Samuel A. Stouffer
Stouffer, Samuel A.
Stouffer, Samuel A.
Samuel A. Stouffer (1900-1960), American sociologist, was a founder of large-scale quantitative social research. His contribution to the analysis of survey data represents a distinctively American approach to sociology. The four volumes of the collaborative magnum opus, Studies in Social Psychology in World War n, of which the two volumes of The American Soldier (1949) are best known, remain a landmark and a model in the new tradition of mass production in research, emphasis upon quantitative evidence, avoidance of theoretical speculation except in close contact with the data, and close connection with applied problems. The work has also been a prime target of criticism for those who deplore this trend (e.g., the book reviews quoted in Lerner 1950). Stouffer‘s career, as his collaborator Paul F. Lazarsfeld has noted (Social Research . . . , p. xv), coincides with the development of empirical social research in the United States, and his writings reflect its strengths and limitations.
Stouffer was born and brought up in the little town of Sac City in western Iowa, where his father was publisher of the local newspaper. He attended the small local Methodist college, Morningside, and after a year‘s graduate work in English at Harvard, he returned to work for a time on the family newspaper. His career as a sociologist began with graduate work at the University of Chicago, where he received his PH.D. in 1930. There he came under the formative influence of L. L. Thurstone and William F. Ogburn, both pioneers in the development and application of quantitative methods in the study of human behavior. Stouffer‘s quantitative bent was strengthened by his work with Karl Pearson and R. A. Fisher during a postdoctoral year at the University of London. On his return from England he proceeded to take up successive academic positions at the universities of Wisconsin and Chicago. During World War n he served as director of the professional staff in the Research Branch, Information and Education Division, of the War Department, where he mobilized a large staff of research personnel and planned and supervised the extensive studies that culminated in the four volumes including The American Soldier. After the war he established the Laboratory of Social Relations at Harvard University, where he remained until his death.
At the outset of Stouffer‘s career, demography and social statistics were the only aspects of sociology in which a quantitative approach was well developed. These were the areas on which Stouffer‘s early contributions were focused, and his interest in them continued. He was, for example, consultant to the U.S. Bureau of the Census; furthermore, during the year immediately preceding his death he was embarking upon a study for the Population Council in New York. His early studies—for example, his analysis of differential trends in fertility in Catholics and non-Catholics (Social Research . . . , pp. 165-184) and his study, with Lazarsfeld, of the effects of the depression on the family (ibid., pp. 134-153)—are, like his later ones, marked by his strong talent for getting social data to speak with minimal ambiguity. The data for these early studies were obtained from public records, and much ingenuity was required to develop relevant indices and to control for obscuring factors. In his analysis, he always considered plausible alternative interpretations and tried to find ways to reduce interpretive ambiguity.
Whether he worked with available social statistics or, later, with survey data created by the investigator, Stouffer‘s forte lay in this ingenious interplay, close to the data, of interpretive hypothesis and empirical check. Theoretical questions tended to arise directly from the data before him and to return him quickly to the data for further analysis. His personal style of research fitted the stage of precomputer technology, when the investigator, running his sets of data cards through the counter-sorter himself, could quickly adapt his tactics of analysis to the emerging results. Stouffer‘s career ended just as the requirements of modern electronic computers were tending to impose a greater separation between the investigator and his data.
Research in Stouffer‘s style is not mere fact-grubbing, but neither does it often result directly in broader integrations of theory. One of his early contributions to demographic research, however, provided a precedent-setting example of the kind of mathematically formalized small-scale theory that was later to attract much attention in American social research: his theory of internal migration in terms of intervening opportunities (ibid., pp. 68-91), which he subsequently elaborated and tested against additional data (ibid., pp. 91-112). Stouffer‘s model seeks to account for the movement of population between cities, and it asserts, in effect, that the number of people moving between two places will be larger the more opportunities there are in the target area and the fewer the opportunities that are interposed between the target area and the place of origin. According to his initial formulation, all locations within the circular area centering on the city of origin (with the target city falling on the circumference) potentially present intervening opportunities; direction is ignored. His later elaboration takes direction into account in the estimation of intervening opportunities and introduces such further refinements as adjustment for the general accessibility of the target city to migration from sources other than the origin under consideration. The ingenuity of Stouffer‘s treatment of migration lies less in the mathematical formulation than in the ways in which he was able to coordinate the formal model with data available in census statistics.
During the 1930s the methods of survey research had been developed in the United States under auspices that were largely journalistic or commercial. Stouffer‘s use of survey methods in the wartime studies under his direction and the later secondary analysis of the data reported in The American Soldier had a substantial impact on the improvement of survey design and analysis as a research approach that generates data convenient for quantitative treatment. His approach to survey research had several distinguishing features. One was largely a matter of historical accident: reliance upon self-administered questionnaires rather than on field interviews. More important were his consistent practice of controlling statistically for the effects of many variables on the relationships under examination and his analytic preference for basing interpretative conclusions upon replicated analyses rather than on strong statistical tests of single sets of observations. Also characteristic was his attention to the need for mathematically sophisticated indices of attitudes, as reflected in his sponsorship of work by Guttman and by Lazars-feld in developing “scalogram” and “latent structure” analysis (Stouffer et al. 1950), and in his own later contribution of the H-technique of scale analysis as a practical improvement upon Guttman‘s method (Social Research . . . , pp. 274-289).
The theoretical concept given the most emphasis in The American Soldier is characteristic not only of Stouffer‘s analytic ingenuity but also of his somewhat data-bound empiricism: the concept of relative deprivation. This notion, not formally developed, was invoked to account for such otherwise puzzling findings as that when rank, educational level, and length of Army service are held constant, the less the opportunity for promotion afforded by a branch of the Army, the more favorably will members of that branch assess their chances of promotion. The explanation is offered that each individual evaluates his chances of promotion in comparison with others who share his situation. The Air Force private, in a service (objectively) characterized by rapid promotion, feels more deprived than the private in Military Police, where promotions are much less frequent. Similar interpretative principles appear to account for a variety of seemingly paradoxical observations. It remained for Robert K. Merton and Alice S. Kitt (1950) to explicate and generalize the theoretical notions that are involved and to relate them to the concept of reference groups; in this way they developed a conceptually powerful approach to the analysis of social influences on individual persons. Stouffer asserted, as does Merton, that modest small-scale theories are more useful than theoretical systems in the older, grand style that rarely generated predictions accessible to empirical test. Stouffer‘s theorizing was more modest, more closely tied to the data than is Merton‘s, but at a cost.
If Stouffer‘s practically-oriented yet intellectually curious empiricism at once reflected and tended to reinforce a characteristically American orientation in the social sciences, so also did his emphasis on prediction as a touchstone of scientific merit. The wartime studies provide several examples of moderately effective prediction of behavior from attitudinal data (Stouffer et al. 1950). What is missing, however, is concern with the theoretical articulation of the steps that intervene between antecedent and consequent. Prediction thus comes to be valued in its own right, as a matter of practical achievement, rather than as a criterion of the correctness or utility of theoretical formulations.
For Stouffer, the ideal research method was the controlled experiment, in which outcomes are predicted by prior hypothesis and in which variations in the independent variables are manipulated by the experimenter. Ironically, Stouffer never conducted experimental studies himself, although in his wartime role he sponsored the important experimental work on mass communications by Hov-land, Lumsdaine, and Sheffield (1949), and later actively encouraged experimental research in the Harvard Laboratory of Social Relations. Always sophisticated and critical in his appraisal of the essentially descriptive and correlational methods of analysis of which he was master, Stouffer‘s admiration for the magic of experimentation was somewhat naive. In his methodological writings (e.g., Social Research . . . , pp. 290-299), he came close to viewing the experimental method as a royal road to truth, neglecting the ambiguities that bedevil the experimenter in his attempts to translate theoretical variables into empirical operations. Yet progress in the sophisticated application of the experimental method in social research has surely resulted from Stouffer‘s enthusiastic sponsorship.
As the merger of research and theory that both he and Merton had called for became increasingly the order of the day, Stouffer proceeded, in two of his later studies, published in 1949 and 1951, to test a small-scale theory concerning social roles and role conflict (ibid., pp. 39-67). Characteristically, the distinctive contribution of these two studies is in the development of empirical indices of such theoretical concepts as universalism and particularism and in revising the view of role expectations as well-defined social norms, thus yielding more complicated conceptions in closer accord with the facts.
Stouffer‘s work reflected a meliorist concern with the social problems of the day. He hoped that social research would be socially useful; he held that in the long run the social justification of investment in basic research is in its applications (ibid., p. 4). Yet he was deeply concerned lest the exaggerated claims of enthusiasts should jeopardize the long-run prospects for the development and application of the social sciences. About these long-run prospects, he was firmly optimistic. In the shorter run, he saw the danger of dissipated effort and conspicuous failure if social scientists were too eager to solve those social problems for which their tools were still inadequate. Stouffer was ever confident about the prospects of social science, but modest about its present resources.
Nevertheless, on three occasions, Stouffer did direct his skills in social research to matters of urgent social import. He played a major role in carrying out the studies undertaken as part of Myrdal‘s monumental analysis of the American Negro problem (1944) and continued to take an interest in problems of race relations throughout his career. His deployment of social research to meet wartime needs has already been noted. During the postwar episode in which civil liberties in the United States suffered serious erosion under the attack of Senator Joseph R. McCarthy, he planned and analyzed an important survey of American attitudes toward communism and civil liberties (1955).
Stouffer‘s influence, in sum, accentuated certain distinctively American trends in social research: toward quantification and small-scale theoretical formalization, toward large-scale collaborative research organization, toward fusion of pure and applied research interests, away from speculative theoretical synthesis. His writings provided models upon which contemporary practice in survey analysis is built. While he valued empirically based theory, his own preference was for a style of inquiry rather closer to the data than is optimal for the development of theoretical conceptions with powerful generality. Stouffer was a master of such empirical analysis, making social data answer to his questioning as they would never “speak for themselves.”
M. Brewster Smith
[For the historical context of Stouffer‘s work, see the biographies ofFisher, R. A.; Ogburn; Pearson; Thurstone; for discussion of the subsequent development of his ideas, seeAttitudes; Public Opinion; Reference Groups; Scaling; Survey Analysis; and the biography ofHovland
1949 Stouffer, Samuel A. et al. The American Soldier. 2 vols. Studies in Social Psychology in World War II, Vols. 1-2. Princeton Univ. Press. → Volume 1: Adjustment During Army Life. Volume 2: Combat and Its Aftermath.
1950 Some Afterthoughts of a Contributor to The American Soldier. Pages 197-211 in Robert K. Merton and Paul F. Lazarsfeld (editors), Continuities in Social Research: Studies in the Scope and Method of The American Soldier. Glencoe, 111.: Free Press.
1950 Stouffer, Samuel A. et al. Measurement and Prediction. Studies in Social Psychology in World War II, Vol. 4. Princeton Univ. Press.
(1955) 1963 Communism, Conformity, and Civil Liberties. Gloucester, Mass.: Smith.
Social Research to Test Ideas: Selected Writings. New York: Free Press, 1962. → Contains papers published between 1935 and 1960. A bibliography of Stouffer‘s writings appears on pages 301-306.
Hovland, Carl I.; Lumsdaine, Arthur A.; and Sheffield, Frederick D. 1949 Experiments on Mass Communication. Studies in Social Psychology in World War II, Vol. 3. Princeton Univ. Press.
Hyman, Herbert H. 1942 The Psychology of Status. Archives of Psychology No. 269.
Lazarsfeld, Paul F. 1962 Introduction. In Samuel A. Stouffer, Social Research to Test Ideas: Selected Writings. New York: Free Press.
Lerner, Daniel 1950 The American Soldier and the Public. Pages 212-251 in Robert K. Merton and Paul F. Lazarsfeld (editors), Continuities in Social Research: Studies in the Scope and Method of The American Soldier. Glencoe, III.: Free Press.
Merton, Robert K.; and Kitt, Alice S. 1950 Contributions to the Theory of Reference Group Behavior. Pages 40-105 in Robert K. Merton and Paul F. Lazarsfeld (editors), Continuities in Social Research: Studies in the Scope and Method of The American Soldier. Glencoe, 111.: Free Press.
Myrdal, Gunnar (1944) 1962 An American Dilemma: The Negro Problem and Modern Democracy. New York: Harper. → A paperback edition was published in 1964 by McGraw-Hill.
Young, Donald 1961 In Memoriam: Samuel Andrew Stouffer, 1900-1960. American Sociological Review 26:106-107.
Samuel A. Stouffer
Samuel A. Stouffer
Samuel A. Stouffer (1900-1960) was an American sociologist and statistician. He was among the leaders in applying rigorous methodology to sociological investigations.
Samuel Stouffer was born in Sac City, Iowa, on June 6, 1900. After receiving a bachelor's degree at Morningside College in lowa (1921), he took graduate work at Harvard University and then at the University of Chicago, where he obtained a doctorate in sociology (1930). He taught at the universities of Chicago and Wisconsin and worked with several governmental agencies during the 1930s. During World War II he was director of the professional staff of the Information and Education Department of the War Department. In that position, he directed an important series of studies on attitudes of servicemen. In 1946 he went to Harvard, where he was director of the Laboratory of Social Relations and professor of sociology until his death on Aug. 24, 1960.
Stouffer's work was marked by a dominant interest in the use of varied research techniques, rather than a sustained focus on one or two topical areas in sociology. His doctoral dissertation dealt with the relative merits of the case-study method and the statistical approach. In the 1930s he critically reviewed data on marriage in his Research Memorandum on the Family in the Depression (1937), and he began to investigate the area of opinion research and mass communications, incorporating his findings in a chapter of Paul F. Lazarsfeld's Radio and the Printed Page (1940).
A sophisticated use of statistical methods by Stouffer was first widely recognized in his analysis of factors in migration (1940). He theorized that the number of migrants between two communities not only was influenced by the opportunities at the receiving community, but was modified or reduced by the presence of opportunities between home community and potential destination. By ingenious use of local rental data, he was able to obtain enough confirmation to stimulate several comparable studies of this basic problem that accurately described and partially explained migrant patterns in the United States.
Stouffer's stature as a research sociologist rests on his experience with survey research techniques, which came to fruition in studies on attitudes and difficulties of American military men during World War II. These were published in four volumes known as The American Soldier (1949-1950), with Stouffer as leading researcher, editor, or contributor. Stouffer and his associates not only developed useful research techniques (such as scalogram analysis), but demonstrated the importance of relativity in people's judgments of both their rewarding and frustrating social experiences.
In the last decade of his career, Stouffer turned to the study of attitudes in situations of conflicting values and roles. He became interested in the actual barriers to educational advancement and mobility among youngsters and, more generally, in the study of the compromises made by people faced by inconsistent moral directives. His major work in this area was the national survey on differences in tolerance of nonconformity, published as Communism, Conformity, and Civil Liberties (1955). Stouffer was able to show that tolerance was connected with education, urban residence, and personal optimism.
Several of Stouffer's articles and addresses are collected in his Social Research to Test Ideas: Selected Writings (1962), which includes a summary of his career by Paul F. Lazarsfeld. □