50 



National Resources Committee 



collection of data national in scope is increasingly 

 recognized as a function of the Federal Government. — 

 There are, it is true, agencies outside the Federal 

 Government which collect and analyze data national in 

 scope. Tiie American Federation of Labor and the 

 National Industrial Conference Board, to use only two 

 examples, have data on employment and unemploy- 

 ment. The latter has data on national income. Other 

 like examples can be cited, but it is at once evident that 

 the collection of all kinds of census data is so compre- 

 hensive ai! undertaking that only the central Govern- 

 ment commands the resources and the authority ade- 

 quate to tiie performance of the task. Wlien the census 

 was taken iji the early years of the national history of 

 the United States, its purpose was solely to supply the 

 facts necessary for the adjustment of representation in 

 the lower House of the Congress. Gradually the 

 fimction of collecting data has expanded until today 

 some phase of tliis activity is an important part of 

 the work of such agencies as the Department of 

 Agriculture, the Department of Labor, the Social Se- 

 curity Board, the Bureau of Internal Eevenuc, the 

 Office of Education, and the Bureau of Foreign and 

 Domestic Commerce. 



It has frequently been assumed that a great deal of 

 duplication existed in the work of the various census- 

 taking agencies and that consultation among the vari- 

 ous agencies would reduce this duplication. It is true 

 that some cases of duplication have been found and 

 dealt with. In the main, however, waste due to dupli- 

 cations is insignificant. The advantages which result 

 from cross checking and the necessity in a rapidly mov- 

 ing civilization of securing census figures more fre- 

 quently than once in 10 years fully justify the collection 

 of statistics by more than one agency. 



Examples of the importance of securing data at short 

 intervals are to be found in the information which the 

 Bureau of Labor Statistics collects on such topics as 

 "trends of employment and pay rolls," "hourly earn- 

 ings, average per-capita weekly earnings, and average 

 hours of work," "labor turn-over," and "building per- 

 mits"; the estimates made of the national income by 

 the Bureau of Foreign and Domestic Commerce; the 

 various educational statistics collected by the Office of 

 Education; and the statistical materials gathered by 

 the Bureau of Mines. 



In many of the cases where information is collected 

 at short intervals the sampling method is employed. 

 The accuracy of the particular sample canvassed is best 

 determined by checking from time to time the sample 

 figures with either over-all census figures or other sam- 

 pling data collected in related fields. 



There is one general problem which arises in con- 

 nection with census taking which is repeatedly brought 



up in the statements supplied by Federal agencies. The 

 form in which this problem is commonly presented is 

 a plea for expansion of the facilities for analysis and 

 interpretation of census data. One of the leading stat- 

 isticians in the employ of the Government put the 

 problem in these terms : "Our Bureau is not interested 

 in mere figures; it is interested in people. I was ap- 

 pointed becau.se of my insistence on interpretation of 

 all statistical data. If we were not to analyze and 

 interpret the statistics which we gather the chiefs of 

 my staff would not be here. Most of them are 

 economists." 



On the other hand, it is the expressed belief of cer- 

 tain leading social scientists outside the Government 

 that the chief function of Federal agencies is to collect 

 statistical data. These agencies, it is contended, should 

 perfonn only a very limited analj^sis of the data col- 

 lected, and should leave the interpretation of the data 

 to individuals who are wholly detached from policy- 

 making authorities. The reason for the position thus 

 taken is that association with policy-making authority 

 is believed to bias analysis and interpretation. 



It is generally admitted even by those who object 

 to extended analyses and interpretations that the re- 

 finement of the data collected requires a certain degree 

 of analysis. The difficulty of drawing a sharp line 

 is evident. There can be very little doubt that, what- 

 ever the academic distinctions urged on one side or the 

 other of this discussion, the Federal agencies should go 

 as far in analysis and interpretation as their energy 

 and resources permit. An examination of the publica- 

 tions of the Bureau of Labor Statistics and the Bureau 

 of Agricultural Economics leaves no doubt on this 

 issue. 



Authority to Secure Information 



The Federal Govenwient has authority, and on occa- 

 sion exercises its authority, to collect information which 

 i<^ iruiccessible to nongovernmental agencies. — Striking 

 illustrations of the exercise of authority in the collec- 

 tion of information are found in the researches of the 

 Federal Trade Commission, the Interstate Commerce 

 Commission, and the Bureau of the Census. Each of 

 these agencies is authorized by law to demand infor- 

 mation. It is stated by administrators and investiga- 

 tors that in most cases there is no occasion to exercise 

 authority. For example, the industrial and commercial 

 corporations with which the Federal Trade Commis- 

 sion has dealings have in general been willing to open 

 their records to inspection without hesitation. 

 Wliether this willingness is due to interest in the out- 

 comes of investigation or to knowledge that authority 

 exists is perhaps an impossible question to answer. 

 Some governmental agencies report that where they 

 have depended on cooperation rather than compulsion 



