TRANSACTIONS OF SECTION F. 821 



rely on the statistician's art, and that the true advancement of knowledge in what- 

 ever path we take depends quite as much on the avoidance of rash conclusions as 

 on the faculty of quick perception of apparent results. 



There is then this lesson to be learned in the discussions to be held in this 

 Section, and it is one on which, in opening our deliberations, I think I am fairly 

 entitled to insist. Since accurate statistical data are fundamental to sound argu- 

 ment and correct deduction? in any sphere of science, too great care cannot be 

 expended in the task of making sure that figures given to the public are really 

 what they claim to be. Where a comparison is to be made it is our business to 

 see a practical identity in the character of the facts to be observed, and to give 

 such warning as is requisite to guard against the possibility of over-strained and 

 illegitimate use of the data by those into whose hands they may ultimately come. 

 Where a deduction is to be made or a conclusion is to he announced by the original 

 compiler himself, it is well, too, he should remember that a statistical decision 

 should have in it something of judicial deliberation and gravity, and should bo 

 given to the world only after the application of a chastened scepticism and distrust 

 to the testing of the first impressions to which the bare numbers that appear on 

 the surface of any calculation seem to point. 



Lastly, let us not overlook the prescriptive cautions of many past masters in 

 statistical work to distrust big totals and dissect general averages. 



We are all of us familiar with the vastly larger space accorded to statistics in 

 debate in the second half of the dying century, how readily the arbitrament of 

 figures is now appealed to by the politician or the journalist, by the man of 

 science or the philosopher. This very fact, however, constitutes in itself a 

 danger, and I trust, therefore, I may be forgiven if I interpose between the 

 Section and its prepared work by preaching from the Chair -with some insist- 

 ence the somewhat trite doctrine that statistical and economic science has few 

 greater enemies than those who fail to apply the most rigid tests to the sufficiency 

 of the elementary figures on which a theory is to be formed or au administrative 

 act accomplished. Nor, indeed, is a much smaller oSence involved in the over- 

 confident use, whether for international comparisons or for those flights of prophecy 

 in which we all like from time to time to indulge, of figures not in their immediate 

 connection themselves erroneous, but which are, nevertheless, not quite strong 

 enough to bear the strain of the superstructure to be reared, or which are devoid 

 of the essential elements of true comparability of condition. 



It is then alike for the makers and the users of statistics to observe much 

 caution in their own utterances and in the manufacture of those missiles of con- 

 troversy which every table furnisbes, and which in the hasty discussions of our 

 day, when mere rapidity is exalted almost to the place of a virtue, are apt at times 

 to prove dangerous to those who wield them, whether in the press, the lecture- 

 room, or the senate. 



Most of all is it incumhent on one who ventures on the duties of this Chair 

 with none of the opportunities of reflection which many professorial predecessors 

 must have enjoyed, and who comes straight from the daily turmoil of executive 

 work and the discharge of continuous official service, to exercise some reticence in 

 venturing on expressions of individual opinion. In what I may say, therefore, 

 by way of preface to your discussions I would endeavour to confine my remarks 

 to a notice of some of the chief statistical investigations now impending and 

 an account of the difficulties to be encountered by the statistician in his work, 

 illustrating, from the class of subjects with which my work has made me familiar, 

 the sorts of obstacles which hinder the accurate presentation of international 

 comparisons of agi-icultural conditions. 



The entire omission of a sectional address — for which there is, I believe, pre- 

 cedent in your records — or the substitution of a simple speech for a reasoned paper, 

 as was allowed to the distinguished statesman who presided at the last Bradford 

 meeting, on the score of the demands of the State on the services of its servants, 

 might, perhaps, have met my case and relieved me of a duty to which I feel far 

 from equal, and you of listening to my crude remarks. This indulgence has not, 

 however, been accorded, and I must, therefore, crave the pardon of the Section I can 



