ECCENTRIC OFFICIAL STATISTICS 83
in fact of anything else save that a less complete return of the rural districts in 1880, and the omission of certain large establishments in 1890, in which it is notorious that the lowest wages are paid, offset the fuller enumeration of the higher-paid hand trades, and in no wise affected the comparability of the wage statistics."
Colonel Wright seems to have been possessed with the same idea that out of this comedy of errors the truth might be evolved ; for in a letter to the writer dated May 3, 1894, he says: "The canvass of the principal cities was undoubtedly more thorough than at 1880, but, on the other hand, it is believed the canvass of the rural districts was more complete at 1880 than at 1890. It, therefore, cannot be said that the entire canvass of 1890 was more complete than at 1880. The change in the form of inquiry, considered by itself, cannot be consid- ered as abnormally increasing the product. The product reported for 1890 is comparable with an exactly similar amount for 1880." Regarding this idea of balancing one error with another. General Walker remarks on p. xxiii of the Ninth Census : " Every error that occurs in the census of a country, or in any statistical result what- ever, is to be regretted as an independent evil, hardly less when it balances another error than when it exaggerates the amount of error already existing. Two wrongs no more make a right in mathematics than in morals, and a falsehood in figures is none the more to be tolerated or excused because it may serve to conceal another false- hood."
It seems never to have occurred to our astute census officials that if the enumeration of the rural districts in 1890 was so deficient as to offset the more complete enumeration of the cities and make the total
■ As to the comparability of the data, Mr. Steuart remarks in the Census (Compen- dium, Pt. II, p. 704) : " No previous census of the United States obtained so complete reports regarding such trades as masonry, carpentering, blacksmithing, cooperage, painting, plumbing, and similar trades using machinery to a limited extent.
" Previous census inquiries omitted the following industries : Bottling ; cars and general shop construction and repairs by steam railroad companies ; china decorating ; clothing, women's dress-making ; coffins and burial cases, trimming and finishing ; cotton, cleaning and rehandling; cotton ginning; cotton waste; drug grinding; druggists' preparations, not including prescriptions; gas, illuminating and heating; hay and straw, baling ; millinery, custom work ; petroleum refining (petroleum refining formed part of a separate report at the census of 1880, and the statistics were not included in the report on manufactures). The inclusion of these industries in the Eleventh Census, together with the changes referred to, renders it impracticable to use the data for 1880 as a basis for comparison without carefully taking these facts into consideration."
