National Resources Committee 



outside group of citizens and also the Federal Govern- 

 ment, their research has both major purposes. 



DiCferences in Definition of "Research" 



To estimate the cost of research in a Federal agency 

 requires, first, a demarcation of activities that may be 

 designated as "research," and, secondly, an allocation, 

 usually somewhat arbitrary, of expenditures for such 

 activities. One diiBculty in making comparisons be- 

 tween agencies is, therefore, the lack of definiteness in 

 the term "research" and differences in its application 

 to the work of the respective organizations. For this 

 study of research expenditures, the following explana- 

 tion of the definition adopted by the Committee has 

 been used in letters to agencies : 



"Research" has been defined to comprise investigations in 

 both the natural and social sciences, and their applications, 

 including the collection, compilation, and analysis of statistical, 

 mapping, and other data that will probably result in new knowl- 

 edge of wider usefulness than aid in one administrative deci- 

 sion applying to a single case. By this definition, for example, 

 "research" would include the work of a statistical division col- 

 lecting and compiling operating data on corporations reporting 

 to a Federal regulatory agency, and also a study by the Navy 

 of the resistance of armor plating to high explosives, since 

 knowledge so acquired might be applicable to uses other than 

 the immediate administrative decisions for which the work 

 was initiated. 



Organizational lines that are roughly coterminous with re- 

 search activity have been respected in compiling the expendi- 

 tures data, so that the entire work of an organization has been 

 considered "research" if nonresearch duties are comparatively 

 few and usually incidental. All expenses connected with re- 

 .search inquiries such as clerical assistance, printing, transporta- 

 tion, etc.. and an allocated amount of general administrative 

 overhead within the bureaus and independent agencies have 

 been included. 



For purposes of this survey "research" excludes, in addition 

 to the more obvious nonresearch activities, testing, regulation, 

 promotion, informational service, education, legal counsel, ac- 

 counting, and routine administration. However, such activities 

 are necessarily included so far as research is carried on as an 

 aid to more eflBcient accomplishment of such purposes, or where 

 some of the.se non-research operations are carried on as inci- 

 dental aids to research investigations in the same bureau or 

 independent agency. 



No unanimity was found among experienced research 

 directors as to the kinds of work that should be in- 

 cluded as "research" in this study. For example, some 

 scientists would exclude the reports of both the Bureau 

 of the Cen.sus and the Weather Bureau; while on the 

 other hand a classification by a Government agency 

 about 15 years ago listed these two bureaus as the only 

 Government agencies whose entire/ work was "research." 

 Wliile preparing data for this study, the chief of a 

 technical oiRce reduced his estimate of research ex- 

 penditures from $140,000 to $25,000 after a reconsidera- 

 tion of the application of the identical definition to the 

 existing activities of his oiBoe, 



The principal conflicts of opinion about the defini- 

 tion used in this study have revolved around the in- 

 clusion of the following activities as "research": 



1. The collection and tabulation of basic data, especially in 

 the social sciences; where, in contrast with the natural sci- 

 ences, analysis is usually carried on separately from collection 

 and initial preparation of the material. 



2. Economic and social studies of all kinds. 



3. Mapping and surveying. 



4. Aids to research, such as library and archival services. 



5. Highly specialized and secret techniques for military 

 purposes. 



6. Studies, otherwise clearly "research," that are specifi- 

 cally required by an act of Congress under another name. 



7. Studies not specifically enumerated in an act of Congress 

 but necessary to carry on authorized work efficiently. 



In an agency where research had been almost elim- 

 inatetl in the previous year by a cut in the appropria- 

 tion originally requested, there was a noticeable effort 

 to define "research" as narrowly as possible. On the 

 other hand, the attitude of most budget officers toward 

 the research function and its definition was aquiescent 

 and understanding, if not sympathetic. One experi- 

 enced administrative officer, not himself engaged in 

 research, suggested the following definition of "re- 

 search," which has not been improved by inclusion of 

 more detail : "Research is the attempt to know the 

 facts in order to do our work and serve the public 

 better." 



Part of tlie difficulty with the adopted definition of 

 "research" is due to attempts to distinguish between 

 what might be designated as the "higher" and "lower" 

 orders of research without admitting the use of those 

 concepts. Research ability of the "higher" order is 

 required to devise and improve the methods for pre- 

 paring compilations of data in useful form, and to 

 analyze the results obtained from such compilations. 

 A lower order of research ability, and sometimes al- 

 most none, is required in the personnel carrying on 

 the routine collection of information and its tabulation 

 and presentation according to general specifications. 

 However, both types of activity are "research" within 

 the definition adopted for use in this report. For the 

 effective collection of data, research of the higher or- 

 der must necessarily precede and follow the routine 

 task, and in most cases planning and analysis becomes 

 an integral and continuing part of the whole operation. 



In connection with the study of research expendi- 

 tures, it would probably be instructive to obtain sepa- 

 rate estimates for these two "orders" of research, 

 especially in the case of economic or sociological statis- 

 tics, which furnish the bases for most analytical 

 research in the social sciences. However, such a sepa- 

 ration has proven impracticable because of the budg- 

 etary indivisibility of the two types of research 

 processes. Wliile analysis of compiled data is often 



