220 



National Resources Committee 



as that of economic advisei- to the Secretary of State. 

 Usuall}-, such quick interpretative analyses are not pub- 

 lished, but are presented to administrators verbally or 

 iii the form of memoranda, often confidential. The 

 I'tfectiveness of the analyses depends on the analyst's 

 <irasp of the immediate administrative issues involved, 

 on his judgment in separating the relevant from the 

 irrelevant facts, and on his personal rapport with the 

 administrator. We can, therefore, safely disregard this 

 type of quick analytical research in the present report 

 on a nonadministrative agency. 



Should the Bureau Do Research 

 of the Second Type? 



If it be agreed that the first type of analytical work 

 listed above, namely, investigations of the accuracy 

 of the basic data, is clearly a function of the Bureau 

 of the Census, and that the third type, namely, quick 

 interpretative analysis of a particular administrative 

 I)roblem, is generally not the function of the Bureau, 

 the ground is cleared for a consideration of the second 

 and possibly most debatable activity, namely: 



Research on the social and economic significance of 

 relationships derived from basic tabulations, with the 

 general aim of making additions to knowledge. 



There is, perhaps, no sharp line of distinction be- 

 tween research described under this category and the 

 mere collection and presentation of basic data in tabu- 

 lar form. Selection of categories for cross tabulation 

 in the ordinary basic census table is presumably a 

 recognition of the social or economic significance of 

 die relationships. But there are various levels or de- 

 grees of I'efinement of the new data. There is a dis- 

 tinction, in emphasis at least, between a census mono- 

 graph such as that by Edmund E. Day and Woodlief 

 Thomas on The Growth of Manufactures and the 

 original cross tabulations in a regular census volume, 

 [ii such a monograph an effort is made to bring to- 

 irether from various general-purpose tabulations se- 

 lected data which, summarized, describe a particular 

 sequence of events or aid in verifying an explicit hy- 

 ()othesis. The main end is that of telling the historical 

 story or formulating and verifying hypotheses, and 

 llie statistics are used as means to that end. 



Practice Heretofore Followed 



The Bureau of the Census has produced a remark- 

 able body of research of this character. Indeed, when 

 one goes back a lialf century to the regular census vol- 

 umes of 1880 and 1890, before the establishment of a 

 permanent Bureau, one is astonished at both the qual- 

 ity and volume of analytical reseai'ch in a period when 

 the social sciences in the universities were still branches 

 of philosophy, with little interest in data. The intro- 



ductions to the regular census \oluiues in 1900 and 

 1910 continued this distinguished tradition. The vol- 

 ume edited by Dr. Walter F. Willcox, Supplementary 

 A7ialysis and Derivative Tables, containing over 1,100 

 pages of analysis of the 1900 population data, perhaps 

 marks the peak of accomplishment. 



Thei'e were two serious objections to a continuance 

 of the tradition of publishing elaborate analytical in- 

 troductions to the routine tabulations. The most seri- 

 ous objection was the delay caused in the publication 

 of the basic data. The other objection was the cost. 

 In view of the wastes due to the character of the field 

 work (Carroll D. Wright estunated it at $2,000,000 in 

 1890),^' the objection to scientific analysis on the 

 gi'ound of cost does not make an impressive argument. 

 The first objection has generally been recognized as 

 valid. The census of 1910 is regarded as one of the 

 best in American history, for which particular credit 

 is due to Dr. E. Dana Durand,^^ the first profession- 

 ally trained social scientist to be appointed Director of 

 the permanent Bureau; but criticism was offered be- 

 cause of delays in publication. That the delays were 

 due in part to the time required for analytical work is 

 doubtless true. 



At any event, it was decided in 1920 to speed up the 

 publication of the regular census volumes by reducing 

 the introductions to a few interpretative pages and to 

 publish detailed analyses of selected subjects in the 

 form of separate census monographs. Thirteen such 

 monographs appeared, the last one coming out in 

 1931, after the next census had been taken. Some of 

 these were prepared by well-known scholars outside of 

 the Government. 



In 1930 the same procedure was plaimed, and publi- 

 cation of the regular census volumes was thcrebj" 

 speeded up. Many of the analytical introductions to 

 the regular volumes, though brief, were able presenta- 

 tions. Twenty-one special monogi-aphs were issued, 

 but many, if not most of these, did not represent ana- 

 lytical research in the stricter sense of the term. They 

 were mainly vehicles for carrying basic tabulations for 

 one reason or another not included in the regular de- 

 cennial publications and they comprised a minimum of 

 explanatory text. A number of important research 

 studies which had been planned were discontinued for 

 lack of funds upon the passage of the Economy Act of 

 1933. Some of the tabulations made for these unfin- 

 ished studies have since been handed over to private 

 research organizations, such as the Scripps Foundation 

 and the Milbank Fund, for analysis and publication. 

 In the opinion of competent critics with whom the 

 writer has spoken, the monographic type of actual ana- 



s' See quotation from Wright in President Theodore Roosevelt's veto 

 message, in part I of this section. 



» Now a member of the Tariff Commission. 



