data archive

data archive, data-bank

A storage and retrieval facility or service for social scientific data . Data archiving has grown in parallel with the development of secondary analysis as a recognized field of social research. The International Federation of Data Organizations for the Social Sciences (IFDO) was founded as recently as 1977 by four North American and seven European data archives. These archives act as lending libraries for machine-readable information collected by academics and government agencies. In the vast majority of cases, the information in question is derived from social surveys (most data archives are in fact survey archives), although there are now archives which store the taped interviews obtained by oral historians and others. Survey archives are often depositories for official statistics , including censuses , as well as non-official data-sets such as opinion polls and academic surveys. In Britain, for example, the Economic and Social Research Council established such an archive (the UK Data Archive) at the University of Essex in the late 1960s. Similar archives exist at many universities in America and Europe. (A representative sample of archives are listed in, Secondary Analysis in Social Research, 1982.)

All data archives have three primary functions: the collection, storage, and preservation of data (usually in the form of magnetic tapes); the dissemination of data within the national social science community and across international boundaries; and the development of facilities and techniques of analysis to encourage the widest and most intensive possible use of data. Different archives attach different degrees of importance to these objectives: some operate a broad acquisition policy whereas others are more selective; some make access to the data-sets easy by publishing detailed catalogues of holdings; others specialize in secondary analysis conducted by in-house researchers.

The major advantage of data archives is that they constitute a rich source of inexpensive and publicly available information. However, unless the data-sets held by the archive are fully documented (with complete and detailed codebooks, for example), then the researcher engaged in secondary analysis may be unaware of important limitations on how the original information was collected and how it may legitimately be used.