Methodological development in contemporary sociology
At the beginning of the 20th century, interest in developing a sociological methodology grew steadily. Methodological approaches outlined in W.I. Thomas and Florian Znaniecki’s Polish Peasant in Europe and America (vol. 5, 1918–20) were recognized as an important advance, not so much in methodology as in committing sociologists to the task of improving methodology. Thomas and Znaniecki systematically gathered longitudinal data through letters, diaries, life histories, and other relevant documents. Intended to gather specific data to help planners solve social problems, this approach soon became popular. The most ambitious of these “community social surveys” was the two-volume work Great Depression, Recent Social Trends (1933), edited by sociologists W.F. Ogburn and H.W. Odum.
Significant advances in scientific methodology occurred at the University of Chicago in the 1920s. Many studies of the metropolis and its subareas were conducted under the leadership of Robert E. Park, Ernest Burgess, and their colleagues. Most important, hypotheses were developed during the research rather than being imposed a priori (a practice later replaced by theoretically guided research). Many students took part in the studies and contributed to methods and findings.
Ecological patterning
A critical aspect of the Chicago School’s urban research involved mapping locations. These included locations of land values, specific populations (racial, ethnic, or occupational), ethnic succession in neighbourhoods, residences of persons who committed certain crimes, or zones with a high incidence of divorce and desertion. Data-collection methods included participant observation, life histories, case studies, historical information, and life cycles of social movements. Sociologists at the University of Chicago paid equal attention to the improvement of methodology as they developed this approach. Here, for the first time, was a large-scale effort in which theory, methodology, and findings evolved together in an inductive process. Growing from its success in the American Midwest, urban research and zone mapping spread throughout the United States and influenced sociology abroad.
Ecological methods such as urban mapping were also first developed at Chicago, having grown out of the research on the metropolitan region, especially that regarding nonsocial patterns that resulted from the movement of populations, businesses, industries, residences, and institutions as each sought more advantageous locations. Most early urban studies mapped distributions that revealed relationships in general patterns of urban ecology. Because of this, multiple indicators of disorganization, stratification, vertical mobility, and population phenomena were found to follow regularities and could actually be considered predictable to some degree (see demography).
Experiments
Experimental methods, once limited to the domain of psychologists and considered inapplicable to social research, were eventually applied to the study of groups. By the 1930s, social psychologists Kurt Lewin, Muzafer Sherif, and their colleagues had begun conducting experiments on social interaction. Sociologists soon followed their example and set up research laboratories. Notably, Robert F. Bales at Harvard systematically observed interaction in small artificial groups, producing useful results that were replicated elsewhere.
As a rule, successful experiments tend to occur in simple situations in which the variables are limited or controlled. Complex experiments, however, are possible. At Stanford, for example, a series of experiments over 30 years contributed to a formal theory of social status building and maintenance set forth by Joseph Berger and Morris Zelditch in Status, Rewards, and Influence (1985). At the University of Iowa, two decades of laboratory and computer-simulated research on power and exchange in small groups advanced theory in networks and decision making summarized by Barry Markovsky in Social Psychology of Small Groups (1993).
Statistics and mathematical analysis
Sociologists have increasingly borrowed statistical methods from other disciplines. Statistician Karl Pearson’s “coefficient of correlation,” for example, introduced an important concept for measuring associations between continuous variables without necessarily defining the nature of the connection. Later, statistical estimates of causal relations were probed by “multiple regression analysis,” employing techniques that estimate the degree to which any particular variable influences a particular outcome.
Patterns of responses to interview questions, once thought to be purely qualitative, have also been subject to mathematical scaling. A method devised by psychologist L.L. Thurstone in the late 1920s gained popular use in sociology. In this approach a list of items is presented to a number of judges who individually relist them in order of importance or of interest. Items on which there is substantial agreement are then reordered to form a scale. Another technique asks participants to respond to statements by strength of agreement (strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree). Social distance may be measured by asking respondents whether they would accept members of other groups as spouses, close friends, fellow employees, neighbours, or citizens.
A method called sociometry, introduced by J.L. Moreno in the 1930s, collects and tabulates information on group interactions. The interactions studied can appear trivial—for example, who confides in whom, which friends eat lunch together—or they may be more businesslike, such as who might be appointed as a group spokesperson. The data may be gathered by direct observation, interviews, or questionnaires. The preferences each individual has for specific others are then mapped with arrows from sender to receiver, and this results in a diagram of choices for the entire group. The person chosen most often is labeled a “star” or, less often, an “isolate.”
The patterns may be quantified and supplemented with other data to reveal a group’s informal structure. A powerful application of the approach, often mathematized, called network analysis, maps different types of interactions between organizations over extended periods and thus exposes a substructure not revealed from organizational charts or public documents.
Factor analysis, an elaboration of Pearson’s coefficient of correlation, significantly reduces the number of complex variables to be considered. For instance, 50 different questions or measures of work alienation may in fact represent only seven or eight dimensions of alienation. Factor analysis reduces the variables to a more practical number of common factors and then determines each factor’s relative contribution to the outcome variables.
Many other statistical methods have been devised to suit the purposes of such specialties as demography, ecology, social stratification, organizational analysis, mass communication, and social movements. Statistical methods have developed so rapidly that they sometimes outstrip scholars’ ability to find data worthy of their application. Computers have accelerated the application of complex measures that were previously limited by the amount of time required for performing the mathematics. Progress in measurement has been so significant that the American Sociological Association in 1969 established an annual volume entitled Sociological Methodology.