Page:American Journal of Sociology Volume 5.djvu/133

 REVIEiVS 119

He may be correct, but the statistical proof that he is correct is not in evidence. Massing the sort of evidence that we have in a way to leave the impression that it amounts to statistical proof is " practical " from some points of view perhaps, but it is neither philosophically nor scientifically nor pedagogically sound. For instance (p. 41): "The number o{ families increased from .... i860 to 1870, 45.45 per cent." Is it not a fact that in the census of i860 the families of three or four million slaves were excluded, while in the census of 1870 they were included ? If so, the proposition is obviously inexact. Again : " One-third of the inmates of prisons coming from one-seventh of the population " (p. 352). For commentary on this reference I merely refer to the article entitled " Immigration and Crime," by Dr. Hastings H. Hart, in the American Journal of Sociology, for November, 1896.

On p. 37 Colonel Wright says : "The average age of the popula- tion of the United States in 1890 was .... 25.11 years, as compared with 24.13, in 1880 This rise of average age of the living popu- lation has long been going on." Perhaps it has, but the United States census is very poor proof. In the one census the age was taken " at nearest birthday," in the other, "at last birthday." Is not this differ- ence in itself sufficient to account for nearly, if not quite, half the apparent increase, and would not immigration and the decrease in the birth-rate explain most of the other half ?

On pp. 253-4 the author claims that the statistics show an increase in the proportion of skilled laborers. The evidence cited is the census of i860 in comparison with that of 1890. Is not the rea- soning made utterly fallacious by the inclusion in 1890 of (say) 3,000,000 descendants of slaves, some 600,000 boys at ages not included in i860, and 2,000,000 more women than were thus accounted for in the earlier census ?

On p. 215 we read: "Not only is the decrease in the number of children [employed] observable in proportion, but in the actual num- ber as well; for in 1870 there were 739,164, and in 1890 only 603,013." Is not this an unwarranted exhibit, from the fact that the larger num- ber at the earlier date included "children under sixteen," while the smaller number at the later date accounted only for those under four- teen and a half ?

On p. 213 we find the cheering conclusion: "There need not be any alarm, therefore, as to the encroachment of women upon the occupations held by men." I cannot reconcile this with the showing