PRESIDENTIAL ADDRESS. 631 



group must comprise all workers in metals and veterinary surgeons together. 

 The old classification, so far as it is retained, still leads to curious anomalies. 

 The word ' postman ' does not occur even in the detailed list of occupations, and 

 it is a doubtful question whether a telegraph-boy is not considered as occupied in 

 * The Government of the Country,' while the post-office clerk is engaged in the 

 ' Conveyance of Men, Goods, and Messages.' 



It is a sad reflection that, while so much care and labour are spent in 

 accumulating and printing statistical tables, so few of them are of any real 

 importance, and so few are intelligible, even to one who studies them carefully. 

 This topic was handled so ably by Professor Mandello in London last year ' that 

 it is only its great and immediate practical importance that leads me to refer 

 to it. "We need a central thinking department in statistics. There is already 

 collected by the various Government departments, partly in their routine work, 

 partly for the dissemination of information, an immense amount of valuable facts: 

 but whenever a scientific inquirer endeavours to describe accurately some social 

 or industrial development, or wishes to bring to the test of statistics the eft'ect of 

 some proposed reform, whether in taxation, regularity of employment, care of 

 young children, or whatever it may be, some essential in''ormation is found 

 lacking, for the reason that it has been no one's business to collect it. The details 

 of returns of income remain uncodified in the offices of the local surveyors. 

 Baxter's first estimate of — -or, rather, guess at — the amount of income below the 

 exemption limit still holds the field, for no inquiry has ever been made, and we 

 continue in our ignorance of the aggregate national income and of its distribution. 

 We have no adequate knowledge of the age, physical condition, or former occupa- 

 tion of the persons who receive public relief. Illustrations such as these could be 

 multiplied by everyone who has tried to use official statistics. Owing to the enterprise 

 of the Board of Trade, we are to have a second Wage Census and an Industrial 

 Census, and thus many important gaps will be filled in ; but there are as yet no 

 signs of the consideration of the general question of what statistical measurements 

 of the wealth, industry, occupations, and physical condition of the nation should 

 and can successfully be undertaken. 



Official publications have in general been restricted to arithmetical statistics, 

 except in the case of the Table of Survivors in the Census reports, and quite 

 recently to the measurement of the significance between the death-rates of 

 diflferent occupations in the Eegistrar-General's report on Scotland. The official 

 view appears to be, quite correctly, that nothing should be published under the 

 sanction of Government which is not an ascertained fact; but the briefest study is 

 sufficient to show that the very nature of the measurements which have to be made, 

 if only because of the necessary arbitrariness of definition, precludes exactness. The 

 result is that the official counting is numerically correct, but the things counted 

 are not coextensive with the quantity that the scientific inquirer needs to measure. 

 He is left in the position of a man who inquires a distance in France, and is told 

 that it is 8'543 kilometres along the high-road, and then some way along a path; 

 the precision of the first measurement is useless to him. It must be recognised 

 that most statistics are necessarily approximate ; and just as in other scientific 

 measurements the quantity is given as correct to so many significant figures, so in 

 statistics the possible and probable limits of error should be estimated, and the 

 false show of so-called mathematical accuracy given up. 



It is in this direction that the application of mathematical methods is necessary; 

 but before dealing with them I wish to consider the provision made for the supply 

 of persons capable of dealing with statistics by scientific methods. 



I have made inquiries at the principal universities of the United Kingdom, 

 with a view to ascertaining what facilities were aflbrded for the study of statistics, 

 whether arithmetical or mathematical. A knowledge of the statistics of trade 

 and some acquaintance with the main sources and ordinary non-mathematical 

 methods of statistics is required in the Faculties of Commerce in Manchester and 



' Pee Joufnal of the lioijal tStati^tipal Sitcieti/, 1005, pp. 725 neq. 



