PROCEEDINGS OF SECTION G (l.). 639 



lengthy periods, whilst to statistics of scientific value we must un- 

 doubtedly assign a permanent as well as universal importance. 



The responsibility of the statistician is a particularly grave one, 

 since it is well known that figures lend themselves almost as easily to 

 manipulation as do legal enactments. The compilation of statistics 

 should therefore be, perhaps more than any branch of Government 

 activity, free from political influence or interference. Without this 

 there may be a certain amount of temptation for those at the head of 

 a government to make statistical truth subservient to political interests. 

 It might, therefore, appropriately be asked whether it would not be 

 well for government statisticians to be safeguarded, by special pro- 

 visions as to their position, such as are made in the case of an auditor- 

 general or a judge, from any undue influence that may imperil the 

 scientific accuracy of their work. Not only is it possible to present 

 the figures collected in an entirely partial or perverse manner, but 

 estimates are often required, and these in particular lend themselves, 

 if the disposition should exist, to misleading presentation. An example 

 has already been given of the possibility of wrong presentation in the 

 case of the respective degrees of fertility of women of different nation- 

 alities among the Australian population. It would be easy to give 

 instances where similar wrong presentation in other branches of 

 statistical calculation might be used for most improper purposes. 



Less dangerous, perhaps, than manipulation or wrong presenta- 

 tion is false deduction, since here, at least, the correction can be made 

 by anyone who has access to the figures presented. It is a means 

 only too frequently resorted to by those who are not too scrupulous 

 about their method in gaining an object, but it also arises in a great 

 many cases from bias or misunderstanding. 



Finally, it remains to say a few words about the right use of 

 statistics. 



First of all it is necessary to use statistics correctly. In this res- 

 pect it is never possible to be too scrupulous. Slovenly use soon 

 becomes a habit. There is a dangerous euphemism called " adjust- 

 ment," which, if once tolerated in its bad sense, may become a per- 

 nicious practice destructive of all accuracy and reliability. We find, 

 for instance, in many tabular statements the convenient heading 

 " all others," which may be very wrongly used for concealing discrep- 

 ancies. It is, therefore, evidently dangerous in making statistical 

 researches to pass over such a heading as this without a full inquiry 

 into the nature of its detail if at all available. 



The second necessity in using statistics is that of using them 

 impartially. It is often easy to leave out certain features of the infor- 

 mation obtained and to give prominence to others, thereby to create 

 certain impressions which are but half the truth. The information 

 should be so presented as to give a correct and proportionate view of 

 the results obtained. The fact that statistics are at present generally 

 collected by some state or local body is a natural temptation to those 

 working in the interests of such state or body to present its conditions 

 in the most favorable light. This may lead to more or less reprehensible 



