SECTION 7, 



PROBLEMS OF THE BUREAU OF THE CENSUS IN THEIR 

 RELATION TO SOCIAL SCIENCE 



By Samuel A. Stouffer 



INTRODUCTION 



The purpose of this report is to illustrate concretely 

 some of the practical problems faced by a Government 

 agency responsible for furnishing social science data 

 and doing social science research. 



The illustrations are taken from the work of a single 

 agency, the Bureau of the Census. Therefore, this 

 document may be considered as a case history. Such 

 a case history is necessary if one is to see vividly and 

 nitimately how certain problems actually arise. But 

 such a case history lacks generality miless the single 

 agency studied is typical. Tiie Bureau of the Census, 

 partly by virtue of its central importance, is not a 

 typical agency. It has several more or less nnique 

 characteristics. Some of its problems are common to 

 many agencies, but many of its problems arise histori- 

 cally out of a unique functional situation. It is not 

 always possible to make a clean-cut separation of the 

 common and the unique problems. 



Three major subjects are discussed: 



(1) Consumer needs and consumer support. 



(2) Selection, training, and promotion of personnel. 



(3) Analytical research on the raw data. 



For purposes of orientation, the discussion of these 

 tliree topics is preceded by a brief background sketch. 

 This is particularly necessary in order to indicate how 

 the Bureau of the Census inlierited certain problems 

 which may differ from those in other research agencies. 



The record of the Bureau of the Census is one in 

 which all Americans may take pride. The leadership 

 of the United States in the development of objectivity 

 in the social sciences would have been impossible with- 

 out the quantity of economic, demographic, and social 

 data i^rovided by this Government agency. The Bu- 

 reau, like various other organizations in the Govern- 

 ment, has had its vicissitudes. It is now in process 

 of rebuilding under the leadership of its director, Mr. 

 William L. Austin. Even though many of its prob- 

 lems have their unique aspects, an examination of these 

 problems hardly can fail to be instructive to those 

 seeking an undertanding of the conditions under which 

 work so fundamental to social science is carried on. 

 Nothing in this report must be construed as represent- 

 ing, necessarily, the official position of the Bureau. 

 The writer accepts full responsibility for statements 

 of fact or opinion. 



I. BACKGROUND 



The role of the Federal Government in research in 

 the social sciences dilfers from its role in natural science 

 research. The difference lies in the fact that social 

 sciences depend on the Go^'ernment for the collection 

 and tabidation of much of their basic data. 



From the decennial census and other statistical com- 

 pilations by the Government come many, if not most, 

 of the raw materials used in economics, sociology, and 

 related disciplines. While Government agencies do a 

 large amount of research with these data, just as they 

 do a large amount of research with natural science 

 data, the direct and primary provision of the original 

 statistical data is a governmental contribution to social 

 science research transcending its relative contribution 

 to other scientific fields. 



There are about 90 different agencies in the Federal 

 Government which collect or compile statistics.^ Of 



•The Committee on Government Statistics listed 02 agencies as of 

 about January 1936. See Committee's Report, Social Science Researcli 

 Council Bulletin 26, Appendix A. 



these agencies the Bureau of the Census, though second 

 in size of full-time staff,- is preeminent, because of 

 the fundamental and general-purpose character of its 

 AVork and because of the magnitude of its periodic 

 enumerations. The Bureau's regular budget in a 

 normal intercensal year is nearly 2 million dollars. 

 The 1930 census cost about 40 million. 



There are two main respects, apart from mere size, 

 in which the Bureau of the Census differs from other 

 statistical agencies: 



1. The permanent staff of the Bureau, like a miniature 

 standing army, is the nucleus of great emergency organizations 

 recruited quinquennially to take the Census of Agriculture 

 and decennially to take the Census of Population. At the 

 point of its greatest expansion, the Bureau has more than 

 125,000 employees. Moreover, within the intercensal periods 

 its work load fluctuates greatly, with the consequence that the 

 average ratio of temporary to permanent employees is large 

 compared with that in other agencies. 



- On July 1, 1938, the Bureau of the Census had 104S full-time em- 

 ployees, as compared with 3.358 in the Bureau of Agricultural Economics. 



197 



