120 THE AMERICAN JOURNAL OF SOCIOLOGY
in the Eleventh Annual Report of the Commissioner of Labor, which, if I read it correctly, reaches the result that in the establishments investigated the women were increasing faster than the men, the boys faster than either, and the girls fastest of all.
On p. 311 Colonel Wright admits the incompleteness of earlier censuses (1850-60), but concludes, nevertheless, that "if the total [of wealth] given for 1850 or i860 should be doubled, the increase is most gratifying." The layman who notices this concession at all may be suspicious enough to ask : " If Colonel Wright thinks it might be in the interest of fairness to multiply the earlier figures by two, how are we to know that three or four would not be the fairest multiplier?" But taking Colonel Wright's own estimate of allowance, and a brief use of the pencil shows that in his judgment a very moderate apparent increase is " most gratifying." If the figures represent only half the total wealth for i860, the increase was from $1,028 per capita in i860 to $1,036 in 1890, or only %%. Moreover, we cannot suppress the surmise that if the conventional values were equalized in the two schedules, even the $8 per head might vanish. In other words, we find so many incomparable tables in the different censuses that our faith in them as scientific demonstrations of anything tends toward the zero point.
Of all the perplexing conclusions drawn or indorsed by Colonel Wright none are more confusing to the layman than those based on the Aldrich report (pp. 228 f.). The author concedes that the report has faults, but to the uninitiated the faults of the report itself are venial compared with the faults of experts who build houses of inter- pretation upon the sands of the Aldrich statistics.
The mass of wage returns may be too much for non-professional intelligence, but to the uninstructed it seems very remarkable to argue from the series that show the greatest increase from i860 to 1891, instead of showing the average increase, or of admitting that the evidence does not permit demonstration of the average.
The inadequacy of the evidence appears in a case like this (p. 230): " In a well-known establishment in the state of Connecticut compos- itors who worked by the day received in 1840 $1.50; in i860, $2; in 1866, from $2.50 to $3, and the same in 1891." The Aldrich report contains data for four printing establishments, but one of them in the state of Connecticut (Rep., Pt. 3, pp. 330-94). In this establishment no data are given for 1840 or i860. Colonel Wright seems to have used for those dates the quotations for 1842 and 1857, respectively
