Descriptive Statistics

The leading representative of descriptive statistics was the astronomer Pehr Wilhelm Wargentin, long-time secretary of the Royal Swedish Academy of Sciences, who played an instrumental role in securing an institutional base for Swedish statistics. Descriptive statistics had as its objective to reveal, describe, and interpret data, but not to prescribe how the data might be used. In parallel with the limitation of its aims, descriptive statistics came to be confined to a subject where data might be gathered without insuperable difficulty—to population studies. Slowly, sound and methodologically conscious

population statistics began to squeeze out the extravagant attempts at precision and the lofty social aspirations of utopian and practical statistics. By the 1770s, population statistics would become the only type of statistical work undertaken in Sweden.

The general outlines of the growth of Swedish population statistics are well known.^[33] In 1749 influential Swedish mercantilists and the Academy of Sciences succeeded in their campaign to establish an Office of Tables (which would become the Central Bureau of Statistics in 1858). Parish priests were required annually to complete printed forms reporting the numbers of births (classed by sex), deaths (by sex, age, and cause), and marriages, as well as the total population (by age, sex, estate, and occupation) for the parish. The tables were forwarded through a series of governmental agencies to the Commission of Tables, whose task was to summarize the results and transmit them to the Riksdag and the king. The Office of Tables thereby compiled the first set of population statistics in the world based on regular counts of total population. An efficient parish registration system that did not miss a single soul, a permanent institutional base, and a population unusual for its ethnic and religious homogeneity, disciplined by an established church with ample opportunity to exercise formal and informal control, contributed to the success of the venture.

But the very success of the Office of Tables represented a retreat from larger ambitions. Its reason for existence derived from the central importance of population in the mercantilist program, but population studies alone were only part of the social analysis urged by the utopian and practical statisticians.

Initially, the staff of the Office of Tables shared the optimism of other statisticians. In particular, the Commission of Tables (dominated by statisticians and civil servants) dreamed of a gigantic survey of all components of the economy. Collated, combined, and

[33] See August Johannes Hjelt, Det svenska tabellverkets uppkomst, organisation och tidigare verksamhet (Helsingsfors: O.W. Backmann, 1900); E. Arosenius, Bidrag till det svenska tabellverkets historia (Stockholm: Nord. Bokh., 1928); and "The history and organization of Swedish official statistics," in John Koren, ed., The history of statistics (New York: American Statistical Association, 1918).

compared, the numbers and tables would constitute a map of Sweden's resources, strengths, and weaknesses, and provide the political authorities with an effective instrument for governing the country. This ambitious program is evident in the highly secret reports delivered by the Commission to the Riksdag and the king in 1755, 1761, and 1765.^[34] Mind-boggling arrays of figures classified Sweden's population under a total of sixty-one headings; virtually all individuals were linked to their work and capital-producing capacities.

In their aim, these reports appear to be a faithful application of political arithmetic. Individuals were assigned categories according to their economic and hence political value to the state. First came providers, then consumers, and finally a category of wholly "superfluous members" (notably tavern staff and servants) numbering 10,336. The figure for emigration—8,059 in 1761—is just as precise; when converted into value using the methods of utopian statistics, it represented an annual capital loss to Sweden of 9 1/2 million dalers. When the potential of the emigrants to produce offspring was figured in, the loss amounted to no less than 19 million dalers.

At first the Swedish parliament showed much interest in the data and their implications. It appointed commissions and ordered certain reforms, especially in the medical field. Soon, however, the initiatives were tabled or defeated. Decisions disappeared mysteriously en route to the king for implementation. The important table of estates and occupations was originally required annually. But already in the 1750s the requirement was changed to reporting once every three years; later this was reduced to once every five years. Quantitative analysis of natural data like births and deaths remained noncontroversial, but attempts to derive social diagnosis or prescribe social therapy from the figures excited objections. Political arithmetic fell out of favor as a political instrument. With the constitution of 1772, parliamentary reponsibility for the Office of Tables formally ceased.

As officials lost enthusiasm for statistics, so too did advocates of descriptive statistics rebel against the use of their subject as an

[34] Reproduced in August Johannes Hjelt, De första officiela relationerna om Svenska tabellverket (Helsingfors, 1899), and N.V.E. Nordenmark, Pehr Wilhelm Wargentin (Uppsala: Almqvist & Wiksell, 1939).

instrument of state power. For fifteen years, population information, broken down by estate, occupation, and age group, had been kept under wraps by the Office on Tables. During the early 1760s this suppression of population figures as a state secret occasioned heated debate. Only in 1764 was the official population of Sweden (2,383,113) first disclosed. Before long the detailed information underlying the estimate was available for study by anyone who wanted it.^[35] As the gap between statistics and state widened, statistics had the opportunity to develop independently of power interests or practical applications. A gradual drop in the number of references in the Academy's Transactions to practical political aspects of statistics reflects this shift.