lo THE JUKES, 



and safest. For these reasons I have given it the preference. In 

 collating its materials it is liable to a fundamental error, that of 

 comparing similar facts which are not identical because they do not 

 occur under similar conditions, as where the frequency of crime 

 among men is compared to that among women, using official rec- 

 ords, when we know that the law and the administration of the law 

 treat women with more leniency than men. But Positive Statistics 

 does not explain the causes or the consequences of facts, therefore 

 conclusions drawn from its figures are inferential and may lead to 

 mistaking coincidences for correlations, as where it is concluded that 

 because criminals show a larger percentage of illiteracy than the 

 average of the community, therefore illiteracy is a cause of crime. 



Conjectural Statistics consists of Political Arithmetic and the 

 Theory of Probabilities. The first is a method of computing esti 

 mates of unknown facts by means of known ones, using the rule of 

 three or other mathematical devices for that purpose. For instance, 

 knowing what proportion of paupers to population there is in one 

 county it assumes that all the other counties in a State have the 

 same ratio of pauperism, while, by actual count, some of them have 

 a higher and some a lower ratio. The Theory of Probabilities is a 

 special application of the above to calculate the chance which a 

 given event has of occurring or not occurring in a given number of 

 times, and requires a profound knowledge of mathematics. Life 

 insurance is based on the probability that, say out of a thousand 

 persons born, a given number will die within one, two or more 

 years. But it cannot tell us which person will die at any time, 

 although it can tell us how many will survive after any given 

 term of years. Its essential process is to reduce all facts to an 

 average, and in doing so it substitutes an abstract mathematical 

 entity of uniform quality and degree in place of the actual concrete 

 facts. The dangers of this method are that facts of the same nature, 

 but differing in intensity, are classed together when their effects are 

 not distributive. Thus heat, which, at one degree warms, at another 

 withers, at another devastates, produces at each extreme effects 

 which are diametrically opposite, but which are nevertheless made 

 compensative by reduction to an average and appear as if they were 



