364 THE AMERICAN JOURNAL OK SOCIOLOGY
According to his own statement, the number reported in 1890 was the average number while the number reported in 1880 was a number greater than the average number, yet in his Record article he takes the difference between the two numbers as repre- senting the decrease in the number of children employed at the two periods. In the case of children who attend school a part of the year and are employed in establishments that run a full force but a few months in the year, the total number of employes might be more than double the average number. Mr. Waite states that the number reported as the average number in many establishments in 1880 was the average number while the estab- lishment was running a full force and was often several times the average number as figured out by the methods employed in compiling the census of 1890. And that for one-half the estab- lishments the number reported was the greatest number at any one time.
There is also another important factor in this problem entirely overlooked by Colonel Wright. The census reports as adults males over 16 and females over 15 classifying the remainder as children. In ascertaining the age, however, the question asked at the last census called for "age nearest birthday" which would include as children males under 16^, females under 15^. At the preceding census age last birthday was called for, which would include males to their I7th and females to their 1 6th birth- day. There is, therefore, a difference of one-half year in the classification and this half year is one that must have included much the largest proportion of children. It thus seems probable that but for the change in methods a very considerable increase instead of decrease in the number of children engaged in manu- factures would have been shown.
An examination of census tables reveals their unreliability in other respects. We find, for instance, in the tobacco industry, as reported for the city of New York, that in 1890 there were 1295 establishments, with employes as follows : Males, 13,352; females, 6772; children, 164. Yet, according to Colonel Wright's report on strikes and lockouts (1887), twice this
