634 PROCEEDINGS OF SECTION G (l.). 



4.— CLASSIFICATION OF DISEASES AND THE CAUSES 

 OF DEATH. 



By W. T. WEEDON. 



5.— THE USE AND SCOPE OF STATISTICS. 



By W. SIEBENHAAR, Compiler of Statistics and Sub-editor of the Year 

 Book, Western Australia. 



(With Three Plates.) 



[Abstract.] 



In spite of the almost universal use of statistics there still appears 

 to exist a great amount of uncertainty about its real nature. When 

 an authority like Dr. Mayo-Smith refers to " the question whether 

 statistics is a science like sociology, or is only a scientific method of 

 investigation like the science of microscopy," without showing any 

 inclination to settle the point, we may be pardoned for concluding 

 that the conceptions as to the scope of this science or scientific method 

 must be somewhat vague. On further investigation we find that the 

 same uncertainty is evidenced in the definitions given of statistics, 

 and in the material published in various quarters as statistical. For 

 instance, the above-named authority says : — " Statistics consists in 

 the observation of phenomena which can be counted or expressed in 

 figures. It gives us the quantitative measurements of social phenomena 

 which are required for the analysis of social organisation." On the 

 other hand the definition found in Webster, no doubt also furnished 

 by an authority, reads as follows : — " The science which has to do 

 with the collection and classification of certain facts respecting the con- 

 dition of the people in a state, especially those facts which can be 

 stated in numbers, or in tables of numbers, or in any tabular and 

 classified arrangement." The condition that the phenomena must 

 be capable of being counted or expressed in figures is here, we see, 

 not rigorously insisted on, though it is mentioned as that of the princi- 

 pal statistical categories. This undoubted evidence of a difierence 

 of opinion is further supported by the various kinds of information 

 contained in statistical publications. Very few of these confine them- 

 selves strictly to phenomena that can be counted or expressed in 

 figures. It is of frequent occurrence, in exclusively statistical com- 

 pilations, that particulars are included which have no direct connec- 

 tion whatever with figures — ordinary information, which it is thought 

 desirable to place on record without even considering whether it is 

 capable of being presented in tabular or classified arrangement. The 

 statistician whose duty it is to compile publications of this nature, 

 either for the use of state or local government, or for some private 

 institution, is constantly faced by the perplexing necessity of having 

 to decide where the line should be drawn between information that 



