AUSTRALIAN MORALITY. 47S' 



and, secondarily, a due discrimination and allowance for those 

 factors bearing on the question which cannot be reduced to statis- 

 tical form. If it is held that the last-mentioned factors over- 

 shadow in importance those that can be expressed statistically, the 

 question resolves itself into a solvitiir ambiilando. 



In regard to the plea for a sufficiency of statistical data, it is 

 of course quite common to meet the trite rejoinder that " statistics 

 can be made to prove anything," or perhaps the old saw about 

 " lies, awful lies and statistics " will be trumpeted forth with much 

 gusto. Strangely enough, these peculiar ideas on the subject of 

 statistics find such comparatively wide acceptance that it is 

 necessary to emphasise that they are based, consciously or uncon- 

 sciously, on one of two assumptions. These are, first, that the 

 statistics are built up from unreliable data, or, second, that granted 

 the accuracy of the data, actual figures can be advanced from the 

 same set of tabulations to support either the pro or the contra of 

 an argument. The latter ingenious misuse of statistics was 

 fairly common in the old Freetrade v. Protection days in New 

 South Wales, when each side " selected " figures which seemed 

 quite admirably to bolster up its arguments. In many cases the 

 disputants were unaware that this was an illegitimate use of 

 statistics ; but if they were aware, they managed to conceal the 

 fact very well from their hearers. Legitimate and intelligent use 

 of statistics imphes not only familiarity with the methods of 

 collection, compilation and presentation of the units, but pre- 

 supposes also the faculty of recognising and making due allow- 

 ance for those elements bearing on the investigation which cannot 

 be directly embodied in tlie tabulations, it implies, also, a 

 recognition of the principle that results for single years cannot fairly 

 be taken to prove or disprove a tendency which may require at 

 least five years for its proper manifestation. One might at this 

 juncture allude also to a pitfall which frequently entraps persons 

 who might naturally claim to possess some statistical facility. 

 This is the fallacy of mistaking cases of parallelism for cause and 

 effect. For example, some years ago, in connection with an 

 investigation into the decline of the birth-rate in New South Wales, 

 one observer produced a set of statistical tabulations shewing 

 that the decline in the birth-rate was accompanied by an increase 

 in the insanity rate, thereby suggesting a causal relation, which 

 analysis proves to be quite unwarranted. 



Sufficient has perhaps been advanced to shew that in dealing 

 with the question of the improvement or otherwise in the moral 

 condition of Australia there is urgent necessity for the dispersal, 

 by the clear, cold light of statistics, of the mists of obscurity in 

 which the problem has been enveloped through hasty generalisa- 

 tion and ex cathedra assertion. 



Space will permit of the detailed examination of two only 

 of the possible elements which bear on the problem. These are 

 the statistics of crime and the statistics of natality, and they have 



