Page:American Journal of Sociology Volume 3.djvu/639

 OFFICIAL STATISTICS 625

say that the more thorough canvass, or the inclusion of additional industries, has destroyed the utility of the totals of 1890 for the purposes of comparison with 1880.

It is asserted that the schedule of inquiry used in 1890 was so entirely different from that of 1880 that, if exactly the same establishments had been enumerated at the two censuses, the results could not be compared, and that the totals for 1890 do not show true conditions. If the schedule used at 1890 has resulted in obtaining such a distorted report that it cannot be compared with 1880, it certainly has not secured a true statement of the conditions of industry.

The differences in the schedules used at the two censuses consisted entirely in itemizing the questions contained in the schedule of 1880 and the incorporation of one additional question. The schedule of 1890 was an advance in statistical methods in that it not only developed the full totals, but enabled a presenta- tion that would show all the items of which the totals were com- posed, so that in the future no question could arise as to the intention of including or excluding certain facts.

The questions concerning employes and wages have been the particular object of criticism. The schedule of 1880 called for the average number of employes men, women, and children respectively engaged during the year and the total amount paid as wages. No class of employes was excepted. The evident intention was to secure a complete return of ail classes. The schedule of 1890 required the employes men, women, and children respectively to be reported in five groups : first, skilled workmen ; second, officers and firm members ; third, clerks ; fourth, unskilled workmen, and, fifth, piece workers. There were other subdivisions of the question at both censuses, but nothing foreign to the inquiry of 1880 was added in 1890, the intention being only to secure a complete and full report. As stated in the census reports, "the tendency of the questions used in 1880 was to obtain a number in excess of the average number of employes, while it is believed the questions used in 1890 obtained the average number. The questions in 1890