official statistics Statistical information produced, collated, and disseminated by national governments, their agencies, and the international bodies which link them. These data are almost invariably nationally representative, because they are obtained from complete censuses or very large-scale national sample surveys, and they usually seek to present definitive information conforming to international definitions and classifications or other well-established conventions. The impersonal character of official statistics, and their resistance to innovation, stand in sharp contrast to statistics and data-sets from other sources: academic research, market research, independent research institutes, commercial organizations, local, regional, and state bodies.

Official statistics used invariably to be published in large tomes and preserved in book libraries as the definitive record. This method of dissemination underlined their inflexibility, forced the results of enquiries to be presented in a relatively small number of selective statistical indicators and indexes, and is disappearing rapidly with extensive use of information technology. In the 1990s and beyond, government statistical information is far more likely to be disseminated as computer tapes of anonymized microdata, as data subsets on diskette, or as specially compiled harmonized time-series on diskette with monthly updates sent through national and international computer networks or the telecommunications system. Instead of finding the statistics one needs in a printed volume, interested parties may now have to create and extract them from a diskette with built-in software, or dial into a national-user-service computer centre from a desk-top terminal and extract the figures required from regularly updated databases.

As yet, national governments have refused to pass legislation permitting large-scale data linkage between government agencies, and such a proposal would require a system of unique identity-card numbers (or other referencing system) for each person in the population, to be used from birth to death. Data protection policies currently impede such a development within the commercial sector also, although area-based social profiles are widely used in Western industrial societies. Until such time as massive government data-banks are permitted, the only information held by central government agencies is that supplied directly to them by citizens in response either to public enquiries, or to regulations that specify events to be formally recorded in public records and the like.

Almost all official statistics originally came from registers in which were recorded the details of specified events as and when they occurred: births, deaths, marriages, divorces, crimes, certain contagious diseases, and later notifiable diseases such as cancer, AIDS, and so on. Similar procedures are involved in administrative records of non-compulsory activities, such as claiming unemployment benefits. A decreasing proportion of data comes from these sorts of records. Their chief advantage is that they constitute complete censuses of the events in question, and are thus reliable, up-to-date, and cheap to use as sources of statistics. The obvious disadvantage is that only a fairly narrow range of information lends itself to being collected through such procedures. While the fact of a death may be easy to record, its cause may be more arguable; and other related factors are simply too complex to deal with in that way, however relevant. Certain statistics are still obtained from compulsory registrations and administrative records—for example hospital records of patient illnesses, police records of crime, and records of people claiming various types of social insurance benefits. But they are supplemented with, and increasingly replaced by, specially designed data collections: compulsory censuses of population, housing, and employment, and voluntary interview surveys with national samples of the whole adult population or particular sections of it.

Censuses are usually carried out only once a decade, and are supplemented by a range of regular surveys that provide statistical information on a quarterly basis, an annual basis, or at less frequent intervals. Most countries now have an annual multi-purpose household survey to collect social and economic data in the intercensal decade. In Germany it is called the microcensus and in the United States the Current Population Survey; in most other countries it is called the Labour Force Survey; and it collects a much wider range of information than the census it replaces. In addition there is a great variety of other data collections using survey methods with personal, postal, or telephone questionnaires to collect information of a kind that can be coded and quantified to produce statistics on a great range of matters: earnings and incomes; trade; illness, health, and usage of the medical services; housing, job change, and migration; household expenditure patterns; the Retail Price Index; national economic accounts; government expenditures; patterns of food consumption and nutrition; any experience as the victim of crime; leisure activities; travel patterns to work, for business, and for leisure; international travel, immigration, and emigration. In addition there is a huge range of ad hoc sample surveys carried out by national governments on a wide variety of topics of public concern, sometimes on a once-off basis, sometimes with repeat surveys every five, ten, or twenty years.

The exact number and variety of regular and ad hoc national government-funded surveys varies in line with local needs and circumstances. In many cases they are funded and carried out jointly with other bodies, such as independent research institutes, international bodies, charitable foundations, or commercial organizations. The dividing-line between ‘official’ and ‘non-official’ statistics and data-sets is being eroded by the change of emphasis from public-sector records and registers, which are necessarily a government preserve, to interview surveys, which are available to all sectors of society, and may even be more successful if carried out by a non-governmental agency.