Items Vol. 21 No. 4 (1967)

Page 1




AMONG the activities undertaken by the European Association for Experimental Social Psychology since its establishment three years ago, the organization of summer research training programs has occupied a prominent place. Although many European students of social psychology are familiar with the relevant literature and _pave mastered many of the technical aspects of experimental research, the Association recognized immediately that in order to hasten the development of active centers of research in experimental social psychology throughout Europe, intensive research training should be made available. The first European Summer School on Social Psychology accordingly was held two years ago in The Hague, as reported in Items) December 1965. Partly on the basis of the recommendations made in the reports on the Summer School, by Maarten R. van Gils and Jaap Koekebakker 1 and by Sibe Souten• The first author, Professor of Social Psychology and Director of the Laboratorium voor Experimentele Sociale Psychologie at the University of Louvain, served as Dean of the Training Seminar, and the second author, Lector of Social Psychology at the University of Leiden, served as Associate Dean. The Seminar was sponsored by the European Association for Experimental Social Psychology in collaboration with the Council's Committee on Transnational Social Psychology (Leon Festinger, Stanford University, chairman; Harold H. Kelley, University of California, Los Angeles; Jaap Koekebakker, Netherlands Institute for Preventive Medicine, Leiden; John T. Lanzetta, Dartmouth College; Serge Moscovici, University of Paris; Ragnar Rommetveit, University of Oslo; Stanley Schachter, Columbia University; Henri Tajfel, University of Bristol; staff, Jerome E. Singer). The Seminar was financed largely by funds granted to the Council by the National Science Foundation. The planning committee consisted of Serge A Moscovici (chairman), Jozef M. Nuttin, Jr., and Stanley Schachter. . . l"The First European Summer School on Social Psychology, The Hague, July 15 - August 11, 1965," /telliS, December 1965, pp. 50-54.

dijk,2 the 1967 Training Seminar was conceived and organized in a somewhat different way: specifically as a residential seminar in which the participants would have the opportunity to develop-in collaboration with a "master"-an original hypothesis, to design an experimental test of it within a laboratory context, to conduct the experiment, and to analyze and present the results in a written report. In essence the purposes of the Training Seminar were: to provide intensive firsthand experience with the entire research process; to help the participants, who were all engaged in research, to solve their own specific problems; and to encourage the development of a professional milieu which would support basic research activities in the field. To accomplish these aims, the core of the training program was conceived as the research apprenticeship system-a scheme implying the continuous and intensive involvement of small groups of participants in the research programs of senior faculty members. ORGANIZATION OF THE SEMINAR The main features distinguishing the 1967 Training Seminar from the 1965 Summer School may be summarized as follows: In the earlier program, lectures and research were organized within the framework of the general theme "social psychological aspects of organizations." In 1967 2 "Evaluation of the European Summer School on Social Psycho· logical Problems in Organizations (1965) Based on Survey Data," Institute of Social Psychology, University of Utrecht, 1966 (mimeo.).


there was no lecturing and no over-all theme; the emphasis was on formal training in experimental research. The faculty of the Summer School changed each week and consisted, apart from the Director and Secretary, of two senior members. In 1967 research teams were directed by a permanent faculty of six members; the number of participants was reduced from 40 to 30; and the duration of the program was increased from 4 to 5 weeks. The participants, who were generally advanced research assistants, were appointed in consultation with the directors of the various research centers in social psychology throughout Europe, without particular regard to geographic representation which had been a factor in the selection of participants in 1965. The preparation for the Seminar was intensive for both faculty and participants. s The faculty held two preliminary meetings, in Los Angeles in November 1966 and in Louvain in April 1967. The prospective participants were required to prepare themselves by reading a number of general works selected by the faculty at its first meeting, as well as specific background 8 The faculty consisted of Harold H. Kelley and Ragnar Rommet路 veit of the committee, and Harold B. Gerard, University of California, Riverside; Jaap Rabbie, University of Utrecht; Robert B. Zajonc, Uni路 versity of Michigan; and Philip G. Zimbardo, New York University. The participants were: Jean-Claude Abric, Assistant Professor of Social Psychology, Universite d'Aix-Marseille, Aix-en-Provence; Jacques Allegro, Research Assistant, Bureau Sociaal Psychologische Zaken, Royal Dutch Navy; Jan Beijk, Research Assistant, University of Amsterdam; Vera Bokorova, Research Worker - Psychologist, Institute of Psychology, Czechoslovak Academy of Sciences; Alistair Chalmers, Lecturer in Social Psychology, University of Sussex; William Cheyne, Lecturer in Psychology, University of Strathclyde; Malcolm Cook, Research Student, University of Oxford; Janina Frentzel-Zagorska, Teaching Assistant, Department of Sociology, University of Warsaw; Roberto Gentile, Psychologist, University of Naples; Nenad Havelka, Assistant, Department of Psychology, University of Belgrade; Zdenek Helus, Scientific Worker, J. A. Komenskeho Pedagogical Institute, Czechoslovak Academy of Sciences; Paul Henry, Chef de travaux, Groupe de psychologie sociale, Ecole Pratique des Hautes Etudes; Werner Herkner, Research Associate, Institute for Advanced Studies, Vienna; Renee Honai, Aspirant, Fonds National de la Recherche Scientifique, Brussels; John Michael Innes, Lecturer in Psychology, University of Birmingham; Helmut Lamm, Research Associate and Lecturer, Institut filr Sozialwissenschaften, Wirtschaftshochschule, Mannheim; Freddy Lange, Scientific Coworker in Social Psychology, University of Amsterdam; Jacques-Philippe Leyens, Assistant, University of Louvain; Wolfgang Manz, Dipl. Psych. Wissenschaftliche Assistent, Bruhl, Germany; Pierre Miollan, Assistant in Psychology, Universite d'Aix-Marseille, Aix-en-Provence; Martine Naffrechoux, Technical Collaborator, psychosociological study group, University of Paris; Michel P~cheux, Research Assistant, University of Paris; Anna Potocka-Hoser, Research Assistant, Institute of Philosophy and Sociology, Polish Academy of Sciences; Marc Servais, Assistant in Experimental Sociopsychology, University of Louvain; Olav SUrdal, Lecturer in Social Psychology, University of Oslo; Betty Hughes Swift, Ph.D. candidate in psychology, London School of Economics and Political Science; Willy Vande Capelle, Ph.D. candidate in psychology, University of Louvain; Dik Van Kreveld, research and teaching in social psychology, Vries, The Netherlands; Peter Veen, Officer, Social Psychology Branch, Royal Dutch Navy; Maryla Zaleska, Research As路 sistant, Laboratory of Social Psychology, University of Paris.


material later sent by each faculty member to his research team, and by increasing their facility in reading, understanding, and speaking English. All participants had at least a master's degree in psychology (obtained..on the average 3 or 4 years ago); 5 of them held the. Ph.D. Thirteen of the participants were research and/or teaching assistants at their home universities; 12 held positions as research workers; and 5 were university lecturers. The faculty at its second preliminary meeting assigned each participant to one of six research teams, mainly on the basis of expressed preferences for research topics or team leaders. Since each team would have to write a report in English and work with Dutchspeaking subjects, the teams were organized to include at least one member whose native language was English and at least one whose native language was Dutch. Besides the team leader and 5 participants, each team had a participant assistant who was familiar with the local facilities. Before the actual start of the Seminar, each participant was invited to communicate with the leader of his team and to comment on a general outline of its research area or a preliminary proposal for an experiment to be carried out during the Seminar. It is hardly possible to describe briefly the work accomplished during the Seminar period. The participants and staff lived at the Leo XIII Seminar, a college near the laboratory where the formal sessions were held, and worked together almost without regard to time. The general work schedule for each team was approximately as follows: during the first week of the Seminar one or more concrete research proposals were planned and discussed; during the second week pilot experiments were conducted; the main experiments and analyses of data were carried out during the third and fourth weeks; by the end of the fifth week the first drafts of the research reports were written. Nine laboratory experiments were actually carried out. Needless to say, this required the cooperation of many persons. The Psychologisch Instituut of the University of Louvain had made available more than 100 experimental and office rooms. In addition to the 6 participant assistants, 10 members of the social and technical staff collaborated full time. 4 More than 1,000 members of the military services served as subjects. 5 Most working days ended after midnight.


4 The Seminar is highly indebted to the Academic Council of the Catholic University of Louvain, Leuven-Nederlands, and its Instituut voor Psychologie en voor Pedagogische Wetenschappen, which contributed in various important ways to the success of the Seminar. IS Thanks are also due to the Belgian Army and especially to the Kazerne kwartier De Hemptinne of Heverlee for its unqualifiable co-. operation in providing experimental subjects throughout the seminar. period.






Team led by Harold Gerard. This team chose to work in the general area of dissonance theory and devoted considerable time to discussion of the literature and to formulation of a particular problem that could be investigated in the Louvain setting. Thorough discussion of the most crucial experiments related to the theory revealed many differences in procedures in these experiments, which could explain contradictory results reported in the literature. The specific focus of the team was on forced compliance. The first order of business therefore was to identify several experimental tasks that would be unpleasant for the subject population available. The relevant literature indicates that a weakness in many experiments has been the failure to provide such a task-a necessary requirement for testing derivations from the theory. Two unpleasant tasks which were quite different in nature were finally selected by the team, which then discussed several possible research designs. The design selected was one in which the degree of justification for working on one of the two unpleasant tasks was varied. Also varied was the degree of choice of task, when justification was offered, and the point in time when the subject'S evaluation of both tasks was measured. The design, which involved a double nesting with repeated measures on half of the subjects, provides for investigation of decision making in general rather than the special case of forced compliance. Preliminary analysis of the data obtained shows that the effect of justification was in line with the dissonance theory derivation only when the subject'S evaluations of the two tasks were measured before but not after he actually engaged in the task for the second time (he had already had experience with both tasks in the initial phase of the experiment). Team led by Harold Kelley. In the original proposal sent to the members of this team the leader suggested the general area of negotiation and bargaining as a topic of research, and more specifically the context of a particular mixed-motive, incomplete information game. The work of the team started, as suggested, with a discussion of this experimental task (game), which despite its limitations was thought to be of great potential for research in this area. In subsequent discussions several research proposals were made and modified, and these were pretested in pilot experiments. The team finally settled upon a 2 x 2 design in which order of difficulty of the bargaining problems (whether an ascending or descending order) was路 compared for two types of sources of the problems (whether the problems were selected by a competing pair of bargainers or by a neutral person, the experimenter). DECEMBER


Certain difficulties encountered in carrying out the experiment were attributable in part to the fact that soldiers who first took part in the experiment could tell future subjects about its nature. Thus preliminary analysis showed that regardless of the experimental condition the subjects very often chose a strategy in which one player let the other win the maximum amount of money, probably with the intention of sharing the gain afterwards. This factor along with other factors associated with the soldier population (intellectual abilities, prior social relationships) seemed to result in extremely high withincondition variability in bargaining behavior. As a consequence, tests of the success of the experimental manipulations were only marginally significant, and the predicted variations in dependent variables were not obtained. Team led by Jaap Rabbie. The area of intergroup relations had been suggested as the general framework for the experiment to be carried out by this team. During the discussions in the first week of the Seminar, it became clear that the effect of expectation of future interaction with other people could be regarded as a minimal condition for differentiation of in-group from out-group. A design was developed in which this expectation was varied as well as whether a group's interaction was observed or not, either by a competitive or a cooperative group. In a pilot test it turned out to be very difficult to differentiate by normal experimental manipulations between in-group and out-group subjects. In a second pilot study an attempt was made to test the hypothesis that the perception of belonging to a group and the sentiments about the group and oneself tend to be in a balanced state. High and low self-esteem were induced by telling the subjects after they had taken an intelligence test that they had performed very well or very badly. Although this procedure was highly successful, the results were unsatisfactory. The general variable of self-esteem was investigated again in the next experiment, but the expectation of future interaction was replaced by the related variable of social comparison. In a third pilot test both the design and manipulations were found to be satisfactory, and in the final experiment the effect of self-esteem in social comparison under conditions of competition and no competition was investigated. It was found that subjects prefer to compare themselves with weaker opponents only in a competitive situation when they themselves are not improving and the weak opponent is improving. In all other conditions (subject improving and/or no competition) subjects prefer to compare themselves with 43

stronger opponents. Further analysis had still to be carried out at the end of the Seminar. Team led by Ragnar Rommetveit. This group was asked to choose between two different, though related, fields of experimental inquiry. The members expressed so much interest in both fields that it was decided to carry out two experiments. The first experiment had to do with the joint contribution of extra-linguistic and linguistic context and utterance in the transmission of messages. Since very little had been done previously in this area, which represents an intersection of social psychology and psycho linguistics, entirely novel methods had to be developed. Furthermore, demands concerning explicit coding criteria were accentuated since both verbal stimulus material and responses (recall data) pertained to a foreign language (Dutch). The second experiment had to do with the "order effects in impression formation," and more precisely with the question: "What happens if specific nouns are added before or after a series of adjectives?" The results of both these studies were so encouraging that several new experiments along the same lines were discussed, and possible plans for following up the study at the home institutions of the team members were considered. Team led by Robert Zajonc. The research undertaken by this team was an attempt to separate the energizing and directive effects of the presence of other persons on the behavior of a subject. According to Zajonc's theory from which the work of the team started, the mere presence of another is a source of general arousal and it enhances and energizes dominant responses. However, the other person can also provide cues which can direct the subject'S behavior. Three experiments were carried out within the framework of this general theory. In the first experiment the influence of a coactor, who received stimuli in synchrony or in asynchrony with a subject in a test of reaction time, under conditions of regular or random presentations of stimuli, was investigated. In the second experiment reaction time was measured under conditions of presence or absence of another person. In the third experiment more and less reliable cues were presented to the subject who performed a vigilance task. Under one condition these cues were presented to the subject by another person working on a similar task; under another condition the sounds of the other person, who was in a different room, were transmitted to the subject. Team led by Philip Zimbardo. This team decided to study the antisocial consequences of loss of personal identity of individuals in group settings. During the first week the team engaged in an intensive analysis of the scant literature available on this phenomenon and 44

also speculated about natural occurrences of it. Because the team had less research experience and less familiarity with English than the other teams, it was decided to begin a pilot study by the end of the first week in order to work on something concrete. The anonymity WI of the subject and of the "victim" were the independent variables manipulated, and the dependent variable was duration of shock delivered by the subject to the victim (with an acceptable rationale). The wearing of hoods, dim illumination of rooms, and related devices were used to increase feelings of lack of individuality. It was predicted that such operations would lower restraints on antisocial behavior and result in greater aggressiveness. This deindividuation effect should be greater when the victim also has little personal identity. Contrary to expectations and also to pretest results obtained with American students, subjects in anonymous condition suffered less shock than subjects whose identity was not masked; and most aggression was shown when both victim and subject were in identifiable condition. Understanding of these results depended upon an explanation of the nature of the subjects, since soldiers in uniform who came to the laboratory in natural groups were already somewhat depersonalized. The laboratory operations perhaps created more self-awareness and anxiety than anonymity in these subjects, and increased their individuation. A model was proposed to clarify and classify the many varieties of deindividuation phenomena, and suggestions for a number of experiments were advanced.



EVALUATION Did the Training Seminar achieve the objectives formulated in advance by the planning committee and the faculty? This is of course a very difficult question to answer, especially if one keeps in mind that the most important long-term aims of the Seminar essentially refer to the future development of active research careers. Although this ultimate question cannot be answered now, there is no doubt that the Seminar quite successfully accomplished its intermediate objectives. In reviewing the proposal submitted to the National Science Foundation and the minutes of the two preliminary meetings of the faculty, one is surprised how closely the planned program was carried out. In general, much more was accomplished than had been hoped for. Nine full experiments, performed with more than 1,000 subjects, are far more than was envisioned in the original proposal: "In the final weeks of the training program, we hope that each team will have pretested an experimental design, modified it as experience dictated, and run at least a few subjects in a final design." VOLUME




It is clear, however, that a 5-week period is still too short for such an enterprise, and one should not be surprised by the general complaint of the participants that there was not enough time to sit back and reflect about the research plans and results obtained. We believe that this complaint is justified, but also that such a situation can hardly be avoided when a process that normally takes months or even years has to be compressed into 5 weeks. Another complaint, namely, the lack of formal contact between teams, also referred to the same pressure of time. Each team became so deeply involved in its own research plans that there was hardly any time for plenary sessions. At the end of the Seminar each team organized one formal presentation of the work it had done. These sessions were the only formal supplement to the informal communication among teams during the preceding weeks. That the level of sophistication of the participants was strikingly high became clear during all phases of their investigations. The research proposals first made by the faculty were in some cases drastically modified through incorporation of original suggestions of participants. In carrying out the experiments, participants made several novel contributions to procedure and instrumentation. The analyses made at the end also showed the high degree of statistical competence of many participants.

Language difficulties, which were expected on the basis of experiences at the 1965 Summer School, did arise to some extent in all groups. In five of the six teams, however, there was only one member who had difficulty in expressing himself in English. In the sixth group a majority of the participants had language difficulties, but this group was composed in this way at the explicit request of the faculty member in charge. T!lis seems to have been a rather happy procedure since language difficulties were similar for most of the members; the only adverse effect was that discussions did not proceed as rapidly as in the teams in which a majority of the members did speak English fluently. The enthusiasm with which the participants and staff worked during the Seminar and the close relationships which were established among all residents seem to justify the expectation that the Seminar would succeed in fostering conditions necessary for stimulating further research in the field of experimental social psychology in Europe. Finally, it should be noted that from a pedagogical point of view this Seminar was, as far as we know, unique as a training device. It was the firm conviction of most of the participants that those interested in the methodology of advanced research training should learn much from the rich experience gained during this highlevel Seminar.


THE Council's Committee on Areas for Social and Economic Statistics was appointed in November 1964 and terminated in September 1967 upon completion and review of the major project it had sponsored: a re-examination of the criteria by which the present Standard Metropolitan Statistical Areas have been defined to provide uniform areas for the publication of Census and other data relevant to metropolitan problems, and an evaluation of alternative principles of classification, such as the concept of functional economic areas. The committee also sponsored an exploratory conference on spatial aspects of human behavior in October 1965, and • Based on the final report of the Committee on Areas for Social and Economic Statistics. The members of the committee were: Karl A. Fox, Iowa State University (chairman); Brian J. L. Berry, University of Chicago; Lester R. Frankel, Audits Be Surveys Company; John Friedmann, Massachusetts Institute of Technology; W. L. Garrison, University of minois at Chicago Circle; Britton Harris, University of Pennsylvania; Donnell M. Pappenfort, University of Chicago; and Conrad Taeuber, Bureau of the Census. DECEMBER


considered means of increasing the quality of Census and similar materials, as well as the flexibility and economy of access to them. STUDY OF PRINCIPLES OF METROPOLITAN AREA CLASSIFICATION The committee's major study, planned and initiated in 1965, was directed by Brian J. L. Berry and conducted at the University of Chicago with support provided by a contract between the Bureau of the Census and the Council. The study resulted in a report, "Functional Economic Areas and Consolidated Urban Regions of the United States," which (after preliminary review of its findings at a conference in Washington in December 1966) was transmitted in May to the Bureau of the Census and Bureau of the Budget for further study and possible future action. The principal findings of the study were formally approved by the committee


at a meeting on September 1, and are summarized in the following paragraphs. Three sets of criteria were used by the Bureau of the Budget in 1960 in an attempt to redefine Standard Metropolitan Statistical Areas that would conform to the general concept that a metropolitan area is essentially a large integrated economic and social unit-a county or a group of contiguous counties-with a recognized large population nucleus. The most basic criteria were (a) that the SMSA include a legal central city of at least 50,000 population, or "twin cities" totaling 50,000; (b) that 75 percent of the labor force of each county included be nonagricultural and live in contiguous minor civil divisions with a population density of at least 150 persons per square mile; and (c) that at least 15 percent of the workers in each county included commute to the central city. Each of the criteria has been the subject of criticism. For example, 50,000 has been said to be both too small and too large, and the use of the legal central city rather than an urbanized area has been challenged. Some have said that the urban-rural distinctions implied in the criteria of metropolitan character have no meaning in a society whose way of life is becoming almost completely urbanized. Similarly, the 15 percent cutoff on intensity of commuting has been said to make little sense since it excludes part of the metropolitan labor market. The study directed by Berry found, however, that the 1960 classification of SMSA's stems not from all three criteria, but fundamentally from only the first two-size and metropolitan character. The size criterion determined how many SMSA's there would be, and that of metropolitan character determined which contiguous counties (if any) would be joined with the central counties. In effect, the map Standard Metropolitan Statistical Areas of the United States, prepared by the Bureau of the Budget, thus presents a uniform regionalization of the country divided between "metropolitan" and "nonmetropolitan" categories and with the former divided into more than 200 segments. The principal import of criticisms of the 1960 area classification is twofold. First, visual criteria, such as density and contiguous subdivision, are no longer regarded as relevant for purposes of area classification because-whatever the outward appearance-society, economy, and way of life are all highly urbanized. Second, meaningful integrated social and economic areas must be far more extensive than the sections of the United States classified as SMSA's in 1960. If labor markets, retail and wholesale shopping patterns, communication by mass media or any other index of integration are examined, one will find that the entire country consists of a set of functional economic areas centered on 46

urbanized areas. Further, with improvements in transportation and communication, these FEA's are being transformed rapidly into urban realms which are characterized not by a single central city but by a specialized, multifocal organization. These criticisms indicated the need for a detailed analysis of the feasibility of subdividing the country into integrated socioeconomic areas. There had been no prior complete, consistent, comparative analysis of the spatial organization of the United States into functional economic areas. Rand McNally produces a map which allocates the counties of the United States into "Basic" and "Major" wholesale trading areas; Bogue and Beale have subdivided the country into state economic areas; 1 and reports dealing with specific parts of the country have been published, for example, by the Upper Midwest Economic Study. Also, federal agencies continue to define exhaustive sets of service areas, and state Labor Departments produce reports on commuting patterns and labor markets. A considerable gap in our knowledge of the country was evident, however. COMMUTING PATTERNS, 1960 An original analysis was needed of the functional regionalization of the United States in 1960, based on criteria of integration. Here, fortunately, the Bureau of the Census provided a rich supply of unpublished journey-to-work data from the 1960 Census. A regionalization was sought that would classify the United States into a set of economic areas based on the commuting behavior of the population in 1960 (i.e., on linkages between place of residence and place of work). In the study a 43,000 x 4,300 data matrix was analyzed in which the workers residing in the 43,000 census tracts and "pseudo-tracts" of the United States (standard location areas) had been cross-classified by place of work according to a list of 4,300 possible workplace areas. Unfortunately, there were problems of both sampling error (the data came from the Census 25 percent sample) and systematic bias to contend with, but with these limitations it was possible to define the "commuting fields" and "labor markets" of the United States. On this basis functional economic areas were defined, first, for the set of central cities of the SMSA's recognized in 1965, and then for additional independent regional centers of less than 50,000 population in the less densely settled areas of the country. In addition, consolidated urban and metropolitan regions were created out of groups of labor markets, to take account 1 Donald J. Bogue and Calvin L. Beale, Economic Areas of the United States, Glencoe: Free Press, 1961.





of cross-commuting. Considerable experimentation led to the following definitions: Commuting field: an area encompassing all standard location areas sending commuters to a designated workplace area. The field varies in intensity according to the proportion of resident employees in each SLA commuting to the workplace, and may be depicted cartographically by contours that enclose all areas exceeding a stated degree of commuting. Labor market: all counties sending commuters to a given county. Central county: the designated workplace area for definition of a labor market. Central city: the principal city located in a central county. Functional economic area: all those counties within a labor market for which the proportion of resident workers commuting to a given central county exceeds the proportion commuting to alternative central counties. Metropolitan economic area: an FEA in which the population of the central city exceeds 50,000, or in which there are twin cities satisfying criteria of existing SMSA definitional practice. Consolidated metropolitan region: two or more FEA's and/or MEA's (at least one must be an MEA) in which at least 5 percent of the resident workers of the central county of one commute to the central county of another. Consolidated urban region: two or more FEA's and/or MEA's in which 5 percent of the resident workers of any part of one commute to the central county of one of the others.

Maps depicting the extent and complexities of interdependence among areas of the United States were prepared. Examination of these yielded the following conclusions: (1) Commuting fields (FEA's that enclose both place of residence and place of work) are far more extensive than the areas classified as SMSA's in 1960. (2) In the more densely settled parts of the country, commuting fields are not mutually exclusive, but overlap in complex and extensive ways. (3) Independent regional centers of less than 50,000 population are the hubs of labor markets in the less densely settled sections of the country, paralleling in their role centers of greater population where settlement is thicker. (4) With the exception of national parks, public lands, and areas with extremely low population densities, the entire area of the United States is covered by the network of commuting fields. It was found that 95.85 percent of the population of the country lived within the set of FEA's and MEA's ultimately defined-86.62 percent in the MEA's --compared with the some two-thirds of the population that was counted in the 1960 SMSA's. Almost the entire population of the United States lived in areas in which at least some portion of the residents had jobs in large urban centers. Further exploration of commuting between outlying areas within the larger commuting fields of central cities led to two further conclusions: (a) A central county containing a central city and other area is an appropriate focus for a single commuting field, because the DECEMBER


individual commuting fields of the two components are virtually identical, and because the commuting fields of all outlying counties nest within that of the central county. (b) Labor markets made up of county units are sound approximations to commuting fields defined on the basis of tract (SLA) data, involving relatively little loss of information. On the assumption that it remains useful to construct labor markets with county units, FEA's can be defined most readily from a county-to-county commuting matrix. To ensure a mutually exclusive allocation of counties to FEA's, the greatest percentage flow seems the simplest and most logical criterion. (If a populationsize distinction is desired, it can be applied by differentiating some subset of the FEA's, e.g., MEA's focusing on an SMSA.) Use of county-to-county commuting data permits allocation of all the settled parts of the United States into a set of functional economic areas. In some parts of the country there is substantial crosscommuting. Recognition of this is possible in a consistent set of consolidated regions. These may be defined by combining MEA's and/or FEA's that evidence significant degrees of cross-commuting. IMPLICATIONS AND RECOMMENDATIONS The uniform regionalization of the 1960 SMSA's and the functional regionalization evidenced by commu ting behavior are significantly different. The Bureaus of the Budget and of the Census thus face a major choice, for the 1960 classification does not produce fully integrated areas with a large population nucleus. Is the intention to classify areas on the basis of how they look? In this case, continuation of present practice will suffice. Alternatively, should the areas embrace people with common patterns of behavior? If so, commuting data which deal with daily behavior and the links between place of residence and place of work are relevant. Comparability is not the issue if county units are used. Besides, there has been no consistency in definitional practice since inception of attempts to define metropolitan areas. Nor should consistency be expected in a dynamic socioeconomy in which patterns of organization and behavior are subject to continuing change. The problem of choice is difficult since there is general agreement that some form of area classification will be required for publication of summary statistics for some time to come. The report of the study concludes with the following recommendations: 1. That counties or equivalent units be retained as the basis of any area classification, in all parts of the country. 2. That county-to-county commuting data be the basis


of the classification of counties into functional economic areas. 3. That functional economic areas be delineated around all counties containing central cities of more than 50,000 population, and also be created for smaller regional centers in the less densely populated parts of the country. 4. Where significant cross-commuting takes place, functional economic areas should be merged into consolidated urban regions. 5. Studies should be undertaken to determine whether additional criteria of integration (for example, wholesaling) might lead to realistic merging of smaller western functional economic areas into larger urban regions, to exhaust the land area of the country, just as the FEA's embrace all but 4 percent of the population; and also to satisfy some minimum total population for an economic region.

tate aggregating data for geographical areas suited to particular problems and to facilitate uses of Census data to augment information collected under areal classifications other than those conventionally in use by the Census. The committee discussed the rich variety of crossclassifications and comparisons that may be made with Census data and expressed interest in programs that would facilitate tabulations and cross-classifications other than those now published by the Census. The committee recommended that the Bureau of the Census extend its plans for use of geographical coordinates to include all censuses and the whole country. Implementation of present plans and programs, subject to the availability of funds, would result in identification by such coordinates of about 60 percent of the respondents in the 1970 Census of Population and Housing. Geographical coordinate coding must also be used in other censuses if its full value is to be realized.



In connection with the analysis of functional geographical areas, the committee discussed uses of Census materials in scientific inquiry in the social and behavioral sciences, in effectuation of social and economic development programs, in urban planning, and in resolution of marketing and business location problems. The committee was well aware of the steady improvements that have been made in the quality and quantity of Census data, but believed far greater developmental efforts to be necessary. It urged that every effort be made to hasten the availability of Census materials and the flexibility of access to them. The pace of socioeconomic change is accelerating, and needs for more and better information at the earliest possible time are felt by analysts and policy makers both in the public and private sectors. Social scientists and others are attacking new kinds of problems or are attacking old problems in new ways. The information base for much of this investigation could be greatly strengthened by improvements in the availability of Census data. The committee strongly endorsed the intention of the Bureau of the Census to make data from the 1970 Census more freely and flexibly available to users. In this connection it recommended that special attention be given to preparation of basic tapes and other documentation, and to general purpose programming which will be necessary for economical access to these tapes. DATA WITH GEOGRAPHIC COORDINATES The committee was particularly interested in identification of data using geographic coordinates, to facili48

The committee's conference on spatial aspects of human behavior brought together experts in a variety of academic and related disciplines that deal with problems in the utilization of space and its effects on human conditions.2 The idea of the conference came from the increasing congestion and general deterioration in the urban environment and its impact on the population, and from the recognition that improvement in the spatial accommodation of people involves extensive commitments of resources. A number of areas of common interest to the participants were identified at the conference, but it made no recommendation as to action. The range of interests considered at the conference spanned two clearly distinguishable aspects of space. Viewed at the micro level, the organization of space and the physical nature of structures have real but not very clearly defined impacts on the psychology of individual users and on the social relations between individuals and groups. However, this view of space is inadequate for dealing with very large aggregations of people and their organization into functioning cities, metropolises, regions, and nations. At this level a macro view of space is the concern mainly of economists, city planners, geographers, and urban sociologists. Since human activities are initiated at the micro level, there are many interactions between these two levels of the viewing of space, but it is recognized that knowledge is lacking and technical languages are inadequate for properly defining these interactions. • Corresponding to the diversity of disciplines involved W-



Cf. Items, March 1966, p. 9. VOLUME




and the variation in levels and foci of interest in space, there is a substantial amount of research in progress. The participants in the conference encountered mini, . mal difficulty in communication despite the diversity ~ of their interests. They recognized that, with the expansion of research and the increasing pressure of population on the environment (both natural and manmade), questions concerning a systematic view of the use of space will become increasingly important. It was

the consensus of both the conference and the committee that no systematic effort to bring these diverse tendencies together is appropriate at the present time, but that all workers in social science fields should maintain a sensitive awareness not only of the existence of problems of space utilization and its impact on people, but also of the substantial difficulties facing research in this field and the ultimate necessity of the active involvement of many disciplines.-

NEW COUNCIL COMMITTEE ON STATISTICAL TRAINING* by Conrad Taeuber, Frederick Mosteller, and Paul Webbink


AT THE request of the American Statistical Association, the Council held a conference in Washington, D.C. on May 19, 1967 to explore problems of statistical training, especially of statisticians to serve local, state, and federal government. 1 The conference was proposed because the growing lleed for statistical data has not been matched by a corresponding increase of personnel trained in data collection and processing, preparation of descriptive summaries and organization of the facts used in much social research and in municipal, state, and national policy-making. As a result of the conference the Council has appointed a new committee to be concerned with means of overcoming this imbalance. The members of the committee are: Conrad Taeuber (chairman), Carl L. Erhardt, John M. Firestone, Anders S. Lunde, John Neter, and Harry C. Trelogan. At the opening of the conference Frederick Mosteller, as Chairman, called attention to the increasing number of persons doing statistical work in federal, state, and local agencies and in the business world, and to the growing concern with the training which such persons need. The demand for persons with training in both a subject-matter field and in statistics has been expanding rapidly. The increasing use of computers has led to a need for persons competent in computer technology and in statistics. The importance of these prob• Reprinted with minor changes from The American Statistician, December 1967, by permission of its editor. 1 The participants in the conference were: Herbert L. Bryan, New York State Department of Correction; Carl L. Erhardt, New York City Health Services Administration; John M. Firestone, City Col· lege, New York; Leon Greenberg, U.S. Bureau of Labor Statistics; Robert E. Johnson. Department of the Air Force; Roye L. Lowry, Bu· reau of the Budget; Anders S. Lunde, National Center for Health Statistics; Philip J. McCarthy, Cornell University; Robert H . Moats, Illinois Cooperative Crop Reporting Service; Frederick Mosteller, Harvard University; John Neter, University of Minnesota; Donald C. Riley, American Statistical Association; Joseph Steinberg. Social Security Administration; Conrad Taeuber. Bureau of the Census; Harry C. Trelogan, U.S. Department of Agriculture; Paul Webbink, Social Science Research Council; and H. Bradley Wells. University of North Carolina. DECEMBER


lems is well recognized; the task of the conference was to consider ways of dealing with them. To initiate discussion John M. Firestone prepared a position paper, "The First Responsibility of a Statisti· cian." This included extensive quotations from eminent statisticians on the importance of statistics for such subjects as population, labor, agriculture, health, and education, and horror stories of the miseries that can occur when the basic properties of statistical indices are not appreciated. The author suggested: (1) a uniform first course in statistics, (2) emphasis on the continuum of training between theoretical and practical and on the importance of scientific method, and (3) the value of good reporting. Changes in employment opportunities since the depression days and changes in technology were said to have reduced the competence of those statisticians who come most directly in touch with the primary data. Carl L. Erhardt, in a report on the Conference on Health Statistics Professional Training (the "Bourbon Street Conference") held at Tulane University in April, noted its recommendation of a 10-fold increase by 1970 in support for departments of statistics in Schools of Public Health (for faculty and for equipment and space). The needs and problems of operating health agencies, especially state and local agencies, which were emphasized at this earlier conference, parallel those in the social fields. Participants in the Council's conference brought out the need for a variety of training courses, for: (1) reo search mathematical statisticians (who require more experience with applications), (2) research analysts with modest statistical training but strong subject-matter training, the two to be closely related, (3) supervisors of enumerators, who must understand the training requirements of their subordinates and the quality control of data collection in the field, (4) statistical clerks, (5) managers and policy makers. A major point was that the field of collection or


acquisition of data is a speciality that urgently calls for development. Statistical training has tended to stress advanced analytical tools, applied mathematics, systems analysis, and computer technology. Operating agencies have a limited need for personnel with such training, and greater need for personnel able to translate a research or operating question into a procedure for collection, tabulation, analysis, and reporting of the relevant data. With the emphasis on advanced analytical methods, too little attention is given to the nature of the data, sources of error, and the limitations of the data consequent upon the conditions in which they are collected and processed. There is need also for personnel competent in the social and health sciences as well as in statistics. Small offices, which cannot afford a high degree of specialization, must resort to consulting services to obtain more specialized assistance. If this is to be effective, the regular staff must be sufficiently competent to utilize the specialized services. Recruitment might be facilitated through better means of informing students early about the possibilities of careers in statistics, perhaps beginning in high schools to create an awareness of the role that statistics plays in the modem world and the importance of the personnel that produces and disseminates statistical information. An early involvement in statistical work is a valuable adjunct to much subject-matter training. It .is essential in this type of training that the student be exposed to "live" data and to an appreciation of the techniques that are applicable in the real world of uncertainty. The agencies that require personnel could assist teachers of statistics by providing brief descriptive statements of how surveys are generated and actually carried out. Conditions under which data are collected, tabulated, and analyzed should be described in realistic and understandable terms. The accounts should include illustrations of how results of analyses are presented for use in administration, and how they are actually used. Re-evaluation of all aspects of statistical work in many agencies is needed, to reassign much that can be done by clerical personnel or by computers. In this way jobs can be made more attractive to persons with some statistical training. It is not realistic to expect the present generation of college graduates to take jobs that are essentially clerical. In-service training has been developed in many agencies as a means for developing needed staff. In some instances students are recruited at the end of the sophomore year, given summer jobs and projects to do on their home campuses during the school year, and con-


tact with the agency during academic vacations. Training programs to acquaint new recruits with the whole range of activities of the agency and to inculcate the techniques that are most frequently applied there may take as much as a year, and include specific assignments WI in various aspects of the work. Combinations of on-thejob experience and academic work are now generally available in federal agencies. The National Center for Health Statistics is establishing ad hoc training programs to assist workers in state agencies. Summer employment and training as a means of recruitment have had mixed responses, depending in part on the ability of the employing agency to provide supervision and stimulation for the short-term assignments which students can carry out. An essential ingredient of training is experience in data collection and firsthand exposure to the conditions under which raw data are collected and to the sources of error in collection and processing. A period of service in a statistical agency is often a valuable prelude to advanced work in statistics. Further development of working relations between the universities and operating statistical agencies would be of value. The experience of a number of statistical laboratories and survey research agencies has demonstrated the utility of such collaboration. The conference recommended that the Social Science Research Council appoint a committee to deal primarily with the issues in training that had been dis- _ cussed. It was suggested that this committee might identify excellent programs, encourage their adoption or proliferation, and disseminate information about them; consider improvement of curricula and develop them if necessary; assess the need for centers for special training in selected areas, and determine where such centers might be especially helpful to state statistical and other agency programs; set up and administer demonstration programs as a means of appraising the value of proposed programs. ASA could provide task forces to assist the SSRC committee. The following lines of attack were mentioned as of possible interest to the committee: 1. What training will be needed, say, 10 years from now for work with the kinds of statistics that must be developed? It is necessary to think not only of the governmental series now available, but also of the social indicator series of the near future, of data banks and archives, and of the larger, more sensitive social inquiries-observational and experimental-that are going to be required. These new inquiries will have to be more extensive because the kinds of social investigation that are in increasing demand contain a great deal of "noise" and only large numbers can filter it. They will be required in spite of opposition to possible invasion of








privacy because the demands of economy will necessitate some sort of cost-benefit analysis for the whole society, although not, let us hope, one based on some single variable such as GNP. The development of such an analysis for, say, the field of education alone promises to require a very substantial effort by excellent personnel over the best part of a decade. 2. Regular collegiate centers. Through universities, colleges, or junior colleges, programs and curricula might be developed to prepare professional personnel for positions in local, state, and federal statistics. A review of current programs and the successes attained in parallel efforts in other areas might indicate the programs that are especially useful, and whether modest changes in programs in certain institutions would meet needs. Although traditionally such programs are viewed

as involving full-time work on the part of the student, the incorporation of summer or even longer programs to give the student some on-the-job experience and bring him back to his studies with more meaningful career objectives was regarded as desirable. Special attention should be given to this work-experience aspect because the degree of satisfaction with it has varied enormously from one training program to another, both on the part of the student and of the organization concerned. 3. SPecial short-course centers. Arrangements might be made for special training centers where personnel already employed full time in state or local agencies might take short courses to bring their training up to date, and indeed ultimately to become expert in a whole area such as data collection.

COMMITTEE BRIEfS KOREAN STUDIES (Joint with American Council of Learned Societies) Edward W. Wagner (chairman), George M. Beckmann, Gari K. Ledyard, Chong-Sik Lee, Glenn D. Paige; staff, Bryce Wood This new joint committee was appointed in June to assist in the development of Korean studies in the United States through a program of conferences and related activities, for which $65,000 has been made available to the Councils by the Ford Foundation for a three-year period. The Foundation has also made grants to five universities for the improvement of their resources for research and training in this field, and it is expected that the committee will be able to facilitate communication and cooperation in the development of their programs. The committee held a first conference at the University of Washington on November 10-11, to consider problems of library resources for research on Korea, especially problems of acquisition of materials; cataloging, classification, and utilization of collections; and possibilities for interuniversity cooperation in dealing with these problems. In addition to the members of the joint committee and staff, the participants in the conference were: Jaehyun Byon, Byong-ik Koh, Ruth Krader, Fred Lukoff, S. E. Solberg, Doo Soo Suh, and George E. Taylor, University of Washington; William E. Henthorn, Princeton University; John C. Jamieson, University of California, Berkeley; Joobong Kim, Columbia University; Kyu S. Kim, C. W. Post College; San-Oak Kim, University of California, Los Angeles; Sungha Kim, Harvard University; Hesung C. Koh, Yale University; Joyce Wright, East West Center Library, University of Hawaii; and Key P. Yang, Library of Congress. The conference resulted in a number of suggestions for fostering cooperative relationships among libraries. At a DECEMBER


meeting held during the conference the joint committee initiated plans for improving Korean language instruction and for development of closer relations with scholars at Seoul National University, Korea University, and other institutions in Korea. SOCIOLINGUISTICS Charles A. Ferguson (chairman), Susan M. Ervin-Tripp, Joshua A. Fishman, John J. Gumperz, Everett C. Hughes, Dell Hymes, Stanley Lieberson; staff, Elbridge Sibley A meeting of the committee on October 29-30 was devoted mainly to review of current lines of research in its fields of interest and evaluation of developments since its appointment in 1963. Three guests were present: William Labov of Columbia University, who reported on aspects of his study of urban dialects; Wayne A. O'Neil of Harvard University, who discussed his earlier work on IcelandicFaroese intelligibility and some of the current "school language" research at the Harvard Graduate School of Education; and Allen D. Grimshaw of Indiana University, whose general review article "Sociolinguistics," to appear in Nathan Maccoby (ed.), Handbook of Communications Research, was discussed along with comparable articles by John J. Gumperz, "Language and Communication," The Annals, September 1967, and Susan M. Ervin-Tripp, "Sociolinguistics," to appear in Leonard Berkowitz (ed.), Advances in Experimental Social Psycholo~ IV. Most of the discussion centered on "small group" kinds of sociolinguistic research which have been productive and influential in recent years. In much of this work investigators have used the speech event (e.g., leave-taking, narrative) as the unit of analysis, but with the kind of sophistication usually applied to the linguistic analysis of sentences and the writing of descriptive grammars. On the one hand this new research presents a challenge to more traditional linguistics in attempting to include nonlinguistic social 51

variables in the linguist's framework of rule-governed behavior, while on the other hand it yields data and interpretations of interest to social scientists but with much of the linguist's terminology and research techniques. From this kind of research new concepts of wide applicability have emerged, such as the notions of linguistic indicators, mm'kers, and stereotypes developed in studies of social stratification, and the more recent classification (by Labov) of sociolinguistic rules as "categorical" (exceptionless, outof-awareness regularities), "semicategorical" (the infrequent exceptions are socially significant and reportable), and "variable" (relative frequencies identify social variables). Another new development is the formulation (by Ervin-Tripp) of "flow-chart" rules for personal address behavior. Closely related research in which the ve1'bal 1'epertoire (e.g., range of style levels in a monolingual community, different languages in a multilingual community) is first used as the unit of analysis has challenged the traditional linguistic preference for structurally homogeneous data by showing important regularities of language uses which may cut across dialect and language boundaries. Both this and the "small group" kinds of research have largely by-

passed the issue of correlations between social structure and linguistic structure and have attacked new problems with new conceptual tools. In spite of the time devoted to these "micro" approaches the committee reasserted its concern for specific areas of "macro" research, especially in reference to its projected conference on social conditions conducive to the establishment and maintenance of creole languages and its ongoing study of language problems of developing countries. The committee noted with satisfaction that a recent Ford Foundation grant to the University of Hawaii will enable four scholars who participated in its 1966 conference on the latter subject to spend the academic year 1968-69 at the East West Center in intensive collaborative research. Although the committee agreed that sociolinguistic research is much more widely recognized as a legitimate and productive field than it was in 1963, it also recognized that research and training in the field have not yet attained the degree of institutionalization needed and that major sub fields of research are hardly touched. It decided therefore to continue for at least two years its sponsorship of research projects, symposia, and other activities, at the same or an accelerated pace. C. A. F.


NEW PUBLICA rlONS Agricultural Development and Economic G1'Owth, edited by Herman M. Southworth and Bruce F. Johnston. Sponsored by the former Committee on Agricultural Economics. Ithaca: Cornell University Press, September 1967. 623 pages. $12.00. The City m Modem Af1'ica, edited by Horace Miner. Product of the conference on methods and objectives of research on urbanization in Africa, sponsored by the Joint Committee on African Studies, April 1-3, 1965. New York: Frederick A. Praeger, December 1967. c. 384 pages. $7.50. Contemporary China: A Resea1'ch Guide, by Peter Berton and Eugene Wu. Hoover Institution Bibliographical Series: xxxi. Prepared for the Joint Committee on Contemporary China. Stanford University: Hoover Institution on War, Revolution, and Peace, September 1967. 724 pages. $22.50.

Ea1'ly Education: CU1'rent Theory, Resea1'ch, and Practice, edited by Robert D. Hess and Roberta M. Bear. Papers prepared for the Conference on Preschool Education, sponsored by the Committee on Learning and the Educational Process, February 7-9, 1966. Chicago: Aldine Publishing Company, December 1967. c. 272 pages. $6.95. Genetic Diversity and Human Behavi01', edited by J. N. Spuhler. Viking Fund Publications in Anthropology, No, 45. Proceedings of a symposium, September 17-25, 1964, jointly sponsored by the Wenner-Gren Foundation for Anthropological Research and the Committee on Biological Bases of Social Behavior. Chicago: Aldine Publishing Company, November 1967. c. 320 pages. $7.50. Social Science in Latin Ame1'ica: Pape1's Presented at the Conference on Latin American Studies Held at Rio de Janeiro, March 29-31, 1965, edited by Manuel Diegues Junior and Bryce Wood. New York: Columbia University Press, September 1967. 348 pages. $4.50.










Incorporated in the State of Illinois, December 27, 192-1, for the purpose of advancing research in the social sciences Directors, 1967:












Officers and StafJ: PENDLETON HERRING, President; PAUL WEBBINK, HENRY W. RIECKEN, Vice-Presidents; ELBRIDGE SIBLEY, BRYCE WOOD, Executive Associates; ELEANOR C. ISBELL, ROWLAND L. MITCHELL, JR., NORMAN W. STORER, Staff Associates; JERO~IE E, SINGER, Consultant; CATHERINE V. RONNAN, Financial Secretary .



Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.