124 REPORT— 1856. 



When, therefore, in discussing social questions, we apply the statistical test, we 

 are really doing nothing more than appealing from imagination to fact, — from con- 

 jecture to certainty — from an imperfect to a perfect method of observation. In the 

 principle, srrictly speaking, there is no novelty : every sensible and observing man 

 who has lived in a civilized state of society, has been to some extent a statistician ; 

 the novelty, consists, first, in the greater accuracy with which, and the enlarged 

 scale on which, facts can be collected in modern Europe ; and, secondly, in the 

 practical application of that theorj', which to philosophers must, from the analogy 

 of inanimate nature, have always appeared probable — the theory, namely, that 

 organized beings, taken in the aggregate, are governed in their acts by determinate 

 and discoverable laws. 



It is obvious that in a science of this kind, unlike many which have occupied the 

 attention of mankind, little room is left for imagination, and as little for error. On 

 the first ground, the study is unattractive even to many who appreciate its value ; on 

 the second, it is eminently and necessarily progressive. " Hypotheses non fingo," — 

 those memorable words of Newton's — should be written over the door of every 

 Statistical Society in Europe. Nor is there any branch of mental exertion so 

 calculated to promote a cosmopolitan habit of thought and feeling. Man is the 

 object studied ; and man, so studied, is seen to vary in difi^erent countries only in 

 consequence of discoverable influencing causes, such as race, climate, food, laws, 

 modes of life, &c. However great, therefore, the external differences between 

 branches of the human family, the tendency of sociology is to eliminate these 

 differences one by one, to refer each of them to its several specific orgin, and thus, 

 finally, to bring to light the essential unity of type which underlies them all. 



I would also observe, that as an experimental science, the progress of statistics is 

 not liable to those delays which impede the advance of manyother branches of know- 

 ledge. Where, as in mathematics, the work to be done is transacted necessarily and 

 exclusively within the mind of the discoverer, — where not the quantity, but the 

 quality of intellect brought to bear is all-important, — great advances are rare, for 

 the plain reason that they can only be made by men of extraordinary capacities. 

 No number of ordinary proficients in mathematics, working jointly, can make up 

 for the absence, or supply the place, of one Newton. But though not one man in 

 ten thousand can be distinguished as an analyst or a geometer, the number is far 

 larger of those who possess the mental requisites for statistical investigation, at least 

 in its simpler forms : and without disparaging the remarkable talent for arrangement 

 and generalization evinced by such men as Quetelet, and by some of our own country- 

 men whom I will not here mention, it may be safely affirmed that the extension of 

 statistical inquiry depends less on the appearance among us of any one mind of 

 more than common power, than on the sustained and cooperative industry, encouraged 

 by the State, of many minds trained to this pursuit, and each taking a separate and 

 distinct department in which to labour. 



It is almost superfluous to point out the sources of those errors which most 

 beset statisticians. They may I think be reduced under two heads : (1) Calculation 

 of mean results from an insufficient number of data ; a fault, from the effects of which, 

 in finances, many provident societies are suffering grievously: (2) Calculation of 

 mean results without sufficient care being taken to eliminate disturbing causes : 

 whether this omission arises from the classing together of phsenomena essentially 

 distinct, and referable to separate laws, or from omitting to make allowance for 

 imperfections in the data supplied, e.g. as though one engaged on criminal statistics 

 were to assume that all oflfeuces committed were actually brought to light, overlooking 

 those in which no detection follows, and, consequently, in which no trial takes place. 

 Neither of the sources of error which I have mentioned are difficult to avoid. The 

 one danger against which they warn us is that of premature conclusions. In all 

 physical science, but in no science more than this of which we treat, is supension of 

 judgement necessary. I mean by the phrase, that temper of mind which says, " I 

 neither believe nor disbelieve ; evidence is wanting to do either. I only wait and 

 hold myself free from bias until further facts are adduced." How easy this is in 

 theory, — how hard and painful in practice, need not be told to any one who has 

 given time, and thought, and toil to the proof or disproof of a scientific hypothesis. 



Time would not allow me to attempt even the most rapid and hasty survey of 



