Richard C. Rockwell

Assessment of Multicollinearity

Interdependence among explanatory variables is a common condition for sociological analyses. It may markedly affect the stability of estimates of parameters obtained from least-squares regression. Multicollinearity is viewed as a problem which poses two questions for the analyst: how severe is the multicollinearity and what is its effect on the analysis? The determinant of the correlation matrix of explanatory variables is a measure of the severity of multicollinearity. Haitovskys chi-square statistic permits the assessment of the null hypothesis that the correlation matrix is singular. This paper demonstrates the need for this test through an examination of published correlation matrices. It is suggested that use of the Haitovsky test be routine in any analysis which attempts the estimation of parameters through regression analysis.

From Preserving the Past to Preserving the Future: The Data-PASS Project and the Challenges of Preserving Digital Social Science Data

Social science data are an unusual part of the past, present, and future of digital preservation. They are both an unqualified success, due to long-lived and sustainable archival organizations, and in need of further development because not all digital content is being preserved. This article is about the Data Preservation Alliance for the Social Sciences (Data-PASS), a project supported by the National Digital Information Infrastructure and Preservation Program (NDIIPP), which is a partnership of five major U.S. social science data archives. Broadly speaking, Data-PASS has the goal of ensuring that at-risk social science data are identified, acquired, and preserved, and that we have a future-oriented organization that could collaborate on those preservation tasks for the future. Throughout the life of the Data-PASS project we have worked to identify digital materials that have never been systematically archived, and to appraise and acquire them. As the project has progressed, however, it has increasingly turned its attention from identifying and acquiring legacy and at-risk social science data to identifying ongoing and future research projects that will produce data. This article is about the projects history, with an emphasis of the issues that underlay the transition from looking backward to looking forward.

An Investigation of Imputation and Differential Quality of Data in the 1970 Census

Abstract This article examines the imputation of data in a one percent sample of the nations ever-married women aged 18–25 from the 1970 Public Use Samples of the Census of Population. Nonwhites are more likely to be imputed data than are whites. Rates of imputation of data are found to be extremely high for some categories of women, and white-nonwhite fertility differentials are affected by imputation. Consequences for study of fertility differentials are noted. It is recommended that researchers routinely use information on imputation of data in any analysis of 1970 Census data.

Social impacts of the HIV epidemic

Since there is no sociological literature on AIDS, this paper is an introduction to the concepts of culture, social structure, population and human ecology, with hypothetical applications of AIDS information to these ideas. As C. W. Mills set forth, basic questions of social reality applicable to AIDS research might include: what is the social structure of the society; where does the society stand in economic development; and how will the pandemic affect the varieties of people in the society? A culture is the set of ideas and material products that a society uses in interpreting their world. AIDS will renew attention to regulation of sexuality and research on sexual behavior. Social structures are patterns of relationships that give predictability to social life. Patterns of discrimination against minorities and drug users will be affected by AIDS. The dependency ratio of some populations, if not the size of the population itself, may be affected by AIDS. Human ecology is the study of how societies adapt to physical environments, organization and technology. How society uses human and monetary resources will be markedly affected by AIDS, considering the high social costs of education, condoms, treatment, death and burial. A list of positive changes that AIDS may evoke is given: new effectiveness of WHO, and of indigenous health personnel in developing countries, universal health insurance in the U.S., medical rather than criminal treatment of drug users, attention to poor minorities and elimination of racial and ethnic prejudices.

Data Archives: International

Archives of quantitative social science data began in the 1940s in the United States and have since spread to much of the world. Two major thrusts underlay this movement: (a) the need for transparency in the conduct of research and (b) the potential of data archives to facilitate comparative and time-series research. There is now a widely shared ethic of data sharing in the social sciences, which is codified in rules of a number of granting agencies. The holdings of data archives consist of microdata and macrodata, the codebooks describing those data, and various supporting materials such as finding aids. Archives have exercised care to protect the privacy and confidentiality of respondents in the surveys in their collections; today, laws often buttress those practices. Secure data enclaves are a means of providing access to highly sensitive data, but they offer less convenience to the researcher as a result of the conditions that they must impose. Archives dealt with major technological changes in information technology and telecommunications over the decade of the 1990s. Modern information technology has brought the services of data archives directly to the desktop of the researcher. Modern telecommunications have permitted the rapid dissemination of data around the world and the effective creation of global virtual data archives.

Social Statistics and Public Policy for the 1990s: [Introduction]

At the core of each of the disparate disciplines called the social sciences there is the relationship-sometimes, the tension-between the individual and society. In theoretical jargon, many substantive problems turn on the connections between macro institutions and processes and micro institutions and processes. Many micro-macro interactions come easily to mind, evoking different disciplines of the social sciences: society and its institutions, the state and its citizens, the law and the organizations that enforce it, the economy and the enterprises that produce its goods, gross national product and individual entrepreneurial traits, organizations and their members, neighborhoods and their residents, birth rates and individual attitudes toward having children, schools and their pupils, social behavior and individual personality, economic growth and individual productivity, and cognitive processes and brain chemistry. One disciplines macro analysis may be anothers micro analysis, and it is at such intersections that the disciplines often meet most productively. The articles in this special section are concerned to a greater or lesser degree with these macro-micro interactions. Together, they suggest the variety of topics in public policy for which researchers will analyze social statistics in the next decade and the variety of analytical methods that will be brought to bear. The article by Wong and Mason opens the section by elaborating on a generalization of the hierarchical linear model for multilevel analysis in comparative studies of different societies. Ethnic identity differentiates members of these societies in ways specific to each context. Their empirical example is of the dependence of fertility on the socioeconomic origins of women in less developed countries during the 1970s. This relationship is affected by the womens ethnic identities in ways specific to each country. Wong and Masons article provides an example of a trend within the social sciences toward quantitative comparative analysis. Far too much of American social science has been built upon the shaky foundation of understanding only this one society, as if the United States in the twentieth century encompassed the whole of human diversity. From studies of this one society (or even of tiny aspects of it, such as groups of college sophomores), grand theories have been developed-theories regularly presumed to apply to all human societies at all times. Comparative analysis permits us to transcend the limitations of that approach, but at a significant price: We must find a means for taking into account contextual differences among the societies we seek to compare. The seeming impossibility of doing so has led some to conclude that comparative analysis is inherently meaningless. Wong and Mason offer a powerful analytical tool for overcoming this objection. That tool is surely needed to tackle the problems outlined by Pearson in his article on the American urban underclass. What are to be compared here are not countries but cities, neighborhoods, races, and ethnicities. The public policy question is difficult: Is there any set of policies that would halt or reverse the separation of a small proportion of the urban poor from the social and economic life of the nation? The analytical questions are perhaps more difficult: How can the macro effects of economic change in the cities be disentangled from the effects of social welfare policies or of neighborhood institutions and organizations, and how can these effects be disentangled from those of social networks, local demographics, discrimination, and rates of drug use and violent crime? Corder and Manton raise the macro-micro problem in a different guise: The aged are an increasing proportion of American society, but federal surveys and censuses do not yet provide fully satisfactory tracking of their health and functioning. The relationships of the aged to the institutions of society-the health care system, housing, families, pensions-are points of economic stress and political debate. The October 1990 federal budget crisis illustrated once again the dilemmas facing a society that seeks to reconcile its contract to provide health care and a decent living income for the aged with the constraint of slowing economic growth and rising needs in other sectors of society, including the urban underclass and children. Finally, Schirm deals with the macro-micro relationship between the estimation of the distribution of the U.S. population and the representation of each state in the U.S. House of Representatives. He shows that adjustment for the differential undercount of particular groups of the population would have minor effects on the apportionment of House seats. In addition, he examines some of the difficult public policy issues in light of his calculations and conclusions, including the sensitivity of different adjustment methods to assumptions about particular groups and the place of adjustment in relation to statistical accuracy and political equity. The impetus for this special section came from the Social Statistics Section of the association, which long has sought a way to employ the pages of JASA to demonstrate how social statistics can be brought to bear on issues of public policy. I express my appreciation to Abbott L. Ferriss, Mary G. Powers, Robert Parke, and Kirk M. Wolter, who served on a preliminary review panel, and to the more than 30 anonymous referees, many of whom gave authors helpful advice toward strengthening their articles.