DATA BANKS AND DEPOSITORIES

CONCEPT AND HISTORY OF DATA-BANKS

The tradition of thought that gave rise to the social sciences was based on the quest to understand the laws or regularities governing the emerging industrial societies with democratic political regimes. The political and economic revolutions that were shaking the world substituted the relative predictability of the traditional ways of handling power and production with the disconcerting uncertainties of political consensus and the new market of commodities.

Finding such laws was not easy. The model for scientific inquiry established by the successful endeavors in the various fields of physical and natural sciences was not applicable to social science research. For a period social scientists believed that the complexity of the social object and the immaturity of the field were responsible for the failure of the natural science model. Social scientists gradually became aware, however, that the epistemological foundations of the social sciences were different: in practical terms, because creating an experimental situation in social matters is extremely difficult; and in theoretical terms, because society is a moving target, readily reacting to changes in circumstances. In the end, mainstream social scientists learned to live with these difficulties.

The Newtonian challenge of formulating hypotheses, collecting the relevant data, and accepting only those hypotheses that fit observational data, has been in one way or another the stimulus and the standard of advances in knowledge in the many fields that later composed the social sciences. The systematic collection of empirical observations has been the ballast that has kept modern social sciences from drowning in second-rate philosophy or outright ideology. Numerous social researchers and thinkers have stood up to the challenge of providing reliable observations on the social world. (Since every observer is also a member of society, it is not easy to stand aside and look at it from a fixed point of view.)

In retrospect it is understandable that different paths to the common goal of collecting systematic reliable data were tested with various methodological and technical tools, not always being understood as part of the same endeavor. In the latest bout of cultural dominance of Marxist theories, Karl Marx and his school were viewed as "grand theorists" squarely opposed to the "abstract empiricism" of contemporary sociology. But this was a misconception that completely overlooked the many years spent by Marx painstakingly collecting data on industrial society, and by desire to be a scientist like Charles Darwin—to whom he dedicated the book produced by his gigantic research effort, Das Kapital.

Despite the apparent confusion and turmoil, there was an underlying paradigm. Social sciences had to be empirical. No matter how radically critical and antipositivistic have been the epistemological conclusions of the various Methodenstreiten, the mainstream has resisted the idea of a data-free social science. Altogether such theoretical orientation has provided this disciplinary area with a fairly resilient ballast against ideological nebulosities. And this can be said safely without underestimating the humongous and repetitive production of irrelevant trivia in tabular or graphic form produced over the years by the low-brows of "abstract empiricism". Data, however, are a peculiar commodity insofar that they have to be produced. Which means most of the time it is costly, painstaking, and time consuming to produce data. By and large, data belong to one of two families. Primary data are collected by the researcher himself. Secondary data are collected by other agencies, mainly public and private large-scale organizations, in which case the researcher can only perform what is known as secondary analysis, and he obviously has no control over the collection of this data. In the latter case it is also useful to further distinguish between data collected for analytical purposes by agencies like the various census bureaus, and those collected for administrative purposes, such as data on health collected by hospitals, or data on education collected by educational institutions, or mortality data collected by governments and other agencies. These are process-produced data; data created as by-products of administrative activities.

There are disciplinary areas that have a high degree of institutional stability and control over their data such as physics, but also engineering, medicine or law. The latter two fields, of course, provided the original kernel for the development of universities and studia generalia, in Bologna, Padova, or Salerno. Other fields are less stable, usually in the area of the humanities and the social sciences. But in all cases the organized character of the type of knowledge provided by academic disciplines is predominant.

Organized knowledge is the practical rather than speculative knowledge accumulated by governments and their administrative apparatuses, by corporations, political parties, trade unions, churches, and other institutions. Thomas Kuhn has been the leading figure in analyzing innovation processes in the institutions of organized knowledge (Kuhn 1962). These types of secondary data are essential for social scientists, and in fact, as noticed long ago by Paul F. Lazarsfeld and Stein Rokkan, among others (Rokkan 1976), the work of sociologists, especially in Europe, is largely based on interactions with the institutions in charge of this type of knowledge.

Le Suicide by Emile Durkheim is an excellent example of the advantages and drawbacks of secondary analysis of process-produced data. This masterpiece of sociological research gave empirical evidence of the theoretical tenet that a cogent collective agent can influence individual behavior even to the extreme gesture of annihilating the genetic commandment. It could not have been written without access and a hard sweated one at that, to secondary data. No individual scientific actor could collect data on events that are so numerous, and so highly dispersed in time and space. On the other hand, relying on data collected by other agencies means that the researcher relies on somebody else's definition of events. One of the most damaging critiques to Durkheim's work is that his sophisticated theoretical definition of suicides, is nullified by the fact that the cases of suicides in the data used were defined by officers or judges according to completely different, and uncontrollably variable, criteria.

The work of historians, too, would probably not be possible without access to the organized knowledge embodied in the archives of agencies of all kind. The early development of economics as a quantitative discipline was greatly favored by the availability of organized knowledge collected by public administrations. The same applies to demography, which could base itself also on data accumulated by the church, and in general to the whole field of statistics, which could today be defined as policy sciences (Cavalli 1972).

The Durkheim syndrome, the need to use data not collected by scholars, but by public employees, is a constant problem for social scientists using information coming from the realm of organizational knowledge. At the beginning of the nineteenth century Melchiorre Gioja expresses the irritation of a scholar dependent on low-quality public data. He took issue with "the many questions that various inept busybodies called secretaries send from the capital to the province. Questions that never produced other than the following three effects:

1. fear that the Government seeks the basis for some aggravation, and therefore opportunistically false answers;
2. ridicule resulting by the silliness, inconsistency, and imprecision of the questions, and thus answers biased by contempt;
3. heaps of paper uselessly encumbering archives, if the government mistrust them, or very serious mistakes if it uses them, not to speak of the time stolen from the municipal or provincial administrators who must prepare the answers." (Gioja 1852, p. 5)

A powerful answer to these problems came from the development of survey research, particularly from the 1940s on by American sociologists who developed the "art of asking why"; namely the inquiry into the rational motivations of individual behavior. It is not surprising that these new methods originally developed in two crucial areas of behavior: the choice of candidates in an electoral process, and the choice of a product in consumption behavior. There has been much criticism of mercantile attitudes in voting, and of course it is quite clear that some of the value motivations that are mobilized in choosing a candidate are evidently not the same as those that are mobilized in picking up one granola package rather than another from a shelf. But the critics miss the point. The two areas of behavior are similar, and the inquiry into behavioral motivations should not be diverted by undue considerations of political or ethical correctness. Little wonder that scholars had to develop new data collection tools on this type of behavior. And equally not surprising is that political and economic elites are willing to invest resources in the arduous enterprise of predicting the aggregate outcome of individual behavior. A prudent politician today constantly monitors the opinion of the electorate. And economists make use not only of large-scale models of the behavior of macroeconomic systems, but also of assessments of consumer behavior. Politicians, bureaucrats, entrepreneurs, and managers would have a hard time doing without the tools provided by the social sciences.

Survey data collection methods, however, presented the social scientist and his institutions with novel problems. Pollsters and survey people in general produce huge quantities of individually uninteresting questionnaires. Punched cards were developed to hold data; at first the cards were processed manually as the "McBee cards," but soon after machine-readable cards were developed, such as those punched with the Hollerith code (universally known as IBM card). Cards were easier to store than questionnaires, but it was easy, in a routine research process, to lose the "codebook" of the research so that in many cases the cards were useless, even if they contained relevant information (Rokkan 1976). Thus the storing, handling, processing, and redistributing of punched cards required specific skills. Furthermore, the traditional institutional structure of universities and research centers is not well adapted to take on these tasks. Originally social scientists turned to libraries to store their data (Lucci and Rokkan 1957), but in the late 1950s and early 1960s libraries were not equipped to handle large masses of data requiring mainframe computers. The big machines were housed in separate structures and tended by IBM technicians. Social scientists had limited access to the mainframe computers and therefore, the data, until the scientists developed their own separate institutions on the model of the Inter University Consortium for Political Research at Ann Arbor, Michigan (ICPR, later turned into ICPSR when social research was added). In Europe the vision and farsightedness of scholars such as Philip Abrams in the United Kingdom, Stein Rokkan in Norway, and Erwin Scheuch in Germany, helped establish the first archives in Essex, the Social Science Research Center, in Bergen, the Norwegian Data Service, and in Koeln, the Zentralarchiv fur Empirische Sozialforschung. Later on these archives developed their own organizations, first IFDO, International Federation of Data Organizations, and later on CESSDA, Council of European Social Science Data Archives. Archives developed in Italy at the Istituto Superiore di Sociologia of Milano (ADPSS), in Denmark (DDA), in the Netherlands (the Steinmetz archive), in Belgium (BASS), and in France (BDSP in Grenoble and in several other places).

SOCIAL SCIENCE DATA ARCHIVES

Cultural and technological changes led to the creation of Social Science Data Archives (SSDA). SSDA are scientific institutions that retrieve, store, and distribute large amounts of data on social science. The oldest and the most important SSDA were established in the United States—where an emphasis on quantitative social research is deeply rooted—when there was growing attention by both public administration and the scientific community to the use of social indicators as standards for the population welfare level. These social actors turned to social indicators for help in planning, applying, and the evaluating public-assistance programs when it became apparent that using economic indicators alone was insufficient and inappropriate.

The first major contribution of social indicators is the Recent Social Trendsstudy, supervised by the sociologist William Ogburn and prescribed by the U.S. Government toward the end of the 1920s (Bauer 1966). In 1946 the Employment Act was published, which was a systematic collection of information on economy intended to affect policies and programs that would sustain employment rates. Most of the studies and undertakings of the 1960s described a "great society," which could overcome the widening economic and social gap.

The consciousness of social problems, together with the necessity of endowing a collection of social data, prompted the publication of Towards a Social Report at the end of the 1960s (Olson 1974). This publication was intended to be "a first step toward the evolution of a regular system in social reporting;" but still, like other similar and contemporary writings, data were used just as illustrations supplementing the text.

Technological developments that contributed to the establishment of SSDA include previously unseen data collection techniques and new quantitative methods to organize and analyze those same data. Throughout the 1960s improvements in computing technologies, specifically in data gathering and storing, allowed researchers to do previously unthinkable levels of analysis (Deutsch 1970).

The first SSDA were born autonomously, unrestricted by publicly administered archives or by the institutions traditionally related to the collection and promotion of data (libraries, museums, and data archives). The earliest SSDA were created in the United States, where in 1947 the Roper Public Opinion Research Center opened, followed by the Inter-University Consortium for Political and Social Research (ICPSR) in 1962 at the University of Michigan in Ann Arbor. In 1960, the University of Koln in Germany developed the Zentralarchiv fur Empirische Sozialforschung (ZA). A few years later, in 1964, the Steinmetzarchief settled in at the Amsterdam Arts and Sciences Real Academy. The Economic and Social Research Council Data Archive (ESRC-DA) was created at the University of Essex in 1967. These archives specialized in public opinion surveys and social research, but they also focused attention on the great amount of data complied by statistical bureaus and public agencies (Herichsen 1989).

SSDA developed a culture of data sharing; data exchange and the repeated use of available data for new research projects intensified with the introduction of statistical packages for the social sciences and more compact media for data transfer. As data production and SSDA grew, more systematic acquisition policies were implemented, and transborder cooperation resulted in the exchange of data processing tools and of emerging archiving and service standards (Mochmann 1998).

In the early 1970s SSDA were created: in Norway in 1971, in Denmark in 1973, and in the United States (at the University of Wisconsin in Madison, U.C.L.A., and the University of North Carolina in Chapel Hill). In the 1980s SSDA were created in Sweden (1980), France (1981), Austria (1985), and then Canada, Hungary, Israel, Australia, New Zealand, and Switzerland (1993).

Currently, the most important SSDA is the ICPSR at the University of Michigan, which has over 40,000 data files and gathers information from more than 300 institutions. Besides data retrieval, processing, researching, and transferring, it publishes an annual catalog and a four-monthly bulletin, cooperates in great projects concerning data collection (e.g. the General Social Surveys and the Panel Study on Income Dynamics), conducts formative training (it organizes a summer school on methodology and statistics for social science), and offers educational activities. The ICPSR has been open to foreign institutions since the early 1970s.

SSDA differ on budget and staff size, functions, amount of data files in acrhives, and technical characteristics (e.g. type of hardware and software used, online data accessibility) (IFDO 1991). The SSDA network is tied, however, by international associations such as the Council of European Social Science Data Archives (CESSDA was instituted in 1976 to facilitate cooperation between the most important European SSDA), and the International Federation of Data Organizations (IFDO was founded in 1977 to enhance the cooperation already started by CESSDA).

CESSDA promoted the acquisition, archiving, and distribution of data for social research throughout Europe, facilitated the exchange of data and technology among data organizations, and supported the development of standards for study description schemes to inform users about the archival holdings, classification schemes for access to variables by subject and continuity guides for coherent data collections. In the 1970s and 1980s, European membership grew continuously and CESSDA started to associate and cooperate with other international organizations sharing similar objectives. Today Europe has a good coverage of CESSDA member archives, while planning processes are under way in several more countries that still lack a social science infrastructure.

Recent trends reveal a type of limitation to the growth and proliferation of national SSDA. This occurs partly through the opening of decentralized bureaus and partly through more specialized data file collections—ones with more well-defined inquiry areas. Exemplary regarding this last issue is the Rural Data Base of the ESRC, or even the CESSDA internal agreement aiming at a study field subdivision between the different European SSDA (Tannenbaum 1986).

SSDA have been useful not only in their specific field (especially retrieving, storing, gathering, and making available data) but also have improved general quantitative research techniques, secondary analysis, statistical and administrative software packages, and have made available the best hardware in social research. Most of the SSDA publish their own bulletins, organize methodology schools for social science, hold updating seminars, and cooperate in research projects. In the second half of the 1980s SSDA were particularly active in defining the standards for the information system to be used among these institutions. More recently, SSDA have been working on international access practices on two levels: developing standard study description procedures (to define each data file) and promoting access to online archives.

SOCIAL SURVEY INSTRUMENTS

Major social survey instruments are those institutionalized initiatives that have produced some of the most interesting results from the autonomous research applied to problems of general interest. The General Social Surveys, the Continuous National Surveys, and the U.S. National Surveys are the most interesting examples of longitudinal studies. Eurobarometers and the International Social Survey Program (ISSP) are discussed here because, although not strictly comparable to general social surveys, they are two of the most comprehensive and continuous academic survey programs. Although it is not discussed here, a student of the social sciences should examine the European Household Panels. This project integrated national household panels from Germany (since 1984), Sweden (since 1984), Luxembourg (since 1985), France (since 1985), Poland (since 1987), Great Britain (since 1991), and Belgium and Hungary (since 1992). These panels include variables on household composition, employment, earnings, occupational biographies, health, and satisfaction indicators. (In order to create an international comparative database for microdata from these projects, the Panel Comparability Project (PACO) was formed).

General Social Surveys. General Social Surveys (GSS) were developed to set out data on demographic, social, and economic characteristics of the population, as well as opinion data on social life (e.g. family, politics, institutions, relationships). This important survey instrument improved over time. In the mid 1940s the Survey Research Center of the University of Michigan conducted panel studies on a national level on specific issues, such as political behavior, socioeconomic status, and consumers' attitudes. In the early 1970s, the National Opinion Research Center (NORC) of the University of Chicago organized the first GSS, which was immediately followed by the Continuous National Surveys program.

Following the consolidation of the research method based on enlarged quantitative surveys on a national level, some institutions committed to the transfer of the results, both within the U.S. academic community and toward universities and research centers worldwide. One should note that data provided by the GSS are given a particular treatment inside the ICPSR, which, besides transferring the data, gathers the sociologists interested in the study and the arrangement of social indicators. The importance of the GSS in U.S. social research is exemplified by the more than 1,000 publications related to GSS data analysis listed in the catalog prepared by NORC.

The GSS is done through interviews of a representative sample of U.S. residents who are over 18 years old and are not institutionalized. Representativeness is guaranteed by a complex sampling mechanism, which relies on a multilevel selection over metropolitan areas, municipalities, regions, and individuals. Sampling criteria have been modified throughout the course of different GSS, but to keep the results comparable between one edition and the following ones correctors have been added to allow for these adjustments.

The GSS questionnaire consists of standard questions that are asked each time, and sometimes groups of questions on specific themes are included. The topics treated are the ones most interesting in the study of society and its trends, with a particular focus on family, economic and social status, gender and ethnic group relationships, and moral questions. Political behavior and working activity are not included because the former is already studied in detail through surveys organized by the Institute for Social Research of the University of Michigan, and the latter is well described in the research on labor forces arranged by the U.S. Bureau of the Census.

Given that one of the fundamental aims of the GSS is to provide a general view of time trends, not only in population characteristics, but also and overall in opinions, evaluations, and behaviors over the most important topics describing the social scenery, these questions are included only from time to time, so the questionnaire is not weighted excessively. Some questions are included only two years in three, others every 10 years, and some only after particular and significant events. Over 100 sociologists worked on the first draft of the questionnaire (in 1972). These sociologists devised a final, definite version by voting for each single question. Every year the selection of questions is done by an ad hoc committee, selected by American Sociological Association members.

Some questions utilized in the GSS come from the national surveys run before 1972 that were promoted by commercial research institutes (e.g. Gallup, Harris), university research institutes (especially ICPSR), and also federal commissions organized to study particular phenomena. Data comparability is assured by the question scheme, which is the same of the original study.

Continuous National Surveys. The Continuous National Surveys (CNS) are national studies (the first one dates back to 1973) conducted monthly, with the aim of supplying the various governmental agencies with the necessary data (e.g. welfare indicators) to schedule social programs. The sample plan consists of persons selected on the basis of their living groups. On this regard, the NORC has carefully prepared a master probability sample of households, that is, a multi-stage sample to collect on a first stage municipalities or else groups of municipalities. From them, all districts or block groups are selected, and then the proper cohabitational groups in which to choose the individuals to interview are selected.

U.S. National Surveys. U.S. national surveys on representative samples of U.S. population have been held since 1974 by the Survey Research Center of the University of Michigan. Some of these studies are held regularly (e.g. the Surveys of Consumer Finances, the Survey of Consumer Attitudes and Behavior, and the Panel Studies of Income Dynamics). In particular, the yearly Panel Study of Income Dynamics is one of the most interesting surveys on income trends, and, specifically, on the possible causes for changes in the economic status both of households and single individuals.

The representative sample initially extracted consisted of 2,930 households, to which has been added 1,872 households that were already survey subjects by the U.S. Bureau of the Census on the income topic in the two previous years. Each year these households are re-interviewed, and the sample has grown as members of the original sample established new households. For the first time ever, phone interviews were tested in these researches, and since then the telephone has become broadly used in social research and panel studies.

Eurobarometers. Eurobarometers are opinion studies that have been held twice a year since 1973 in what are now the European Union countries. These studies sample about 1,000 individuals for each country, which represents the population over 15 years-old. The aim of these comparative surveys is to learn about the attitudes of European citizens on some broad interest topics. Questions regard public attitudes toward European integration, but sometimes also address specific problems of a single country or more generally economic, political, and social conditions. Two important functions of Eurobarometers are: being cross-national and easily comparable surveys, they helped integrate social research throughout Western Europe; and, they allowed (and still allow) analysis on social changes in Western Europe. As a matter of fact, there is no survey similar to eurobarometers in what concerns a regular check over time (more than 20 years to date) and space (every single European country). Eurobarometers are now at the disposal of the academic community thanks to the ICPSR and the ZA of Koln.

The International Social Survey Program. The International Social Survey Program (ISSP) combines a cross-national survey with a longitudinal time dimension by replicating particular question modules, ideally in five year intervals. The first survey on the "role of government" started in 1985 in four countries (the United States, Great Britain, West Germany, and Australia). Since then the ISSP has grown rapidly and now covers more than 30 countries around the world, including Bulgaria, the Czech Republic, Hungary, Latvia, Poland, Russia, and Slovenia. Topics of the ISSP have included social networks (1986), social inequality (1987, 1992), family and changing gender roles (1988, 1994), work orientations (1989, 1997), religion (1991), and environment (1993). Role of government was replicated in 1990 and 1996. The official data archive of the ISSP is the ZA, which makes the integrated data sets available via the archival network.

DEPOSITED, ACCESS PROCEDURES AND USE

The existence of SSDA in Europe and in the United States has had a positive effect on the scientific community because SSDA allow access to some data that is particularly useful in secondary analysis. In other words, researchers can collect and use data from different surveys (with particular hypothesis and conceptual frames) to support their own works.

SSDA vary on the type and volume of data that is held and delivered and on the exact services offered. The data differ within and between SSDA in terms of subject matter, time period, and geographical area covered. The data may be individual or aggregate, and the variables may be cross-sectional or time series and suitable for comparative or longitudinal analysis, or both. Charging policies differ from one SSDA to the next and depend upon the type of service being provided, the specific data set demanded, and the institutional affiliation of the requester.

Many data formats are used by SSDA to preserve and deliver the data that is increasingly required to be machine readable, and there is also an increasing demand for the archiving and disseminating of machine-readable metadata.

Despite the heterogeneity of SSDA, there are common goals and tasks. Although we know more about the SSDA in Western Europe, we hope to represent all SSDA in the following list of goals and tasks:

• To promote the acquisition, archiving, and distribution of electronic data for social science teaching and research and to exchange data and technology;
• To exploit the potential of the Internet by the expansion of web-based services and the improvement of dissemination to allow users more direct and immediate access to the data;
• To develop and use metadata standards for the management of data;
• To provide users access to comparative data and to multiple data sources across national boundaries;
• To allow users to receive all or subsets of data via download or on portable media in one of a number of formats;
• To extend the users' base beyond traditional boundaries;
• To better serve the needs of an increasing number of novice users, particularly in terms of data analysis software and improved searching aids through the development and use of social science thesauri and user-friendly interfaces.

The technological changes that contributed to the creation of SSDA have had far-reaching significance for social scientists. The changes throughout the 1990s were considerable: from simple writing (word processing) to data production (especially computer assisted interviewing technique); from archiving (from punch cards to floppy disks and CD-ROMs) to data analysis (the creation of various new software); from bibliographic reference (online access to bibliographic databases) to communication (e-mail) (Cooley and Ryan 1985). The boundaries of technological innovations are not yet clear. The compound archive of integrated data, text, images, and sounds (Garvel 1989), and the Geographical Information System, by which we can analyze data concerning space are examples of previously unimagined products that are possible because of computers (Unwin 1991). These instruments require high-level storing proceedings so that data can be a real resource in the development of empirical knowledge.

While social research in earlier decades lamented the lack of reliable data (in the 1960s the social sciences were considered data poor, and infrastructural support for social research was lacking) the situation has changed. Thousands and thousands of data sets, some of them dating back to the 1940s, are stored for secondary analysis in SSDA. These data sets represent a vast potential for comparative research on historical developments and social change.

PROSPECTS FOR THE FUTURE

In 1994, the European Commission recognized Social Science Data Archives as legitimate institutions to be considered as applicants to the large-scale facility program for scientific research. This is tantamount to recognizing that the social sciences have infrastructure needs equal to those of the "hard" sciences. This goal was originally pursued through the work of a study panel created by the General Directorate for Science and Technology of the European Commission and particularly its Training and Mobility of Researchers program. The study panel met twice, and was entrusted with the tasks of: a) identifying the future priorities of European scientists concerning access to large installations in the social sciences, both in the short and in the medium term (i.e. in the period 1994–2004); b) suggesting ways to meet these priorities—with an indication of their respective costs and benefits—in light of the present and expected availability of large installations, taking into account existing and planned future international cooperations, both within the community and outside as well as the particular needs of researchers working in regions where such installations do not exist; and c) consulting representatives from member states and reporting on what facilities, if any, might be available for possible support in the frame of successor programs. A large-scale facility has, to date, been defined as a large research installation that is rare or unparalleled in Europe and that is necessary for high-level good-quality research. Such facilities tend to have high initial costs of investment and comparatively high operating costs. The large-scale facilities program is intended to provide scientists with access to large installations within and especially outside their own countries, thereby promoting the mobility of researchers and encouraging the creation of a Europe-wide research community. And likewise, these installations are better utilized when their services, facilities, and knowledge are available to a wider community of users. Thus the beneficiaries of the large-scale facilities program are researchers who are provided with access to the facilities and the organizations that receive support for the use and improvement of their equipment. The level of support for such facilities has, to date, been based on the quality and unique features of the facility and the value—particularly in advanced training—to potential users.

As a sequel to the work and recommendations of the study panel, between 1994 and 1999 three Social Science Data Archives, the SSRC Essex archive, the German ZA, and the NSD in Bergen have been included in the large-scale facilities program together with other social science institutions. In addition the General Directorate for Science and Technology created a round table on large-scale facilities in the social sciences. This round table was constituted by the European Union commission to suggest actions needed to support a given field of scientific activity. The prospects of Social Science Data Archives are thus quite positive, and for the first time they have been given a tool capable of furthering their original aims.

references

Bauer, R. A. (ed.) 1966 Social Indicators. Cambridge, Mass.: MIT Press.

Cavalli, A. 1972 "La Sociologia e le Altre Scienze Sociali. Prospettive di Integrazione Interdisciplinare." In P. Rossi, ed., Ricerca Sociologica e Ruolo del Sociologo. Bologna: Il Mulino.

Cooley, R. E., and N. S. Ryn 1985 "An Information Technology Strategy for Social Scientists." UniversityComputing7.

Deutsch, K. 1970 "The Impact of Complex Data Bases on the Social Sciences." In R. L. Bisco, ed., DataBases, Computers and the Social Sciences. New York: Wiley.

Garvel, S. 1989 "National Archives and Electronic Records: Where Are We Going?" IASSIST Quarterly 3–4.

Gioja, M. 1852 Filosofia Della Statistica. Torino.

Henrichsen, B. 1989 "Data from the Central Bureau of Statistics to the Social Science Community: The Norwegian Experience." IASSIST Quarterly fall–winter.

IFDO 1991 "IFDO Survey of Computerized Catalogues." IFDO News September.

Kuhn, T. 1962 The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Lucci, T., and S. Rokkan (eds.) 1957 A Library Center ofSurvey Research Data. New York: Columbia University Press.

Mochmann, E. 1998 "European Cooperation in Social Science Data Dissemination." Working paper, round table meeting on infrastructures for social economic research. Koln, 29–30 June.

Rokkan, S. 1976 "Data Services in Western Europe." American Behavioral Scientist 19:443–454.

Tannenbaum, E. 1989 "Sharing Information Begets Information." IASSIST Quarterly, Summer.

Unwin, D. 1991 "Geographical Information Systems and the Social Science Research." ESRC Data ArchiveBulletin 48.

Guido Martinotti

Sonia Stefanizzi