The influence of the Royal Statistical Society on the census
Matthew Woollard
What is now called the Royal Statistical Society was founded in March 1834. Its name at the time of foundation was the Statistical Society of London — a name it retained until it gained a Royal Charter in 1887. Almost all of the men involved in the statistical analysis of the census and registration data were at one time or another members of the Society. William Farr (1807-1883) held the post of President from 1871 to 1873, and was only the second President who was not a member of parliament. Sir Bernard Mallet (1859-1932) held the post between 1916 and 1918. Many officials in the General Register Office either held other posts or were members of the Society.
Before each of the decennial censuses, commencing with the census for 1841, the Royal Statistical Society appointed a committee to advise the Government on the method of taking the census, and on the information collected. The first success of the Society was in 1841 when a new Census Act (Example of Census Enumerators' Books, 1871) was hurriedly passed to give responsibility for the taking of the census to the newly formed General Register Office.
The recommendations made by the SSL's committee were:
use the administrative machinery set up by the Poor Law Commissioners and the Registrar General to take the census in England;
collection of nominal information (as the "only security against fallacious returns" (p.97) (and thus individual information);
centralised tabulation of information;
collection of information about age (to combine with figures from registration), sex, marital status, occupations, place of birth, "religious persuasion" and health, i.e., healthy or "sick or permanently infirm" with the sickness or infirmity.
These recommendations were published in the Journal in April 1840. At this time, John Rickman was in the process of preparing a bill for the census. Rickman's bill was deposited in early June 1840 and provided for a census along the same lines as in 1831 (Bill for taking an account, 275--), but it was withdrawn within a month and replaced with a second bill with the same title, but bearing clear signs of influence from the Statistical Society's committee. This bill deposited on 13 July became law in August, but lacked the detail to properly carry out the census, resulting in an amending Act in the following year.
The statistical society's committee which made these recommendations which remained of significance to census taking throughout the twentieth century was made up of George Porter, T. R. Edmonds, William Farr and Leonard Horner (the chief factory inspector), but Cullen (96–97) suggests that the actual influence of the Society on the bill was limited.
By 1850 the SSL was taking full credit for the introduction of the full method. In the SSLs sixteenth annual report (published in vol 13 (1850), 97–102): "The immediate effect of its [i.e., the SSL's report] production was to cause the Government to withdraw the bill actually before the House of Commons....A bill drawn up on the plan contained in your Committee's report was then substituted, and became law." As Cullen makes clear the truth was not so simple as the Registrar-General Thomas Lister Thomas Henry Lister (1800-1842) was greatly involved in the drafting of the bill and while his envisaged method of administering the census was in concordance with the SSLs his idea of the scope of the census was strictly limited. However, buoyed by some success, the SSL unwittingly inaugurated a decennial institution of making a series of recommendations for the forthcoming censuses. While these may have been considered by the General Register Office very few of their suggestions were implemented until the end of the nineteenth century.
The references below show the sources for the principal recommendations made by the society, and for the period to 1891 they can be summarised rapidly because they generally speaking they only ask for more detailed information to be collected rather than any change in the administration. Before the 1851 census they suggested the number of rooms in each house should be collected and some agricultural statistics should be collected simultaneously; before 1861 they suggested that the educational and religious censuses should be carried out again and that an additional (optional) question on religion should be added to the householders' schedule; before 1871 the new question proposed related to literacy. Before the 1881 census some more radical proposals were made including publication of consistent reports across the nation and than the census should be quinquennial rather than decennial. For the 1891 round of recommendations, the society suggested more strongly the necessity for a quinquennial census, as well as for the formation of a permanent census office. The committee's proposals in 1900 were essentially the same as for the previous census, though again the proposal for a quinquennial census stands out.
The 1911 series of recommendations were the most detailed and published as a separate pamphlet. They demonstrate considerable thought, and include the proposal to abandon the enumerators' books, and to tabulate directly from the schedules along with more detailed proposals relating to the 'imperial' census.
Was the statistical society an effective lobbying body? In short, no. For various reasons, the large majority of their recommendations were not taken forward, but it can not be said that they were without influence. Further research is necessary on the decision-making processes relating to the administration of the census to clarify the extent of their influence.
REFERENCES
E. Wyatt-Edgell, 'Remarks on the plan adopted for taking the census in 1841, with suggestions for its improvement', Journal of the Statistical Society of London, 12 (1849), 71–77.
Michael J Cullen, The statistical movement in early Victorian Britain: the foundations of empirical social research (Hassocks, 1975).
[Henry W. Macrosty and James Bonar], Annals of the Royal Statistical Society, 1834–1934 (London, 1934).
Bill for taking an account of the population of Great Britain; and of the parish-registers, and annual value of assessable property in England, BPP 1840 III.275–.
'Report to the council of the Statistical Society of London, from the Committee appointed to consider the best mode of taking the census of the United Kingdom in 1841', Journal of the Statistical Society of London, 3 (1840), 72–102.
'Proceedings of the census committee of the Statistical Society of London, 1850', Journal of the Statistical Society of London, 13 (1850), 267–269.
'Recommendations of the Council of the Statistical Society as regards the census of 1861', Journal of the Statistical Society of London, 23 (1860), 222–223.
'Recommendations of the Council of the Statistical Society with respect to the census of 1871', Journal of the Statistical Society of London, 33 (1870), 113.
'Report of a committee with reference to the census of 1881', Journal of the Statistical Society of London, 43 (1880), 134–139.
'Report of a committee with reference to the census of 1891', Journal of the Statistical Society of London, 51 (1888), 816–818.
'The census of 1901', Journal of the Royal Statistical Society , 62 (1899), 385–389.
'Report of the census committee 1908–09' Journal of the Royal Statistical Society, 72 (1909), 574–593.