second, more philosophical or programmatic approach, was introduced by George Lundberg under the heading of « operationism ». A review of these three strands, their institutional location and the opposition that they aroused sets the scene for an evaluation of the significance of wartime research for sociology as a discipline.

Franklin P. Giddings and a qualified empiricism


Gidding’s contribution to sociology lay in his particular definition of science and his concern to make sociology scientific [26][26] For an overview of Giddings’ philosophy of the social.... His model of the natural sciences was taken from Karl Pearson and more specifically from The Grammar of Science[27][27]  The Grammar of Science, 1895.. Like Pearson, Giddings associated science with mathematical formalization and with empirical tests of theory. He also adopted a strong nominalism, whereby science was about the classification and manipulation of sense perceptions and the achievement of practical ends, rather than about reality. Finally science, and thus sociology, was deductive as well as inductive and, at times, called for speculation. While Giddings himself produced little statistical or empirical research, a number of his students went on to develop his program for a statistical, yet loosely theoretically informed sociology. Two of the most influential were William F. Ogburn and Stuart Chapin. A central feature of this approach was the substitution of correlations for causal explanation.


In the mid-twenties Chapin accepted a position at the University of Michigan and Ogburn moved from Columbia to Chicago. An examination of the range of their organizational activities maps out the institutional basis for the Giddings brand of statistical sociology. Both men served as representatives to the Social Science Research Council, an organization that mediated between private foundations and academic-based researchers [28][28] The Social Science Research Council was created in.... In 1936 Ogburn and Chapin joined a small group of more professionally minded academic sociologists in their secession from the American Sociological Society. The group, that included scientistic sociologists and grand theorists (in opposition to the more reform-oriented, Chicago dominated leadership of the American Sociological Society) founded the exclusive Sociological Research Association. They also established the American Sociological Review as the official journal of the American Sociological Society, thereby dethroning the Chicago leadership and their organ, The American Journal of Sociology[29][29]  Abbott, 1999 ; Lengermann, 1979 ; Kuklick, 1973..


As this brief list of positions and activities indicates, by the 1930s statistical sociology was strategically located in a network of organizations responsible for education, publication and funding. The 1930s and early 40s were a low point for sociology both financially and demographically. In the wake of the Depression, the Rockefeller foundation restricted the amounts and use of their funds. Membership in the American Sociological Society dropped by almost a half and few PhD’s found academic positions [30][30] The exceptions, however, were important for they formed.... At the same time, the Depression and associated recognition of the inefficacy of community-level aid, created a new source of demand for social research, one that favored large-scale survey work and national level statistics.


President Hoover’s 1927 Committee on Social Trends was only the most visible of a series of new opportunities for statistically minded, academic sociologists to work on federally funded projects. The Committee illustrates the influence of the Gidding’s student network. Of the twenty-nine principal contributors to the final report, eleven were PhD’s in sociology, all of them were university professors and most were Columbia trained. All except one also had some connection with the SSRC, then or later, either as committee members or grant recipients [31][31]  Bannister, 1987, 182.. While the project was unprecedented in its scope, it remained quite conventional in method. The final report [32][32]  United States. President’s Research Committee on Social... involved a descriptive account of trends. The failure of the report to meet up to the promise of a new scientific sociology renewed the ongoing debate within the discipline over the epistemological value of the search for statistical correlation and probabilistic laws.

Paul F. Lazarsfeld, « statistical gadgets » and contract research


The professional difficulties of the post-Depression era led sociologists to seek alternate sources of employment and funding, including private contracts. Paul F. Lazarsfeld, who pioneered this new arrangement, had come to the United States from Austria in 1933 on a Rockefeller grant [33][33] There is an extensive literature on Paul F. Lazarsfeld.... Before his arrival, he had been involved in a large-scale research project to study the causes of unemployment in a small industrial town, Marienthal. He had also published a textbook on the use of statistics in psychology [34][34]  Pollack, 1980 ; Capecchi, 1978.. It was this combined expertise in methodology and the empirical study of unemployment that attracted the attention of the Rockefeller Foundation and the Roosevelt Administration. In 1934, he was given an opportunity to open a Research Center at the University of Newark to study the effects of the Depression on youth. The job eventually led to his appointment as director of the Bureau of Applied Social Research at Columbia University.


To support the Bureau, Lazarsfeld actively sought commercial research contracts. His work introduced a new form of empirical research into sociology involving a quasi-permanent research staff and private contracts to fund empirical research. In many instances the same piece of research was conducted first for immediate practical ends, to meet the demands of the client, and later re-analyzed for its « scientific » contribution. Substantively, the Bureau cultivated a type of social psychological research that focused on individual level decision-making and the relation between individual decisions, interpersonal relations and broader social contexts. Technically, it drew extensively from market research and polling and used relatively contained statistical techniques (or « gadgets » as Stouffer described them) to test hypotheses concerning relations between data elements. This empiricist, realist approach contrasted sharply with Giddings’ students’ nominalism and their consequent theoretical concerns.

George Lundberg and the turn to operationism


The third strand contributing to the ascendance of quantification as an authoritative sociological method involved a programmatic campaign to establish sociology as a science. George Lundberg was one of the most vocal proponents of sociological scientism [35][35] Discussions of Lundberg’s epistemological position.... His model of science drew explicitly on physics ; more specifically he associated his approach with the Harvard physicist Percy Walker Bridgman’s theory of operationism. According to Lundberg, Bridgman’s importance lay in the recognition that scientific generalization is always quantitative and that the key to science lay in the incorporation of processes of measurement into the definition of concepts, such that hypotheses could be constructed and empirically tested (and metaphysics avoided) [36][36]  Lundberg, 1936, 44.. Whereas Bridgman insisted on the importance of theory building, Lundberg and the sociologists who followed him gave the philosophy an empiricist twist. Measurement, in this approach, became a precondition for concept definition. In promoting his philosophy, Lundberg cited Stuart Chapin, Samuel Stouffer and Paul F. Lazarsfeld as all conforming to the operationist definition of science. While many quantitative sociologists rejected the label, most respected his call for the methodological unity of science and sympathized with his position [37][37]  Platt, 1996, 103..


The visibility of Lundberg’s philosophy in the late thirties was undoubtedly heightened by the opposition which his and related pronouncements aroused. The debate circled around the relation between theory, method and criticism. Critics of the scientific approach attacked scientific sociologists’ excessive empiricism and the failure of statistical analyses to produce new or meaningful results. The public debate began in 1929 when Ogburn delivered the presidential address to the American Sociological Society. In the speech he called on sociologists to abandon their non-scientific procedures in favor of an objective statistical practice in the service of prediction and social engineering [38][38]  Ogburn, 1929.. The following year Herbert Blumer, who was later to develop symbolic interactionism, dismissed statistically-oriented sociologists as « mere artisans » rather than scientists [39][39] Cited in Bannister, 1987, 176.. The publication of Recent Social Trends that same year provided another occasion for critics of the scientific approach to vent. Pitirim A. Sorokin, then chairman of the Harvard Department of Sociology, attacked the work for its excessive and absurd use of statistics (especially when applied to the arts, religion and intellectual life), its fragmented perspective, the disjuncture between « facts » and analysis and its banal perspective [40][40]  Sorokin, 1933.. Sorokin’s criticisms of Lundberg’s book, Foundations of Sociology in 1939 echoed a similar position. While he agreed on the need for explicit statements he argued that Lundberg failed in his task. The problem, Sorokin explained, lay in a faulty model of science and an absence of empirical work [41][41]  Sorokin, 1940, 796..


One of the interesting features of these debates is that, by the late thirties, few academic sociologists rejected the attempt to render sociology more scientific. Instead their criticisms focused on the empiricism of statistical practice, the ad hoc quality of the hypotheses and the banality of the findings. It was these points that were to reappear in the response to the first volumes of The American Soldier and which informed the intellectual elaboration and symbolic valorization that surrounded its subsequent development.

III - The Corpus


The creation of the Research Branch in 1941 reflected the military’s recognition that social science could aid in the management of its soldiers. In World War I the Army Command had adopted psychological testing to guide them in military recruitment and placement ; in World War II they enrolled social scientists to explore the problem of motivation and morale. The basic premise was that attitudes could be studied scientifically and that morale directly influenced performance. Brigadier General Frederik H. Osborn, the director of the Morale Division, played a central role in gaining support for the Research Branch and in appointing its directors. Osborn had been President of Carnegie Corporation before the war and had many personal connections with leading figures of the SSRC.


The actual research was directed by three scholars, Samuel Stouffer (project director), Carl I. Hovland and Leonard I Cottrell (heads of the Experimental and Survey sections respectively) all of whom had participated in SSRC supported research previously. The majority of the 130 or so researchers employed by the Branch were either students of the three directors or were personally recommended to them (most notably by Paul F. Lazarsfeld). Many were in their twenties at the time and had been recruited directly from graduate school or entry level teaching positions ; others were employed in market research or polling [42][42]  Clausen, 1984, 207-208.. The narrowness of the recruitment reflects both the paucity of skilled survey researchers at the time and the tight social network around statistical and survey research. Techniques of representative sampling and scaling had just begun to be developed and the number of people familiar with these techniques was relatively small. It also reflects the strength of the academic-foundation-SSRC network and the role of The American Soldier in further solidifying that circle.


In a strange confluence of events, the Research Branch conducted its first survey the day after the Japanese attacked Pearl Harbor in December 1941. By the war’s end they had produced over three hundred reports based on the attitudes of more than half a million young men. Topics included, among others : variations in personal adjustment by background characteristics and military experience, job assignment and job satisfaction, attitudes toward leadership and social control, attitudes before combat and behavior in combat, combat motivation among ground troops and problems of rotation and re-conversion. The work was especially notable for its explicit attention to differential attitudes between Blacks and Whites and the role of class and race in determining attitudes. Studies generally began with pretesting of questions and sampling decisions. This phase was followed by the administration of questionnaires in the field. Once collected data was sent to Washington, coded and transferred to IBM punch cards. Final reports relied on machine tabulations, analysis of tables and complimentary qualitative data [43][43]  Stouffer, 1949, 20..


After the War, the Carnegie Foundation provided the SSRC with money to re-analyze and publish the data. The War Department provided the committee duplicate copies of the entire set of IBM cards, over 600,000 interviews and all declassified reports. The SSRC appointed Samuel Stouffer as director of the project. The work began in Washington, but moved to Harvard in the Fall of 1946 when Stouffer was appointed professor in the newly created Department of Social Relations.

« The American Soldier » as applied sociology


Volumes I and II, published in 1949, comprised over 1,000 pages. Volume 1 was subtitled « Adjustment during Army Life » ; volume II covered « Combat and its Aftermath ». In the introduction to the two volumes Stouffer contrasted the actual wartime research, that was shaped by the punctual demands of the Army Command, with The American Soldier which was produced after the war for sociologists and social psychologists. In discussing the relation between « science » and « social engineering » Stouffer oscillated between a model of theoretically driven science in the service of social engineering and the secondary analysis of applied knowledge for theoretical purposes. Whereas the first corresponded to his professed model of science, the latter described The American Soldier. Faced with the decision of what to include in the ostensibly scientific volumes, the authors of The American Soldier adopted « a compromise position with respect to introduction of explicit conceptualisation ». The book, they explained, was written to provide factual data that might be useful to scientists and engineers, « even if the data are in no sense definitive in resolving conflict between alternative hypotheses which might now be advanced or even if the data do not seem relevant to any current hypotheses » [44][44]  Ibid., 32-33.. Hypothesis testing clearly constituted an important index of scientificity.


In many ways volumes 1 and 2 can be seen as a culmination of the strictly empiricist version of scientific sociology rather than the proclaimed exemplar of a new approach. The book synthesized and analyzed an unprecedented amount of individual level data. Its strengths and substantive innovations lay in the attempt to bring together attitude data with structural indicators (such as class, education, position in the army) and to compare attitudinal and behavioral indices (as in the famous study of the relation between attitude and non-combat casualties). A few of the analyses employed simple scales, though nowhere was the same scale employed twice, making comparisons all but impossible. Most of the chapters avoided collective indices and models of any type. Instead the data was analyzed in terms of the percentage of different subgroups responding positively or negatively to single questions. As in the 19th century descriptive tradition, the art of the analyst lay in the selection of subgroups and the comparison of parallel sets of tables. The novelty of the empirical analysis lay in the pretesting of questionnaires and in the extensive use of replication techniques to verify initial conclusions.


For example, the first four empirical chapters of volume one dealt with the relation between verbal and non-verbal measures of personal adjustment to military life. Verbal indices included a set of attitude questions, grouped into four general areas including : personal spirit, personal commitments, satisfaction with status and job and approval or criticism of the Army. The clustering was designed to explore the possibility that the set of questions designed to capture adjustment corresponded with an attitudinal profile rather than a single linear scale. Non-verbal measures were defined from the perspective of the Army Command. Rank and more specifically promotion in rank was taken as a positive, non-verbal measure of adjustment, while psychoneurosis and AWOL status served as evidence of negative adjustment. The study reported findings on the relation between verbal and non-verbal measures of adjustment, according to level of education, type of military corps and assignment overseas or at home. It then went on to explore the effect of background characteristics on both types of adjustment. These included level of education, marital status and age. Finally, the researchers explored the hypothesis that pre-army adjustment and relative deprivation mediated the effect of background characteristics on adjustment. They also studied the predictive effect of verbal attitudes on subsequent non-verbal adjustment (in other words, they wanted to know how much good morale served to predict subsequent promotions).


The data for this analysis combined materials from questionnaires administered to a wide variety of different groups at different points in time in the course of the war. As such, the findings concerned purely statistical populations, rather than actual communities. The size of the total database was particularly important given the reliance of the researchers on matched comparisons to explore the relation between different variables. This method limited the investigation to the relation between two variables at a time (e.g. the relation of age on personal spirit or rank). The only way to control for the effect of other variables was to construct sub-populations which had similar values on all other variables. Once the subgroups had been constructed, values were weighted by the proportion of the subgroup in the population as a whole. For example, a study of the effect of marital condition compared the responses of married men to those of unmarried men of a similar age and level of education. Findings were reported in terms of the percentage of each subgroup responding positively or negatively to each individual question and displayed in the form of bar graphs corresponding to each variable. In contrast to more formalized methods, analysis in this approach depended on extensive examination of raw data, classified in a variety of different forms.


As this brief description indicates, the method of matched comparisons reflected and re-enforced an atomistic model of statistical populations and a variable form of explanation. Mathematization was limited to constructing scales to measure the consistency of the different attitude measures, though, as Stouffer pointed out, few were actually developed. For the most part, the data was presented as a purely descriptive account of military life. In those rare cases of counter-intuitive findings, the authors appealed to an ad hoc form of psychologizing and more specifically to the theory of relative deprivation. According to Stouffer, the theory was first introduced as an attempt to account for the absence of differences between the attitudes towards personal adjustment of Northern (American) Negro soldiers stationed in the North and the South. The finding was surprising given Negro soldiers consistently stated preference for Northern postings and resentment at discrimination in Southern buses and by Southern police. To reconcile these findings, the authors introduced a theory of relative deprivation, suggesting that the Northern Negro soldier had a significant advantage over Northern Negro civilians [45][45] Cited in Jahoda, 1962, 496.. Once adopted, the authors used the theory to explain other surprising findings. For example, they found that location (at home versus abroad) had little effect on the personal spirits and job satisfaction of married men. To explain this finding, the authors suggested that the opportunity to be with one’s wife made adjustment to army life more difficult, thus accounting for the absence of effect.


From the perspective of sociology, The American Soldier placed the problems of the extent and limits of secondary analysis and the relation between theory and empirical analysis directly on the table. The questions were : « what use could scientists or sociologists make of data that had been collected for extra-scientific problems ? » and « to what extent could explanatory science develop on the basis of descriptive studies produced for commercial or policy purposes ? » The problem, it should be noted, was the opposite of the physicists who argued for the application of basic research (once completed) to practical problems of political import.

The response from the discipline


Reviews of The American Soldier were of two types : humanist critiques that questioned the authors’ model of scientific sociology and those who accepted the model, but challenged the exemplary status of book. While the first appeared in intellectual magazines directed at the general (educated) public such as Partisan Review and Commentary, the latter figured in mainstream sociological journals, most notably in The American Journal of Sociology, The American Sociological Review, Public Opinion Quarterly and Social Forces. The absence of more fundamental criticisms of the project as a whole within the professional literature attests to the emergence of a disciplinary consensus concerning the scientific goals of sociology and the role of quantification and survey research in securing that end.


Two public reviews stand out for their endurance in later collective memory. The first, by Robert S. Lynd [46][46]  Lynd, 1949., addressed the moral implications of applied social re-search. More specifically, Lynd underlined the extent to which the substantive findings of The American Soldier had been shaped by the interests of the Army Command. The book, Lynd argued, was as an example of social science in the service of social control and a threat to both science and democracy. Lynd was an outspoken advocate of a humanistic model of sociology as a force for social criticism and change. His criticisms directly challenged the editors’ own political claims that the objectivity of scientific sociology provided an important safeguard for democratic regimes.


The second critical review to address professional sociologists came from Nathan Glazer [47][47]  Glazer, 1949. who was an independent researcher at the time. His article, « "The American Soldier" as Science : can Sociology Fulfill its Ambitions » criticized the book for failing to live up to its scientific promise. Glazer presented The American Soldier as a particularly strong example of the attempt to transform sociology in the image of the physical sciences. In this model, Glazer explains, « science is a technical enterprise which involves, among other things, the observation of phenomena, often indirectly through technical devices, the discovery of uniformities – of patterns – in these phenomena, the formulation of hypotheses about these uniformities and the determination, by experiment, prediction, and further observations, which of two or more hypotheses best describes the phenomena and is worthy of being called a law. Science is also an enterprise which… is cumulative » [48][48]  Ibid., 489..


Given this definition, Glazer asks, how well does the book deliver ? To assess the work, and thereby scientific sociology, Glazer examined The American Soldier for examples of observation, hypothesis testing and prediction. In each case, he argued, the analysis failed to realize its promise. In the case of observation, the technology of IBM punch cards and machine calculations led the authors to produce mounds of data on uninteresting and irrelevant topics. « The question is », Glazer argues, « do questionnaires give us good indexes ? Do they really reflect the social reality they are used to study ? » [49][49]  Ibid., 491.. According to Glazer they did not. Questionnaires, he concluded, do not capture underlying attitudes and attitudes don’t predict behavior. Turning to hypotheses, Glazer found few actual examples. The most striking example concerned the use of « relative deprivation » to explain personal adjustment. Whereas true scientific hypotheses could be tested and eventually refuted, « relative deprivation » could be used to account for all possible relations – and their obverse. Finally, the absence of adequate conceptualization and a theory of the connections between items rendered prediction impossible. For Glazer, the failure of The American Soldier to produce scientific knowledge called into question the entire endeavor. Scientific method, applied to human affairs, does not produce science.


Turning to the more mainstream evaluations, all began by praising the book as an important event and significant scholarly contribution. Features mentioned included : the scientific study of attitudes, concept of intervening variables, and use of replication. At the same time, many of the reviewers expressed their concern over the paucity of conceptual and theoretical work. Two reviews in particular stand out for their length, and thus more in-depth analysis, and for their location in mainstream sociological journals. These include the reviews by Neil Demerath and Ethel Shanas, in Social Forces and The American Journal of Sociology respectively. Demerath questioned the wisdom of Stouffer’s professed « compromise position ». Distinguishing between theorizing and conceptualization, he insisted on the importance of the latter for the scientific value of the research.


Conceptualization is inevitable, explicit or implicit. It always rests upon a theory of knowledge and a methodology in the broad sense of that term ; but conceptualization does not necessarily entail a theory of relationship or sequence of events. Just as medical scientists employed the concepts of organism, nervous system, health and disease independently of the various theories of disease so do social scientists now employ certain descriptive concepts or categories out of which is developing something like a « common language » [50][50]  Ibid., 89..


By failing to explicate and systematize the use of concepts throughout the volumes, he argued, the authors impoverished their study, limiting its value for subsequent theorizing.


In a more far-reaching criticism Ethel Shanas [51][51]  Shanas, 1950. reproached the authors for the disjuncture between data, hypotheses and conclusions. Shanas distinguished between two types of surveys (found in the book) :

  1. those in which the interpretative hypotheses seem to have directed the scouring of the data used in the analysis,


  2. those in which the hypotheses appear to have been developed independently of or after the analyses of the data.

Most of the analyses were of the second kind such that conclusions were drawn independently of the data. In those few occasions where the data was collected to test a hypothesis, as in the study of the relation between professed willingness for combat and non-combat casualties, emphasis on computational techniques led to « spurious verifications ». In a re-analysis of the same data, Shanas showed that it could be used to support opposite conclusions. « One can only conclude that in the chapter all possible sorts of computations were done in an attempt to verify the hypotheses advanced. Some of these computations ultimately gave the result desired » [52][52]  Ibid., 592.. In closing, Shanas hoped that young scholars would not adopt the book as a model of practice. « Scientific research, despite permissible meandering, must concern itself with the interplay of theoretical premise and empirical observation. This has marked the advance of modern science. To trust that scientific achievement will come through a Baconian method of assembling vast quantities of precise data is to place reliance on a hope long abandoned by other sciences » [53][53]  Ibid., 594..

Operationism and the construction of indices


Taken together this set of reviews captures the different features that led contemporaries to uphold The American Soldier as a new form of social science. The book (claimed to) use applied research and more specifically policy work in the service of basic science and it (attempted to) apply natural science methods to human affairs or sociology. Negative reviews dismissed the entire scientific enterprise as amoral or misconceived, while positive reviews praised the book, but questioned its role as an exemplar of scientific sociology. The problems lay not in the sophistication of techniques or research design, but rather in the fit of theory and empirical research embodied in the first two volumes and in the value of verification as a criteria of sociological knowledge.


One of the more striking features of the discussions surrounding the volumes involves the lack of consensus (amongst « scientific » sociologists) concerning the type of theory that should properly inform quantitative or scientific sociology. The juxtaposition of Demerath and Shanas’ remarks points to two distinct positions. Whereas Demerath called for a type of operationism, in which conceptualization took the place of theoretical work, Shanas’ argument rested on a more empirico-deductive conception of science involving the empirical testing of theoretically informed hypotheses. It was this division, rather than the more general debate over science versus humanities, that distinguished the Gidding’s tradition from the Lazarsfeld and Lundberg models and that would seem to have been up for grabs at the time [54][54] Stouffer himself straddled the two worlds. He was a....


The publication of Measurement and Prediction[55][55]  Stouffer, 1950. the following year did little to address these criticisms. Instead the book offered a variety of technical solutions to the absence of adequate conceptualization and theorizing. The volume presented two types of attitude scale : Guttman’s scalograms and Lazarsfeld’s latent structure analysis. It also contained a number of chapters on the problems of prediction. The substantive problem addressed in Measurement and Prediction concerned the relation between opinions, as evidenced by answers on questionnaires, and an underlying abstract concept, « attitude ». The problem, Stouffer explained, was in large part one of definition and measurement. While theorists such as Pareto, Thomas and Mead had written extensively on « attitudes », their analyses failed to discuss the passage from theory to measurement. The challenge, he suggested, was similar to that faced by students of intelligence in the 1920s and 30s. In both cases, a single concept referred to a wide variety of behaviors or traits. In the case of intelligence, the development of concrete measurements in the form of I.Q. testing and factor analysis had shifted the focus of debate from a verbal theoretical level to empirical research around alternate models. In a similar way, surveys and attitude scales promised to transform the discussion around morale and attitudes into an empirically grounded scientific enterprise [56][56]  Stouffer, 1949, 35.. The contribution of Measurement and Prediction lay in the provision of two such models around which empirically grounded debate could proceed. Both Guttman Scaling and Latent Structure Analysis were concerned with testing whether a cluster of different questions could be said to index a single attitude.


Volume four can thus be read as an operational response to the problem of theorization and conceptualization. Whereas earlier sociologists sought to define attitudes substantively, in terms of either civilizational and social structures or social processes, The American Soldier offered a quasi-empirical, technical approach to the problem. As in the first volumes of The American Soldier, the craft of sociology lay in the construction of categories or variables. Only here the problem was to find a scientific grouping of questions, not individuals. In the original research data had been produced to respond to punctual problems posed by the Army Command. In The American Soldier vols. one and two, the same data had been re-organized thematically from the perspective of the individual soldier, their integration in the army and military experience (as seen through the lens of their attitudes and expectations). In volume four it was re-organized yet again, according to the formal features of the data. Hypotheses, in this volume, revolved around the internal coherence of the data. The aim of the exercise was to reduce index-ambiguity and index-instability.

Addressing the criticisms : theory and quantification


Whereas Measurement and Prediction was not written in response to scholarly criticisms of The American Soldier, Continuities clearly was. The book was the product of a symposium held to examine the contribution of The American Soldier to sociology and social psychology. Far from a unique event, it was one of a number of such symposia organized by Robert K. Merton and Paul F. Lazarsfeld to « stimulate continuity in social research » [57][57] Other books subjected to a similar treatment included.... The aim of the book was thus « to examine selected hypotheses, methods and findings with an eye to their specific implications for further advancement of certain part of social sciences. Through such re-analysis, it is hoped that these materials will be the more likely to enter into the mainstream of future research and theory, rather than being let to find their own way » [58][58]  Merton, Lazarsfeld, 1950, 11.. The authors justified the need for such a book by the novelty of the approach contained in the first two volumes of the book and the size of the volumes.


Of the six papers included in Continuities, two quickly became a point of reference for future researchers. These include an article by entitled « Contributions to the Theory of Reference Group Behavior » by Robert K. Merton and Alice S. Kitt and on « Problems of Survey Analysis » by Patricia L. Kendall and Paul F. Lazarsfeld. The first demonstrated the theoretical contribution of the Research Branch’s work, while the latter highlighted its methodological significance. A brief examination of each of these papers illustrates the way in which Continuities redefined and valorized the original text.


Kendall and Lazarsfeld’s article used the Research Branch data to discuss the art of survey analysis. The discussion is significant in that it shifted the art of social research from problems of data collection and category construction to that of analysis (as did volume 4). As such it paved the way for later developments where official computer based data sets allowed the statistical sociologists to distance themselves from the messy work of data construction, shifting the problem of uncertainty from data to models [59][59]  Clogg, Dajani, 1991, 8.. The article presented the analysis of survey data as a problem of how to clarify the statistical relation between two variables. The task, they argued, was twofold : « In the first place, we want to determine how legitimate it is to draw inferences of cause and effect. Secondly, we want to examine the process through which the assumed cause is related to its effect » [60][60]  Ibid., 136.. Topics treated in the discussion included :

  1. the approximation of survey results to controlled experimentation, including problems of spurious factors and time order variables,

  2. interpretation and its place in a general scheme of (statistical) elaboration,


  3. the treatment of complex phenomena, including problems of ambiguous answers, index construction and the relation of personal and unit data.

The importance of the chapter can be seen in its subsequent adoption as a model for survey textbooks in the 1950s, most notably The Language of Social Research[61][61]  Lazarsfeld, Rosenberg, 1955. by Lazarsfeld himself and Morris Rosenberg and Survey Design and Analysis[62][62]  Hyman, 1955. by Herbert Hyman.


Whereas Kendall and Lazarsfeld used The American Soldier data to outline a variable approach to data analysis, Merton and Kitt took on the critics by discussing the theoretical contributions of the two volumes. The authors framed their discussion with an assertion of the « two-way » relation between social theory and empirical research [63][63]  Merton, Kitt, 1950, 40.. Their article focused on the contribution of findings in The American Soldier to the theory of reference group behavior. It also provided an opportunity to examine the integration of social structural indices with individual level attitudinal data. The aim of the article was to shift the level of analysis to a higher level of generality or abstraction than that found in the original volumes and to relate it to a broader literature. A striking feature of the discussion was the extent to which theoretical problems were discussed in the language of variables and variable analysis. Relative deprivation was presented as an « interpretative intervening variable » that served to explain statistical relations between attitudinal and social structural variables. From the perspective of sociology as a discipline, the significance of the essay lay in the link that it established between the statistical analysis of survey data and emerging functionalist theory, thereby anticipating and contributing to the dominant sociological tradition of the fifties and sixties [64][64] For a contemporary discussion of this convergence cf.....


From the perspective of The American Soldier, the impact of Continuities was inestimable. The edited collection basically established The American Soldier as an exemplar of survey research and a classic in the field. Most, if not all, subsequent references to the book highlighted those features discussed in Continuities, and many cited the edited collection rather than the original. Whereas reviewers in 1949 and 1950 sharply criticized the original volumes for their absence of theory and conceptualization, Merton and Kitt held them up as « a useful occasion for examining the interplay of social theory and applied social research » [65][65]  Merton, Lazarsfeld, 1950, 40.. While the effect of this work of valorization is obvious, the problem of why it was undertaken is less clear. What was it about the position of sociology and the social sciences in general and survey research in particular that led the Carnegie Foundation and the SSRC to support a four-year project of re-analysis and pushed Merton and Lazarsfeld to engage in a subsequent re-analysis of the re-analysis ? Why work so hard to valorize an admittedly enormous, but theoretically weak study ? And why promote it as an exemplar of empirically informed theoretical research when even its strongest supporters agreed that it failed to meet up to those standards ? The question, of course, extends beyond the significance of The American Soldier per se to the conditions that elevated this model of sociological analysis to a paradigmatic status against which all other approaches were subsequently defined.

IV - Sociology in the late forties


An examination of the professional position of some of the most visible actors (and most well known sociologists) involved in this saga provides a preliminary indication of the variety of overlapping conditions accounting for the effort invested in the construction of The American Soldier as a classic. In 1946 Stouffer moved from his prewar position as professor of sociology at the University of Chicago to Harvard where he was invited to found the Research Laboratory of the newly created Department of Social Relations. As indicated above, he brought with him The American Soldier project and a number of young researchers, including Leland DeVinney and Brewster Smith. Carl I. Hovland and Leonard I Cottrell, the directors of the two analytic sections, returned to their positions at Yale and Cornell respectively. Volume III was produced at Yale, while work on Volume IV proceeded at Cornell where Louis Guttman and John Clausen were located and at Columbia where Paul F. Lazarsfeld held the position of Professor of Sociology and Director of the Bureau of Applied Social Relations. Continuities was edited by Lazarsfeld and Robert K. Merton who was also at Columbia and was also affiliated with the Bureau.

A shift in the (institutional) center


The location of the editors of the two main volumes of The American Soldier at Harvard and of that of Continuities at Columbia immediately situates the book in the shift in the locus of elite sociology from the University of Chicago to Harvard and Columbia universities. An examination of the positions of the broader group of researchers also points to their central role in developing interdisciplinary research institutes organized around the conduct of survey research in a number of universities [66][66]  Bulmer, 1992 ; Winsborough, 1992.. Of the two types of development, the first was more important in the institutionalization of a new mainstream tradition, as such it provides the focus of this discussion. The latter, in contrast, influenced sociology indirectly through its support for the consolidation of social psychology as a discipline situated between psychology and sociology and through the provision of resources of sociologists in social psychology and area studies.


Recent theoretical debates in the sociology of sociology highlight the importance of three distinct factors in the ongoing development of the discipline. These include local, interdisciplinary relations [67][67]  Camic, 1995., networks of teacher-pupil relations [68][68]  Turner, 1994 ; Collins, 1998. and material resources [69][69]  Turner, Turner, 1990 ; Turner, 1990.. In the case of postwar sociology, variants of all three factors can be seen to have been at play. Instead of interdisciplinary relations, the relations between department and administration seems to have been critical for the ascendance of scholars such as Stouffer, Lazarsfeld and Merton [70][70]  Lowen, 1997.. Instead of, or rather in addition to, pupil – teacher chains, research networks such as that formed around the Research Branch and The American Soldier became important channels of transmission and sources of jobs and funding. Finally, in terms of resources, the immediate postwar period was marked by a diversification in the sources of external funding. Whereas prior to WWII most extra-university funding came from the Rockefeller Foundation, in the postwar period the Carnegie and Ford foundations provided extensive support for social research, as did the US military and later the federal government. In the immediate postwar period most of this money went to elite universities, most notably Harvard, Columbia and Chicago [71][71]  Turner, Turner, 1990, 96..


An important factor in the entry of applied social research into elite Northeastern universities involved the commitment of university presidents to transform their institutions from centers for elite training to centers of scientific research. This transformation was particularly marked at Harvard where President James A. Conant under-took a major reform of Harvard’s entire curriculum. His book, On Understanding Science, summed up Conant’s scientific philosophy and had a major impact on university education across the country. In addition to strengthening the natural sciences and introducing the history of science into the curriculum, Conant supported the more scientific branches of the social sciences. From 1941 on, a group of Harvard professors, including Talcott Parsons, Gordon Allport and Clyde Kluckhorn, had lobbied President Conant to create a new interdisciplinary department, combining sociology, social psychology and cultural anthropology [72][72] The initial proposal grew out of their joint participation.... The wartime contributions of the social sciences helped to persuade Conant that inter-disciplinarity was indeed the wave of the future. In 1946, the Department of Sociology was replaced by the Department of Social Relations with Talcott Parsons as its chair and an associated Laboratory of Social Research headed by Stouffer [73][73]  Nichols, 1998 ; Johnston, 1998.. Contrary to the received wisdom, Parsons and the Department more generally actively promoted a model of theoretically informed empirical (scientific) sociology [74][74]  Lidz, 1986 ; Vidich, 2000.. Students were made vividly aware that the disciplinary status of the sociology depended on its scientific credibility [75][75]  Johnson, Johnson, 1986. and « basic research » was an organizing concept in the graduate curriculum.


While the organizational changes in the Columbia Department of Sociology were not so radical, they also involved a significant shift in leadership and intellectual orientation. Ongoing conflicts between the older professors, notably MacIver and Lynd, and the new orientation of the university (fueled by obvious financial opportunities in the form of private and military funding) opened the way for a take-over by the younger, empirically oriented members of the Department [76][76]  Halas, 2001.. In the years after the war Lazarsfeld and Merton effectively assumed control of the department. Merton was rapidly promoted to Associate Professor in 1944 and to full Professor in 1947. The change was signaled by the integration of the Bureau of Applied Social Research into the Sociology Department. While it remained a separate organization, the academic curricula was reshaped to include participation in the Bureau’s projects [77][77]  Clark, 1996.. The Continuities volume on The American Soldier was only one of many projects designed to elevate, transmit and standardize the Bureau’s largely empiricist approach to social research. The coupling of Parsons and Stouffer at Harvard was thus echoed in the Merton-Lazarsfeld pairing at Columbia. At both institutions, sociology was defined by a relation between structural-functionalism on the one hand and a problem-oriented variable approach to empirical research on the other, with a heavy reliance on survey data.

Social Science and the NSF


In the mid-forties, pressures within the university to appear more scientific were re-enforced by a disciplinary drama concerning the inclusion of the social sciences in the proposed National Science Foundation. The campaign for federal funding of science began in 1944 when President Franklin Roosevelt asked a group of elite scientists, led by Vannevar Bush, the then head of the Office of Scientific Research and Development, to outline a proposal. The report, entitled Science – The Endless Frontier, was delivered to President Harry Truman in 1945 and received national attention. It called on the government to create a federal agency, run by scientists, to support basic research. Drawing on wartime achievements, the report argued that fundamental research (not directed at any immediate, practical end) would strengthen national security and further economic growth [78][78] The Office of Scientific Research and Development had.... The report launched a legislative debate that lasted seven years, culminating in the creation of the NSF in 1950. The years of The American Soldier were thus also the years of the debate over national funding for basic research.


For Bush and his colleagues the key issue concerned who would control the proposed agency, scientists or administrators. Bush used the wartime achievement of the physicists to argue for professional self-regulation. For social scientists, in contrast, the question was whether they would be included in this grand scheme [79][79]  Klausner, 1986 ; Solovey, 2001.. The possibility of including the social sciences was first raised by President Truman. In the Fall of 1945 Congress invited the SSRC to argue their case. A number of representatives from different disciplines were chosen to defend the social sciences ; sociology was represented by William F. Ogburn. In a particularly unfortunate strategic move, the sociologists argued in favor of government, rather than scientific, control of the proposed agency. This position put them at odds with the elite group of scientists led by Vannaver Bush who, in turn, opposed their inclusion in the new organization. To support their case, the latter pointed to the absence of social laws comparable to physical laws, the absence of experimentation and scientific method more generally, and the political ideological character of social knowledge. In the end the social sciences were excluded and scientists gained full discretionary control over the distribution of federal funding. The 1950 legislation stated that NSF would support « basic scientific research in the mathematical, physical, medical, biological, engineering and other sciences ». While the social sciences were not definitively excluded, they were not explicitly included.


Historians agree that the exclusion of the social sciences had less to do with the scientificity of the social sciences and more to do with administrative, political and institutional considerations. That said, Vannevar Bush’s report clearly influenced social scientists’ model of science. It also re-enforced their sense of disciplinary weakness and concern for problems of legitimation [80][80]  Larsen, 1992, 7.. The lesson participants in The American Soldier would seem to have taken home was that scientific legitimacy depended on their ability to persuade the world that sociology involved apolitical, basic science. It was thus neither the critical (and thus politically-suspect) project of Lynd nor the simpleminded empiricism of interwar quantifiers. Instead, following the model of the natural sciences, it involved a theoretically informed application of scientific method to fundamental problems, with significant practical applications. The passage from the Research Branch data to the first volumes of The American Soldier to the methodological volumes and finally to Continuities can be seen as a series of successive attempts to reconstruct the wartime experience as precisely the type of work that was worthy of scientific and political recognition. It was only with the fifth book, Continuities, that scientific sociologists would seem to have gotten it « right » in the sense of conforming to the model of science embodied in Conant and Bush’s influential statements. However, as the above discussion suggests, the actual practice was closer to Lundberg’s operationism than Giddings and Ogburn’s model of theoretically informed empirical research.

IV - The new scientific sociology


What then can we say about the consequences of The American Soldier and of wartime research for the subsequent development of sociology ? The most tangible consequences were clearly related to the location and status of survey research. Wartime research definitively linked survey research with quantification. It also catapulted it from its former position in market research and electoral polling into university-based research centers. The Bureau of Applied Social Research at Columbia, the National Opinion Research Center (which moved from Denver to Chicago in 1947) and the Survey Research Center at Michigan were among the more visible of a number of interdisciplinary research centers that contributed to the quantification of academic sociology. These organizations, along with the Department of Social Relations, also contributed to the consolidation of an interdisciplinary project around sociology, social psychology and cultural anthropology [81][81]  Brick (2000) argues that it also prepared the way.... The success of these centers was clearly related to their ability to attract foundation funding, and this, in turn, depended on the strength of the wartime research networks and the movement of individuals into positions of influence in both the centers and the funding organizations.


The second major achievement of the wartime experience, at least from the perspective of sociology, was the ascent of statistics in general and survey research in particular as the dominant methodology. The mid-fifties witnessed the introduction of a new genre, that of the methodology textbook, most of which followed the model of survey and variable analysis indicated in Continuities[82][82]  Platt, 1996, 33-37.. This shift was accompanied by the disappearance of debates over the scientific status of sociology. Whereas scientism had been a divisive issue in the thirties, after the war few scholars engaged in polemics concerning the scientific status of sociology. Instead they took for granted a loose definition of science as method and of statistics as the authoritative method. As the case of The American Soldier suggests, this closure was more a result of institutional and political developments than genuine intellectual achievements. Scientism and thus statistics did not win out because they had been successfully integrated into an epistemological project combining theory and empirical research. Instead, they triumphed because their carriers used their wartime experience to monopolize disciplinary resources and because Cold War politics successfully silenced proponents of a more critical stance.


This observation leads to a third less obvious, but possibly more lasting outcome of wartime research, namely the consequences of the elevation of scientific methodology and statistics for the nature of sociological theory. In some ways, the different pre-war programs for scientific sociology can be distinguished by the place they accord to theory. Giddings called for the use of quantification and statistics to explore general theories. His work and that of his students was informed by evolutionary theories of civilizational development. Lazarsfeld’s approach, in contrast, was ultimately practical and atheoretical. His work was alternately problem and tool-driven. Finally, Lundberg’s meta-discussions affirmed a theoretical model of quantitative sociology, but ultimately focused on the problem of operationalization. This position is clearly exemplified by Demerath’s call for concept formulation, rather than theorizing, in his review of The American Soldier.


As indicated above, the successive volumes of The American Soldier contain elements of each of these approaches. Volume one opened with a statement of the Giddings’ vision of statistical science. The actual analyses in volumes one and two, however, were largely empiricist, with a very minimal use of tools such as attitude scales and ad hoc hypotheses. Volume four, in contrast, introduced formal mathematical modeling of data. Hypotheses concerned, not the relation between individuals or variables, but rather the fit between questions and attitude scales and nowhere were these tools used to draw substantive conclusions. It was only in Continuities that Lazarsfeld, Merton and their colleagues re-framed the work of the Research Branch in a form that began to resemble dominant models of science. But even there, the discussion of reference group theory and survey research were largely distinct. The main relation being their common use of the language of variable analysis.


The institutionalization of the « new approach » in the postwar network of elite universities and new foundation money did little to clarify the relation between theory and quantitative research. At the level of public image, sociologists at Harvard and Columbia actively promoted the social sciences as true sciences in the model of the physical sciences. Students in the Department of Social Relations were exposed to extensive meta-discussions concerning the scientific method ; Conant’s book On Understanding Science was required reading and the core curricula was organized around the different stages of scientific research. Topics included : assumptions of basic social science, fundamental concepts defined, fundamental propositions and hypotheses, analysis and evaluation of current theories, strategic problems, social science and the realm of value and applications of social science [83][83]  Vidich, 2000.. At the same time, there were few examples of actual integration between the different topics. Parsons’ systems theory was not amenable to quantitative formulation and testing. Stouffer’s hands-on approach to data analysis was informed by « ideas », not theories [84][84]  Hyman, 1962, 328.. Lazarsfeld persisted in his program of applied social research. The only one of the leading quartet to directly take on the problem was Merton and he too has been accused of adopting a hodgepodge approach to concept formation, importing definitions according to the logic of the data rather than an overarching theoretical scheme. In retrospect his middle-range theories, while extremely fruitful as tools to organize empirical research, did not add up to theoretical system or even an explanatory framework [85][85]  Turner S., 1993, 24 ; Turner J., 1974, 71-73..


While a full discussion of the place of theory in postwar elite American sociology is beyond the scope of this paper, the wartime experience would seem to have moved sociology away from theoretically informed analysis to a more sophisticated type of empiricism buttressed by operationally defined concepts and formal mathematical models designed to test the strength and internal coherence of already coded data. Theory in this story lost out twice. In the Parsonian model it was divorced from empirical research (despite Parsons’ intentions to the contrary) and in the empirical model it was replaced by operational definitions and a variable approach. In conclusion, the more interesting finding of my analysis may have less to do with quantification, which, while valorized was not significantly transformed, and more to do with theory. At the very least, it raises the question of what counts as theory and the influence of quantification thereupon [86][86] For discussions of the problematic relation between....


  • Abbott A., 1988, Transcending general linear reality, Sociological Theory, 6 (Fall), 169-186.
  • Abbott A., 1999, Department & discipline : Chicago Sociology at One Hundred, Chicago, University of Chicago Press.
  • Bannister R.C., 1987, Sociology and Scientism : The American Quest for Objectivity, 1880-1940, Chapel Hill, University of North Carolina Press.
  • Barton A.H., 1979, Paul Lazarsfeld and applied social research : invention of the University Applied Social Research Institute, Social Science History, 3, 3-4.
  • Brick H., 2000, Talcott Parsons’ « Shift away from economics, 1937-1946 », The Journal of American History, 87, 2, 490-514.
  • Bryant Ch.G.A., 1985, Positivism in Social Theory and Research, Houndmills, Basingstoke, Hampshire, Macmillan.
  • Bulmer M., 1992, The growth of applied sociology after 1945 : The prewar establishment of the postwar infrastructure, in Halliday T.C., Janowitz M., Sociology and Its Publics : The Forms and Fates of Disciplinary Organization, Chicago, University of Chicago Press, 317-346.
  • Camic Ch., 1995, Three departments in search of a discipline : localism and interdisciplinary interaction in american sociology, 1890-1940, Social Research, 62, 4, 1003-1033.
  • Camic Ch., Xie Y., 1994, The statistical turn in american social science : Columbia University, 1890 to 1915, American Sociological Review, 59, 5, 773-805.
  • Capecchi V., 1978, Paul F. Lazarsfeld : A link between american and european methodology, Quality & Quantity, 12, 3, 239-254.
  • Clark T.N., 1996, Clientelism and universalism : Columbia sociology under Lazarsfeld and Merton, The Tocqueville Review/La Revue Tocqueville, 17, 2, 183-205.
  • Clausen J.A, 1984, Research on « the American Soldier » as a career contingency, Social Psychological Quarterly, 47, 2, 207-213.
  • Clogg C.C., Dajani A.N., 1991, Sources of uncertainty in modeling social statistics, Journal of Official Statistics, 7, 1, 7-24.
  • Collins R., 1998, The transformation of philosophy, in Heilbron J., Magnusson L., Wittrock B., The Rise of the Social Sciences and the Formation of Modernity : Conceptual Change in Context, 1750-1850, Boston, Massachusetts, Kluwer Academic Publishers, 141-161.
  • Converse J.M., 1984, Attitude measurement in psychology and sociology : the early years, in Turner Ch.F., Martin E., Surveying Subjective Phenomena, New York, Russell Sage Foundation, 3-39.
  • Converse J.M., 1987, Survey Research in the United States : Roots and Emergence 1890-1960, Berkeley, California, University of California Press.
  • Danziger K., 1992, The project of an experimental social psychology : historical perspectives, Science in Context, 5, 2, 309-328.
  • Daston L., 1993, Reply to James Chandler’s Commentary on « Marvelous Facts and Miraculous Evidence », Chicago Evidence Conference, 21-24 May 1992.
  • Demerath N.J., 1949, « The American Soldier », Social Forces, 87-90.
  • Glazer N., 1949, « The American Soldier » as science : can sociology fulfill its ambitions ?, Commentary, 8, 5, 487-496.
  • Hacking I., 1992, « Style » for historians and philosophers, Studies in History and Philosophy of Science, 23, 1, 1-20.
  • Halas E., 2001, How Robert M. MacIver was forgotten : Columbia and american sociology in a new light, 1929-1950, Journal of the History of the Behavioral Sciences, 37, 1, 27-43.
  • Halliday T.C., Janowitz M., (eds.), 1992, Sociology and Its Publics : The Forms and Fates of Disciplinary Organization, Chicago, University of Chicago Press.
  • Herman E., 1995, The Romance of American Psychology, Berkeley, California, University of California Press.
  • Hinkle R.C., 1994, Developments in American Sociological Theory, 1915-1950, New York, State University of New York Press.
  • Hollinger D.A., 1995, Science as a weapon in Kulturkämpfe in the United States during and after World War II, Isis, 86, 440-454.
  • Hyman H.H., 1955, Survey Design and Analysis : Principles, Cases and Procedures, Glencoe, Illinois Free Press.
  • Hyman H.H., 1962, Samuel A. Stouffer and social research, Public Opinion Quarterly, 26, 3, 323-328.
  • Jahoda M., 1962, Research Methods in Social Relations ; with Especial Reference to Prejudice, New York, Holt, Rinehart and Winston.
  • Jardini D.R., 2000, Out of the blue yonder : the transfer of systems thinking from the pentagon to the great society, in Hughes A.C., Hughes T.P., Systems, Experts and Computers, Cambridge, Massachusetts, The MIT Press, 311-357.
  • Johnson B., Johnson M.M., 1986, The integrating of the social sciences : theoretical and empirical research and training in the department of social relations at Harvard, in Klausner S.Z., Lidz VM., The Nationalization of the Social Sciences, Philadelphia, University of Pennsylvania Press, 131-140.
  • Johnston B.V., 1998, The contemporary crisis and the social relations department at Harvard : a case study in hegemony and disintegration, The American Sociologist, 29, 3, 26-42.
  • Katz D., 1978, Social psychology in relation to the social sciences : the second social psychology, American Behavioral Scientist, 21, 5, 779-792.
  • Klausner S.Z., 1986, The bid to nationalize american social science, in Klausner S.Z., Lidz V.M., The Nationalization of the Social Sciences, Philadelphia, University of Pennsylvania Press, 3-40.
  • Klausner S.Z., Lidz V.M., (eds.), 1986, The Nationalization of the Social Sciences, Philadelphia, University of Pennsylvania Press.
  • Kuklick H., 1973, A « scientific revolution » : sociological theory in the United States, 1930-1945, Sociological Inquiry, 43, 1, 3-22.
  • Lanier L.H., 1949, The Psychological and social sciences in the national military establishment, The American Psychologist, 4, 5, 127-47.
  • Larsen O.N., 1992, Milestones and Millstones : Social Science at the National Science Foundation, 1945-1991, New Brunswick, New Jersey, Transaction Publishers.
  • Lautman J., Lécuyer B.P., (eds.), 1998, Paul Lazarsfeld (1901-1976) : la sociologie de Vienne à New York, Paris, l’Harmattan.
  • Lazarsfeld P.F., Rosenberg M., (eds.), 1955, The Language of Social Research : a Reader in the Methodology of Social Research, Glencoe, Illinois Free Press.
  • Lengermann P.M., 1979, The founding of the « American Sociological Review » : The anatomy of a rebellion, American Sociological Review, 44, 2, 185-198.
  • Lidz V.M., 1986, Parsons and empirical sociology, in Klausner S.Z., Lidz V.M., The Nationalization of the Social Sciences, Philadelphia, University of Pennsylvania Press, 141-182.
  • Lowen R.S., 1997, Creating the Cold War University : the Transformation of Stanford, Berkeley, California, University of California Press.
  • Lundberg G.A., 1936, Quantitative methods in social psychology, American Sociological Review, 1, 1, 38-60.
  • Lundberg G.A, 1955, The natural science trend in sociology, American Journal of Sociology, 61, 3, 191-202.
  • Lynd R.S, 1949, The science of inhuman relations, New Republic, 22-27.
  • Lyons G.M., 1969, The Uneasy Partnership : Social Science and the Federal Government in the Twentieth Century, New York, Russell Sage Foundation.
  • Merton R.K., Lazarsfeld F., (eds.), 1950, Continuities in Social Research : Studies in the Scope and Method of The « American Soldier », Glencoe, Illinois Free Press.
  • Mills C. W., 1959, The Sociological Imagination, Oxford, Oxford University Press.
  • Nichols L.T., 1998, Social relations undone : disciplinary divergence and departmental politics at Harvard, 1946-1970, The American Sociologist, 29, 2, 83-107.
  • Oberschall A., 1978, Paul F. Lazarsfeld and the history of empirical social research, Journal of the History of the Behavioral Sciences, 14, 3, 199-206.
  • Ogburn W.F., 1929, The folkways of scientific sociology (presidential address), American Sociological Society-24th Annual Meeting Papers and Proceedings, 24, 1-11.
  • Platt J., 1996, A History of Sociological Research Methods in America : 1920-1960, Cambridge, Cambridge University Press.
  • Pollack M., 1979, Paul F. Lazarsfeld, fondateur d’une multinationale scientifique, Actes de la Recherche en Sciences Sociales, 25, 45-59.
  • Pollack M., 1980, Paul F. Larazsfeld : a sociointellectual biography, Knowledge, 2, 2, 157-177.
  • Riley J.W., 1949, « The American Soldier » : adjustment during army life, American Sociological Review, 14, 4, 557-559.
  • Schrecker E.W., 1986, No Ivory Tower : McCarthyism and the Universities, Oxford, Oxford University Press.
  • Sewell W.H., 1989, Some reflections on the golden age of interdisciplinary social psychology, Social Psychological Quarterly, 52, 2, 88-97.
  • Shanas E., 1950, « The American Soldier », The American Journal of Sociology, 55, 6, 590-594.
  • Solovey M., 2001, Project Camelot and the 1960s epistemological revolution : rethinking the politics patronage-social science nexus, Social Studies of Science, 31, 2, 171-206.
  • Sorokin P.A., 1933, Recent social trends : a criticism, The Journal of Political Economy, 41, 2, 194-210.
  • Sorokin P.A., 1940, Foundations of sociology, The American Journal of Sociology, 45, 5, 795-798.
  • Stouffer S.A., 1949, The American Soldier, Princeton, Princeton University Press.
  • Stouffer S.A., 1950, Measurement and Prediction, Princeton, Princeton University Press.
  • Turner J.H., 1974, The Structure of Sociological Theory, Honewood, Illinois, Dorsey Press.
  • Turner S.P., 1987, Underdetermination and the promise of statistical sociology, Sociological Theory, 5, 2, 172-184.
  • Turner S.P., 1990, Forms of patronage, in Cozzens S.E., Gieryn Th.F., Theories of Science in Society, Bloomington, Indiana, Indiana University Press, 185-211.
  • Turner S.P., 1993, The end of functionalism : Parsons, Merton and their heirs, Philosophy of the Social Sciences, 23, 2, 228-242.
  • Turner S.P., 1994, The origins of « mainstream sociology » and other issues in the history of american sociology, Social Epistemology, 8, 1, 41-67.
  • Turner S.P., Turner J.H., 1990, The Impossible Science : an Institutional Analysis of American Sociology, Newburry Park, California, Sage Publications.
  • Turner S., Turner J., 1994, Symposium on the history of american sociology portrayed in the impossible science, Social Epistemology, 8, 1, 3-67.
  • United States. President’s Research Committee on Social Trends, MitchellW.C., 1933, Recent social trends in the United States, report of the President’s research committee on social trends.
  • Vidich A.J., 2000, The department of social relations and « systems theory » at Harvard : 1948-1950, International Journal of Politics, Culture and Society, 13, 4, 607-647.
  • Warshay L.H., 1975, The Current State of Sociological Theory : a Critical Interpretation, New York, D. Mckay Co.
  • Wilner P., 1985, The main drift of sociology between 1936 and 1982, History of Sociology : an International Review, 5, 2, 1-20.
  • Winsborough H.H., 1992, Sociology departments and their research centers : an essential tension ?, in Halliday T.C., Janowitz M., Sociology and Its Publics : the Forms and Fates of Disciplinary Organization, Chicago, University of Chicago Press, 269-296.



Platt, 1996, 37 ; Wilner, 1985.


Platt, 1996, chap. 2.


The problematic builds on Ian Hacking’s call for historians of science to focus on the history of styles of reasoning. Cf. Hacking, 1992.


Bryant, 1985, chap. 5.


Mills, 1959, chap. 3.


For a review of US military support for the social sciences during and immediately after the war, cf.: Lanier, 1949. A more historical and insightful analysis can be found in Ellen Herman’s excellent study of American psychology (1995). The impact of the impact of the Cold War on American social science is explored in Schrecker (1986), Lowen (1997), Jardini (2000) and Solovey (2001). For an overview of the changing relation between the federal government and the social sciences cf. Lyons (1969) and Klausner, Lidz (1986).


Sewell, 1989.


For discussions of the development of social psychology as a branch of psychology and more generally cf. Danziger (1992), Katz (1978) and the special issue on the topic in the Journal of the History of the Behavioral Sciences, 2000, 36, 4.


For a more general discussion of sociologists formulations and responses to this dilemma cf. Halliday, Janowitz (1992).


Daston, 1993 ; Hacking, 1992.


Cited by Lerner in Merton, Lazarsfeld, 1950, 217.


Demerath, 1949, 87.


Riley, 1949, 557.


Lynd, 1949.


Glazer, 1949.


Mullins, for example, cites The American Soldier as one of the three "successes" of the postwar mainstream tradition. The other two books being The Structure of Social Action (1937) by Talcott Parsons and Social Theory and Social Structure (1949) by Robert K. Merton (1975, 39).


Converse, 1987, 35-38.


Platt (1996, 99) specifically notes the curious association of survey research with scientism in the immediate post-war period.


The third volume was produced by Carl I. Hovland at the Yale Institute of Human Relations and focuses on experimental studies of communication. It was presented as a contribution to the new field of communications. As such, its reception had little direct impact on the consolidation of sociology as an academic discipline and it will not be discussed herein.


Oberschall, 1978.


Key book length works on American sociological practice, rather than theory, include Converse (1987) and Platt (1996). Turner, Turner (1990) also pay considerable attention to empirical social research in their institutional history of sociology as a discipline. The discussion that follows draws extensively on these texts as well as on Bannister’s excellent analysis of scientism during the interwar period (1987). The history of empirical sociology has also been extensively treated in articles, which will be mentioned where relevant.


For an overview of this discussion, cf. the symposium on « The Impossible Science » by Turner, Turner (1994).


Camic, Xie, 1994.


Turner, Turner, 1994, 42-46.


In making this argument Turner distinguishes between theoretical or intellectual and methodological tradition. Whereas sociology is notoriously fragmented when it comes to theoretical traditions, Turner argues that it is dominated by a mainstream methodological tradition involving « the understanding and use of statistical and measurement techniques to formulate researchable problems and conclude that they have been previously answered » (Turner, Turner, 1994, 49).


For an overview of Giddings’ philosophy of the social sciences and his contribution to sociology, cf. Bannister (1987, chap. 4 and 5).


The Grammar of Science, 1895.


The Social Science Research Council was created in 1923 with funding from the Rockefeller Foundation to facilitate interdisciplinary work and stimulate research. The organization did not conduct research directly, but rather distributed private foundation money to individual projects. In contrast with the National Bureau of Economic Research and the Brookings Institute – two other extra-university research organizations that were established in the same period – the SSRC supported work in sociology and psychology. It was a major actor in encouraging the development of a scientific sociology ; interwar activities included a committee on the aims and methods of research and on social statistics.


Abbott, 1999 ; Lengermann, 1979 ; Kuklick, 1973.


The exceptions, however, were important for they formed the leadership of the postwar period. Scholars obtaining posts in this period include the key figures in the debate around The American Soldier include Samuel Stouffer, David Riesman, Arnold Rose, William Sewell, Paul Lazarsfeld, Talcott Parsons, Herbert Blumer and Robert Merton (Turner, Turner, 1990, 85).


Bannister, 1987, 182.


United States. President’s Research Committee on Social Trends, MitchellW.C., 1933.


There is an extensive literature on Paul F. Lazarsfeld and his contributions. Cf., for example, Barton, 1979 ; Pollack, 1979 and 1980 ; and the recent collection edited by Lautman, Lécuyer, 1998.


Pollack, 1980 ; Capecchi, 1978.


Discussions of Lundberg’s epistemological position can be found interspersed in Hinkle (1994) and Bryant (1985).


Lundberg, 1936, 44.


Platt, 1996, 103.


Ogburn, 1929.


Cited in Bannister, 1987, 176.


Sorokin, 1933.


Sorokin, 1940, 796.


Clausen, 1984, 207-208.


Stouffer, 1949, 20.


Ibid., 32-33.


Cited in Jahoda, 1962, 496.


Lynd, 1949.


Glazer, 1949.


Ibid., 489.


Ibid., 491.


Ibid., 89.


Shanas, 1950.


Ibid., 592.


Ibid., 594.


Stouffer himself straddled the two worlds. He was a student of Ogburn (who was a student of Giddings) and had spent a year in London studying mathematical statistics with R.A. Fisher. At the same time he collaborated extensively with Lazarsfeld.


Stouffer, 1950.


Stouffer, 1949, 35.


Other books subjected to a similar treatment included The Authoritarian Personality (1950) and The Lonely Crowd (1950).


Merton, Lazarsfeld, 1950, 11.


Clogg, Dajani, 1991, 8.


Ibid., 136.


Lazarsfeld, Rosenberg, 1955.


Hyman, 1955.


Merton, Kitt, 1950, 40.


For a contemporary discussion of this convergence cf. Lundberg, 1955.


Merton, Lazarsfeld, 1950, 40.


Bulmer, 1992 ; Winsborough, 1992.


Camic, 1995.


Turner, 1994 ; Collins, 1998.


Turner, Turner, 1990 ; Turner, 1990.


Lowen, 1997.


Turner, Turner, 1990, 96.


The initial proposal grew out of their joint participation in the Harvard based Committee on National Morale, further attesting to the importance of perceived links between science and democracy for the developments under consideration. For a discussion of aspects of this culture cf. Hollinger, 1995.


Nichols, 1998 ; Johnston, 1998.


Lidz, 1986 ; Vidich, 2000.


Johnson, Johnson, 1986.


Halas, 2001.


Clark, 1996.


The Office of Scientific Research and Development had been responsible for the successful development of radar, synthetic rubber and new weapons such as the proximity fuse and, towards the war’s end, the atomic bomb. These accomplishments allowed Bush to argue for the contributions of basic research for military and practical ends. It also gave him enormous influence over the federal government’s new science policy.


Klausner, 1986 ; Solovey, 2001.


Larsen, 1992, 7.


Brick (2000) argues that it also prepared the way for the subsequent focus of sociology on race, gender and ethnicity, at the expense of politics and economics.


Platt, 1996, 33-37.


Vidich, 2000.


Hyman, 1962, 328.


Turner S., 1993, 24 ; Turner J., 1974, 71-73.


For discussions of the problematic relation between theory and statistical models in sociology see: Turner (1987) and Abbott (1988). Warshay makes a similar suggestion in his discussion on the dominance of « small theories » in postwar sociology (Warshay, 1975).



« The American Soldier », a four-volume study published on the basis of war-time research, provides a window into a turning point in the history of American sociology. The book heralded the ascendance of statistics as the authoritative method and an associated rise of instrumental positivism as the dominant style of reasoning. An examination of the history of this exemplar of sociological research situates the association of statistics and survey research in the context of military and government patronage, the emergence of the behavioral sciences and the associated creation of interdisciplinary university-based research centers oriented towards applied social research. The book was written during the prolonged post-war debate over the creation of the National Science Foundation and the inclusion of the social sciences therein. The rhetorical emphasis on the relation of pure theory and applied research and the practical limitation of theory to methodological definitions of concepts found in the book reflects the image of science that emerged from those debates. An important consequence was the consecration of a radical disjuncture between grand theory on the one hand and a form of operationism associated with quantitative sociological research on the other hand.


  • American Soldier
  • survey research
  • quantification
  • instrumental positivism
  • sociology
  • styles of reasoning


L’étude des quatre volumes de « l’American Soldier », ouvrage restituant les recherches conduites durant la Seconde Guerre mondiale par le « Research Bureau of the Information and Education Division » du War Department American, permet de saisir un moment important, en fait un virage, de l’histoire de la sociologie américaine. L’ouvrage annonce l’établissement des statistiques comme méthodes faisant autorité et le développement d’un « positivisme instrumental » comme mode de raisonnement dominant. Le soutien militaire et gouvernemental, l’émergence des sciences du comportement et la création de centres de recherches universitaires interdisciplinaires consacrés aux recherches appliquées constituent le contexte du mariage des statistiques et des recherches par enquêtes empiriques. La rédaction de l’ouvrage s’est faite durant l’après-guerre, au moment des débats sur la création de la National Science Foundation et sur l’intégration des sciences sociales en son sein. L’ouvrage met l’accent, de façon rhétorique, sur la relation entre la théorie et la recherche appliquée alors que l’analyse elle-même réduit la théorie à la définition méthodologique des concepts. Ce contraste annonce l’avènement d’une coupure entre la haute réflexion théorique et les recherches quantitatives en sociologie pénétrées par l’opérationnisme.


  • recherche empirique
  • quantification
  • opérationnisme
  • positivisme
  • sociologie
  • styles de raisonnement

Plan de l'article

  1. I - The Book(s)
  2. II - The specificity of American sociology
    1. Franklin P. Giddings and a qualified empiricism
    2. Paul F. Lazarsfeld, « statistical gadgets » and contract research
    3. George Lundberg and the turn to operationism
  3. III - The Corpus
    1. « The American Soldier » as applied sociology
    2. The response from the discipline
    3. Operationism and the construction of indices
    4. Addressing the criticisms : theory and quantification
  4. IV - Sociology in the late forties
    1. A shift in the (institutional) center
    2. Social Science and the NSF
  5. IV - The new scientific sociology

Article précédent Pages 65 - 94 Article suivant
© 2010-2017