Page images

The remainder of this paper is concerned with the development of a strategy for improving each of these types of indicators. At the outset, it is necessary to differentiate two basic areas of social indicator development: the preparation of social indicator reports and the conduct of social indicator research and development. These areas are the subject of the following two sections.

Social Indicator Reports The presentation of descriptive statistics in order to inform different users of current conditions and trends is of course the basic reason why such data are collected in the first place. The most comprehensive single example is the Statistical Abstract of the United States, whose centennial edition will appear in 1979. These publications, together with similar compendia dealing with particular subject-areas, usually consist of printed tabulations without interpretive text. Their basic use is as reference documents for a variety of users. Other Federal agencies issue annual reports containing a broad range of descriptive statistics together with some interpretive text. Examples are the Annual Statistical Supplement to the Social Security Bulletin, issued by the Social Security Administration, DHEW (since 1955); the Employment and Training Report of the President, issued by the Department of Labor since 1963; The Condition of Education issued by the National Center for Education Statistics; Health, United States, 19761977 issued jointly by the National Center for Health Statistics and the National Center for Health Services Research.

Over the years several efforts have been made to develop comprehensive reports which approach the idea of social indicator reports by including assessments of current conditions and trends and recommendations for needed policies and actions. By far the most comprehensive of these efforts was a 4year research project commissioned in 1929 by President Hoover and carried out under the direction of William F. Ogburn. This project culminated in the publication of Recent Social Trends in the United States (1933). The separate volumes of this report provided detailed examinations of some 30 areas of American life and were a useful source of factual information in guiding the development of policies and programs during the depression years.

Nothing approaching this magnitude has been undertaken since then, but a number of more modest reporting efforts reflect intermittent concern with the

potential usefulness of social reports which seek to cover a wide range of topics. Beginning in 1961, for example, the U.S. Department of Health, Education, and Welfare issued a monthly report entitled Indicators, supplemented by an annual compilation of selected time-series, entitled Trends. Both reports presented selected statistics relating to program developments in the areas of health, education, and welfare together with selected descriptive statistics reflecting the condition of the population in these broad areas of concern. Brief articles treating selected topics were also included.

The resurgence of interest in societal assessment which accompanied the “great society” programs of the mid-1960's gave rise to the establishment, in 1966, of an advisory group of social scientists to advise the Secretary of DHEW on the measurement of social change and the preparation of a social report. The only published result of this large-scale effort was the appearance, in 1969, of DHEW's Toward a Social Report.' Despite its brevity (101 pages), this report is an important prototype of providing assessments of current status in seven major areas of concernhealth and illness; social mobility; the physical environment; income and poverty; public order and safety; learning, science, and art; and participation and alienation.

An important supplementary undertaking was begun in July 1969, when President Nixon established a National Goals Research Staff within the White House, under the direction of Leonard Garment. The work of this group was terminated with the issuance of a report in July, 1970, entitled Toward Balanced Growth: Quantity with Quality.* Taken together, these two reports illustrate both the potential and the limitations of social reports. Toward a Social Report was intended to serve as a prototype for annual assessments of "social well-being” as a basis for making "informed decisions about priorities and directions in this Nation's social programs." The analysis it contained was deliberately normative and the criteria employed in selecting the few social indicators it contained emphasized that these

'U.S. Department of Health, Education, and Welfare, Toward a Social Report (Washington, D.C.: U.S. Government Printing Office, 1969). The original intention of this effort was to inaugurate an annual social report. See the covering letter from Secretary Wilbur J. Cohen to the President, dated January 11, 1969. The principal author of "Toward a Social Report" was Mancur Olson.

*National Goals Research Staff, Toward Balanced Growth: Quantity with Quality (Washington, D.C.: U.S. Government Printing Office, 1970). President Nixon's announcement of the establishment of this research staff on July 13, 1969 also made reference to the preparation of “a public report, to be delivered by July 4 of next year, and annually thereafter....” (italics added). The principal author of this report was Raymond A. Bauer.

?President's Research Committee on Social Trends, Recent Social Trends in the United States (New York: McGraw-Hill, 1933).


indicators should have normative significance. The statistics and who would be expected to develop their notion of “normative significance” is controversial. own interpretations in any case. This approach was Mancur Olson's statement expresses the intended patterned after the long-standing practice of meaning when he argues that it should be possible to presenting economic indicators without interpretive infer that "if (a given indicator) changes in the text. “right" direction, while other things remain equal,

Social Indicators 1973 was issued in February, things have gotten better, or people are “better off." 1974-after more than four years of development. Toward Balanced Growth had a different focus. It was

The reception accorded this first publication offers designed to explore possible alternatives for future

few clear-cut lessons as yet, but a number of tentative development by proposing a number of "debates of

conclusions can be offered. About 5,000 copies were emerging issues" for public consideration. Thus both

distributed among the legislative and executive reports shared a normative orientation and both of

branches of the Federal Government and selected them made use of selected descriptive statistics in

governmental agencies at the State and local level. their analysis and interpretation.

Another 14,000 copies were purchased by libraries The limitations of both reports were well known

and private individuals. From the limited feedback, it and expressed by their authors. In preparing Toward

is possible to distinguish four broad audiences for rea Social Report, it soon became evident that the bulk

ports of this kind: governmental policymakers and of the statistics available for any given subject were

their staffs, librarians and other reference sources, designed for administrative purposes, not for societal

private researchers and research organizations, and assessment. Data relating to the well-being of

interested members of the general public. individuals or population groups or to the results or

Interviews with a small sample of high-level outcomes of particular policies or actions were, and

government officials, conducted by Nathan Caplan remain, extremely hard to come by. Those who

and associates of the University of Michigan, prepared Toward Balanced Growth encountered an

indicated that while Social Indicators 1973 was well equally serious obstacle: the paucity of normative

received, it was generally regarded to have promised projections tracing alternative paths toward the

far more than it could deliver. It was felt that the attainment of specified goals. Although the staff

bulk of the data presented did not satisfy the criteria carefully avoided prescribing goals, it could not

initially set forth for social indicators-that the escape the need to offer alternative goals for public

measures selected should reflect the well-being of consideration, in the form of issues raised in its

individuals or families and that they should relate to "emerging debates."

outcomes or results rather than inputs of resources. Shortly after the publication of Toward a Social Furthermore, the high level of aggregation required Report, work was begun within the Office of to cover a range of subjects in a single publication Management and Budget on the first comprehensive

meant that the information supplied was seldom national social indicators report, Social Indicators

directly relevant to any particular set of policy 1973. This renewed effort was motivated largely by

deliberations. Finally, descriptive statistics, whether the belief that the careful process of data selection

they relate to “inputs” or “outputs,” cannot inform which characterized Toward a Social Report could

the policymaker as to whether a particular program readily be extended to provide more adequate

has brought about its intended outcome or whether coverage of the major areas of concern. Whereas the

its benefits outweigh its costs. Thus the report was preceding reports consisted primarily of interpretive

generally regarded as an exceptionally attractive essays, the social indicator report was designed to reference document, to be used as a handy source of provide a statistical portrait of the society, stressing

background information in speech writing—a service the graphical presentation of descriptive statistics

which many other reference works, such as the without analytic or interpretive text. This focus was

Statistical Abstract, can also provide. justified, in part, by the notion that statistics, if

"Nathan Caplan and Eugenia Barton, “Social Indicators 1973: A properly selected and presented, “speak for

Study of the Relationship between the Power of Information and themselves." It was further justified by the fact that

Utilization by Federal Executives," published by the Center for the primary audience for which the report was Research on Utilization of Scientific Knowledge, Institute for designed consisted of individuals in policymaking or

Social Research, the University of Michigan, September 1976. For policy-influencing positions within the Federal

insight into the informational needs of policymakers with respect

to social issues, see “Social Research and Development of Limited Government-persons who would be relatively

Use to National Policymakers,” a report of the Comptroller experienced in gleaning useful intelligence from raw General of the United States to the Congress, April 4, 1977. Also

useful is Martin Rein, Social Science and Public Policy

(Harmondsworth, Middlesex, England; Penguin Books, Ltd., "Toward a Social Repori, p. 97.


The response of academic researchers and other specialists is also fairly well documented, thanks to the publication of a review symposium organized by the Social Science Research Council shortly after the appearance of Social Indicators 1973. Besides the numerous technical criticisms that were offered by these reviewers, the major general criticism was the absence of analysis and interpretive text. Most of these specialists recognized that the report was not primarily intended to meet their own more stringent data requirements, but they were nevertheless concerned with both the data selection process (what was included, what was left out, and why) and with the paucity of analysis-particularly in regard to the quality of the data, possible misinterpretations, and the like. Whereas both the policymakers and the technical specialists were commonly critical of the absence of interpretive essays, the latter group tended to be especially concerned with the absence of adequate warnings as to the limitations of the data presented.

Much less information is available as to the reactions of individual private purchasers of the report, but occasional comments from librarians indicate that the social indicator report has received extensive use both for reference purposes and in connection with the preparation of students' themes, term papers, and the like. The graphic presentations in particular have won almost universal praise.

Work was begun on the preparation of a second social indicator report, Social Indicators 1976 in July, 1974. The general format of this report is similar to that of its predecessor. It again features the graphic presentation, in color, of summary descriptive data on the socioeconomic characteristics of the population, with very limited geographic detail but with considerable disaggregation by age, sex, color, and other background variables. The coverage of the second report has been enlarged considerably, from 8 to 11 social indicator chapters plus an introduction to the report as a whole. In addition, each chapter contains a section of opinion data relating to public perceptions and a section providing a few international comparisons. Summary statistics relating to the socioeconomic characteristics of a number of ethnic groups, drawn of necessity from the past three decennial censuses, are also included in the introductory chapter, together with a brief discussion of the nature of sampling error and sources of nonsampling error which may affect the quality of the data presented.

The report does not provide interpretive text or detailed analysis of the data shown; however, a special issue (Volume 435, January 1978) of The Annals (of the American Academy of Political and Social Science) has been issued, containing a number of interpretive essays based on the chapters of the report. This publication also contains selected statistical material presented in Social Indicators 1976 and therefore provides a variety of interpretations which could not be included in the report itself. In the spring of 1976, work was also begun at the Bureau of the Census on the preparation of a monthly chartbook, STATUS. Four issues of STATUS were issued to a number of individuals both within and outside of the Federal Government, starting with a July 1976 issue and ending with an October issue. STATUS was originally an outgrowth of the weekly briefing notes which were prepared for the White House beginning in the spring of 1975. These weekly reports contained selected current statistics, in graphic form, relating to a wide variety of domestic conditions and developments. STATUS was designed as a public version of these notes.

Despite the fact that STATUS was discontinued for lack of funding, it may still be regarded as an alternative mode for the presentation of social indicators to a wider audience. Although no single issue of STATUS could possibly offer the comprehensive coverage of different subjects that is provided in a general social indicator report, its cumulative content, over a 12-month cycle of publication, could include considerably more detailed information and could, in addition, retain greater flexibility in providing coverage of special topics or particular population groups which would cut across several subject areas. It would also permit publication of data from recently released topical reports on a timely basis.

In summary, Social Indicators 1976, designed as a biennial publication, and STATUS, originally intended to be a monthly chartbook, are both representative of the current state of the art with respect to the graphic presentation of descriptive statistics. Neither publication has enjoyed the resources or the lead time to permit extensive analysis or retabulation of available data, but both of them are appropriate vehicles for the presentation of comprehensive statistical data relating to a variety of subjects and designed to meet the interests of nonspecialists.

Periodical publications, such as STATUS, enjoy two major advantages over annual or biennial re

Roxann A. Van Dusen (ed.), Social Indicators 1973: A Review Symposium (Washington, D.C.: Social Science Research Council, Center for Coordination of Research on Social Indicators, 1974).

'The special editor for this issue of The Annals is Dr. Conrad Taeuber of Georgetown University.

ports of a more comprehensive nature. First, they can provide much more current data and second, they can provide a quick and flexible means for presenting materials of current interest on an ad hoc basis. For example, a periodical report can capitalize on the availability of occasional studies relating to special population groups or to other subjects which cut across a number of areas of social concern. Their disadvantages include their inability to provide comprehensive treatment of several subjects in any single issue, their limitations with respect to the depth of treatment that can be given to any topic, and their limited ability to operate with a formal organizational framework.

and trends. These facts, in turn, provide essential guidelines in formulating improved questions for further research. A strategy for social indicator research and development, aimed at generating measures that would provide more adequate descriptions and greater analytic power, is discussed in the following section.

These two types of social indicator reports—the monthly chartbook and the annual or biennial social indicator report-are not the only vehicles for communicating selected statistical data to the general public. Several major statistical agencies have begun publication of comprehensive annual reports covering the subject matter under their jurisdiction in a format similar to that of the OMB report. Two examples may be mentioned here: the National Center for Education Statistics (DHEW) produces an annual report entitled The Condition of Education and the National Center for Health Statistics (DHEW) with its first comprehensive report, entitled Health, United States, 1975.' Because these types of reports focus on a particular subject area, they provide more detailed coverage than is possible in any report covering several areas.

All of the above reporting activities may be regarded as experimental efforts at public communication. They share a common reliance upon descriptive statistics selected from sample surveys and administrative records. They also share a common limitation with respect to the depth of analysis and interpretation that can be developed on the basis of such data. Continuation of these reporting efforts may be expected to yield improvements in the selection and organization of materials, in modes of presentation, and perhaps in responsiveness to emerging public issues and interests. Such efforts cannot be expected to generate deeper levels of understanding of the factors underlying the developments which are observed. Even less can they be expected to provide clear-cut evidence of the effectiveness of particular programs or policies.

In short, descriptive statistics provide useful factual information concerning current conditions

Social Indicator Research and Development

With respect to descriptive indicators, two main lines of research and development appear to hold promise. First, the improvement of the data base from which indicators are derived, both in terms of more comprehensive coverage of different areas of concern and in terms of greater comparability of concepts and classifications, would in time yield more adequate indicators of the status and conditions of different population groups in the society. The various subject matter chapters of this Framework set forth a number of proposed improvements in the general structure of sound statistics. Second, the ongoing efforts of the Social Indicator Development Program of the Organization for Economic Cooperation and Development (OECD) are aimed at producing a basic core of social indicators which would reflect conditions and trends with respect to each of the major areas of social concern which have been agreed upon by member countries.

The achievement of greater comparability of concepts and classifications is requisite to developing tabulations and analyses which utilize data from different sources or covering different subjects. Some of the major surveys whereby statistical information on different aspects of our social condition is collected include the Current Population Survey (CPS); the Annual Housing Survey (AHS); the Survey of Income and Education (SIE); the Survey of Income and Program Participation (SIPP); the National Crime Victimization Survey (NCS); the Consumer Expenditure Survey (CES); the Health Interview Survey (HIS); the Health and Nutrition Examination Survey (HANES); and the National Assessment of Educational Progress (NAEP). Among these, the current population survey is unique in serving as a general-purpose data collection mechanism which is employed in obtaining a wide variety of information on the social, economic, and demographic characteristics of the population. It is possible, at least in principle, to increase the analytic usefulness of the data collected in these several major surveys by achieving agreement on a common set of concepts, definitions, classifications and procedures to be employed in all of them. A number of considerations militate against such agreement, except perhaps for a very limited subset of variables. To begin with, the

'U.S. Department of Health, Education, and Welfare, National Center for Education Statistics, The Condition of Education, 1976 edition and National Center for Health Statistics, Health, United States, 1975 and 1976/77 editions: U.S. Government Printing Office).

several areas of social concern do not share with systems of national economic accounts the property of being represented by transactions expressed in terms of common monetary values.

On the contrary, the variables of social interest necessarily relate to different units of observation. When these units are people, they may consist of individuals, households, families, work teams, or other formal or informal associations. When they are geographic locations, they may consist of housing complexes, neighborhoods, school districts, towns, cities, or larger administrative units or geographic areas. In still other cases, the units of observation may consist of events or transactions, such as traffic accidents, crimes, or births, deaths, and other vital events. This diversity does not imply that common concepts and classifications are unattainable, but it does suggest that any commonality that is achieved will be both complex and of limited coverage and applicability.

A second limiting factor is the extreme variation in the amount of detailed information required to meet the purposes of different social surveys. Except for a few items of information collected from all persons in a decennial census, all survey data (and many administrative statistics as well) are collected on a sample basis. The size of the sample to be drawn in a particular survey depends on a number of factors, including the heterogeneity of the population to be surveyed, the frequency of observations called for, whether successive observations should be obtained independently, the level of disaggregation required to provide the required information, and, most critically, the level of reliability needed for information relating to the smallest groups or cell frequencies for which separate estimates are desired. At one extreme, a nationally representative sample of perhaps 1,500 persons may suffice to yield acceptably precise


"For an appreciation of the technical difficulties to be overcome in establishing a workable data system of the kind envisioned, see Edgar S. Dunn, Jr., Social Information Processing and Statistical Systems: Change and Reform (New York: John Wiley & Sons, Inc., 1974). Also informative in this regard are articles by H.H. Winsborough, “Age, Period, Cohort, and Education Effects on Earnings by Race-an Experiment with a Sequence of CrossSectional Surveys,” in Kenneth C. Land and Seymour Spilerman (eds.), Social Indicator Models (New York: Russell Sage Foundation, 1975); “Connections within the System of Social and Demographic Statistics: An Introductory Discussion of their Thematic and Methodological Basis," a discussion paper prepared for the Working Party on a System of Social and Demographic Statistics of the Conference of European Statisticians, the United Nations Statistical Commission and the Economic Commission for Europe (May 6, 1976); and Kenneth C. Land and Marilyn M. McMillen, “The Demographic Approach to Social Indicators," Working Papers in Applied Social Statistics (WP7807), Social Science Quantitative Laboratory, University of Illinois at UrbanaChampaign, March 1978.

estimates of certain population characteristics for the country as a whole. At the other extreme, a 20percent sample, comprising some 15 million households, may be required to provide acceptably precise estimates of unusual combinations of characteristics for small geographic areas or for special population groups. Here also, no single conceptual framework can hope to satisfy such diverse information needs without considerable sacrifice of efficiency or loss of information.

Finally, there remains the problem of the different concepts, classifications and procedures which may be employed in collecting information on the same subjects for different purposes. Some of these differences may readily be resolved by proper planning and coordination. But in many instances, these differences reflect alternative perspectives and research interests and are therefore less likely to prove amenable to common treatment. This final limitation applies as well to the emergence, over time, of unanticipated problems, new theories and perspectives, policy innovations, and the like. No set of concepts and classifications can hope to anticipate future needs, insights, and approaches.

These considerations suggest the need for a shortterm strategy, designed to make optimal use of available statistics in meeting the continuing need for comprehensive information on social conditions and trends, and for a long-term strategy, designed to improve the data base from which such information may be derived in the future. The development of the Framework is one step in this direction, but much detailed planning, research and implementation is required.

In the short term, further efforts should be made to establish comparable concepts and definitions of basic background variables and common characteristics of the units of observation so that data linkages at intermediate levels of aggregation could be made. For example, several of the major surveys now conducted collect data relating to individual members of particular households. Some of these data can usefully be interpreted as a characteristic of the household in question. Common definitions of certain of these characteristics would permit comparisons of data from the CPS (now covering over 60,000 households each month), the Consumer Expenditure Survey (20,000 families), the Annual Housing Survey (60,000 housing units in the national sample), and the Health Interview Survey (42,000 households or 140,000 persons). Data linkages at intermediate levels of aggregation are of course beset by familiar problems of aggregation bias. Comparisons of the characteristics of specified

« PreviousContinue »