The Horace H. Rackham School of Graduate Studies~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~::

PROCEEDINGS of the FOURTH ANNUAL CONFERENCE ON THE ADIMINISTRATION OF RESEARCH UNIVERSITY OF MICHIGAN September 1 1-13, 1950 ENGINEERING RESEARCH INSTITUTE -UNIVERSITY OF MICHIGAN

FOREWORD The first Conference on Administration of Research was held at The Pennsylvania State College School of Engineering, State College, Pennsylvania, on October 6 and 7, 1947. It was initiated by a number of men who, not previously accustomed to administering large organized research groups, had been placed in charge of new research laboratories established during World War II. Some of these men, in talking with others in a similar situation, conceived the idea that an exchange of information regarding the procedures and practices which they had followed in the administration of their laboratories might be of benefit to all concerned. The conference proved to be such a success that it was decided to hold another the following year. This, in turn, led to the idea of an annual conference. The second and third conferences were also held at The Pennsylvania State College. The organizing group, which was made up of representatives from industrial, governmental, and educational institutions, formed the Advisory Committee, the only formal organization of the conference. Participants in the conference were invited on the basis of experience in the administration of research. This resulted in a profitable exchange of facts and ideas. The Advisory Committee felt that the conferences would be of even greater service if they were held at different institutions. However, in order to safeguard against any unforeseeable adverse circumstances during the infancy of the project, it was proposed to hold both the fourth and fifth conferences at the same place. The choice fell upon the University of Michigan for the conferences of 1950 and 1951, whereafter it is expected they will be at a different institution each year. The 1951 conference will be held September 24, 25, and 26. For carrying out the actual arrangements, preparing the program, etc., an executive committee composed of members of the Advisory Committee is chosen each year. To have someone in charge of local arrangements, a representative of the host school is appointed to the Executive Committee. The proceedings of the fourth conference, edited and published by the host institution (as were those of the first three conferences), are herewith presented as a record of the official meetings. The Engineering Research Institute of the University of Michigan, under whose sponsorship these proceedings were prepared, has had plentiful evidence that the fourth conference was a success. This success was due not only to the excellent program arranged by Dr. Harold K. Work, Chairman of the College of Engineering, New York University, and to the facilities which the University

of Michigan has for holding such conferences, but also to the spirit of those who attended and, above all, to the excellent cooperation of everyone who contributed a paper and took part in the discussions. Those who attended the Invitational Luncheon on the first day of the conference heard the welcome extended by President Ruthven. This welcome expressed the thoughts of all the University people cooperating. We hope that the pleasure of those who attended the meetings equals ours in making and carrying out the necessary arrangements. May next year's conference be even more successful! January, 1951 C. W. Good Assistant Director Engineering Research Institute University of Michigan

CONTENTS FIRST SESSION, Monday, September 11, 9:00 A.M.-H. W. Work, presiding. CALCULATED RISK-Its Place in the Selection, Control, and Termination of Research Projects.......................................................... 1 I. T. H. Vaughn............................................................... 1 Discussion................................................ 9 II. E. D. Reeves................................................................ 10 Discussion.................................................................. 13 III. D. H. Loughridge............................................................. 14 Discussion................................................................... 18 SECOND SESSION, Monday, September 11, 2:30 P.M.-H. M. O'Bryan, presiding. MEASURING THE RETURN FROM RESEARCH................................. 22 I. Allen Abrams................................................................. 22 II. W. S. Parsons................................................................ 24 III. C. G. Suits............................................................ 27 Discussion................................................................... 31 DINNER MEETING, Monday, September 11, 6:45 P.M.-J. P. Adams, presiding. ENGINEERING RESEARCH IN MODERN CORPORATIONS, J. C. Zeder............ 37'J HIRD SESSION, Tuesday, September 12, 9:30 A.M.-M. A. Williamson, presiding........ 41 WHAT IS NEEDED IN A RESEARCH EXECUTIVE I. J. C. Flanagan................................................................ 41 Discussion............................................................. 45 II. A. E. Lombard, Jr............................................................. 48 Discussion................................................................. 51 III. R. D. Stevens.......................................................... 53 Discussion........................................................ 56 FOURTH SESSION, Tuesday, September 12, 2:00 P.M.-A. E. White, presiding............ 58 OVERHEAD-As a Factor in Sponsored Research, W. K. Pierpont................... 58 Discussion................................................................... 63 FIFTH SESSION, Wednesday, September 13, 9:30 A.M.-J. I. Mattill, presiding NEW GOVERNMENT SERVICES TO RESEARCH.......................... 70 I. National Science Foundation, T. J. Killian........................ 70 Discussion....7................................................... 74 II. Building Research, R. U. Ratcliff.............................................. 74 Discussion.................................................................. 78 III. Research for Small Industry, J. C. Green..............................81 SUMMARY OF CONFERENCE E. A. Walker................................................................. 86 ROUND-TABLE DISCUSSIONS............................................. 20, 34, 57, 69 v

COMMITTEES EXECUTIVE COMMITTEE I. C. CRAWFORD, Chairman C. W. GOOD O. C. MAIER GEORGE L. HALLER H. J. MASSON H. P. HAMMOND C. G. WORTHINGTON MAURICE HOLLAND G. H. YOUNG ADVISORY COMMITTEE C. G. WORTHINGTON, Chairman S. L. BASS D. H. LOUGHRIDGE hW. R. BRODE O. C. MAIER D. B. CHAMBERS H. J. MASSON I. C. CRAWFORD H. M. O'BRYAN F. M. DAWSON D. L. PUTT P. D. FOOTE A. M. ROTHROCK GEORGE L. HALLER H. A. SCHADE H. P. HAMMOND N. A. SHEPARD K. L. HOLDERMAN M. H. TRYTTEN MAURICE HOLLAND E. A. WALKER D. B. LANGMUIR A. S. WATERMAN G. H. YOUNG PROGRAM COMMITTEE H. K. WORK, Chairman C. W. GOOD FRED OLSEN H. M. O'BRYAN M. A. WILLIAMSON vi

REGISTRANTS Abrams, Allen Marathon Corp., Rothschild, Wis. Adams, J. P. Provost, University of Michigan Adams, R. R. Battelle Mem. Inst., Columbus, Ohio Anthony, R. N. Bus. Adm. Graduate School, Harvard Univ., Boston 63, RMass. Asbury, W. C. Standard Oil Development Co., New York, N. Y. Barker, E. F. Dept. of Physics, University of Michigan Bartell, F. E. Dept. of Chemistry, University of Michigan Bartholomew, Earl Res. Labs., Ethyl Corp., 8 Mile Rd., Detroit 20, Mich. Beall, P. R. Research and Development Board, Washington, D.C. Beals, R. P. Burroughs Adding Machine Co., Detroit 32, Mich. Begun, S. J. Brush Development Co., 3311 Perkins Ave., Cleveland, Ohio Bennett, W. I. Dean, Coll. of Arch. &8 Design, Univ. of Michigan Binkerd, E. F. Armour and Co., Union Stock Yards, Chicago 9, Ill. Bittenbender, W. A. Merck and Co., Inc., Rahway, N. J. Blake, W. T. Pillsbury Mills, Minneapolis, Mlinn. Blanc, L. A. Caterpillar Tractor Co., Peoria, Ill. Boston, O. W. Metal Processing Dept., Univ. of Michigan Boyce, E. F. Civil Engineering Dept., Univ. of Michigan Boyce, Joseph Argonne National Labs., Box 5207, Chicago 80, Ill. Boyer, Raymond Dow Chemical Co., Midland, Mich. Briggs, R. P. Vice President, University of Michigan Brothers, L. A. Hdqrs., USAF, 5D-866 Pentagon, Washington, D.C. Brown, G. G. Chem. and Met. Eng. Dept., University of Michigan Callender, Arch. B. Glenn L. Martin Co., Baltimore 3, Md. Carpenter, W. H., Jr. Babson Inst. of Bus. Adm., Babson Park, Mass. Carr, A. R. Dean, College of Eng., Wayne University, Detroit, Mich. Champlain, W. P. AF Cambridge Res. Labs., Cambridge, Mass. Cole, R. I. Watson Labs., AMC, Red Bank, N.J. Coles, H. L. Michigan College of Mining and Tech., Houghton, Mich. Cooper, A. H. Bucknell University, Lewisburg, Pa. Crane, N. D. Research and Development Board, Washington, D.C. Crawford, I. C. Dean, College of Eng., University of Michigan Crowe, J. M. Exec. Ed., American Chemical Soc., 1155 16th St., N.W., Washington, D.C. Curry, R. B. Applied Physics Lab., Johns Hopkins University, Silver Spring, Md. Daniels, C. E. Dev. Eng. Div., E. I. DuPont de Nemours and Co., Wilmington, Del. Dannemann, H. F. AF Cambridge Res. Labs., Cambridge, Mass. Davis, Thomas E. Ohio State University Res. Found., Columbus, Ohio Dawson, F. M. Dean, College of Eng., University of Iowa, Iowa City, Iowa Doan, R. L. Phillips Petroleum Co., Bartlesville, Okla. Donaldson, H. C., Jr. Cluett, Peabody and Co., Inc., Troy, N.Y. Dow, Douglas Detroit Testing Lab., 554 Bagley Ave., Detroit 26, Mich. Dow, W. G. Electrical Engineering Department, University of Michigan Drysdale, Taylor USAF, 719 Henry St., Ann Arbor, Mich. Dunn, G. W. Defense Research Board, Ottawa, Ontario, Canada Dyke, H. Gordon RCA Laboratories, Princeton, N.J. VUi

Farrar, P. H. Engineering Research Institute, University of Michigan Fisher, F. R. Res. and Dev., Sinclair Refining Co., Harvey, Ill. Fisher, Lyman C. Underwater Ord. Dept., U. S. Naval Ordnance Lab., White Oak, Md. Flanagan, John C. American Inst. for Res., University of Pittsburgh, Pittsburgh, Pa. Fleming, John G. Res. and Dev. Div., Bristol Co., Waterbury 20, Conn. Fontaine, A. P. Director, Willow Run Research Center, University of Michigan Foote, Paul D. Gulf Research and Development Co., Pittsburgh 21, Pa. Furnas, C. C. Cornell Aeronautical Lab., 4455 Genesee St., Buffalo 21, N.Y. Furth, F. Capt., USN, Naval Research Lab., Washington 20, D.C. Gibson, R. E. Applied Physics Lab., Johns Hopkins University, Silver Spring, Md. Good, C. W. Assistant Dir., Engineering Research Inst., University of Michigan Graham, R. B. Bendix Aviation Corporation, Fisher Bldg., Detroit, Mich. Green, J. C. Office of Technical Services, Dept. of Commerce, Washington, D.C. Green, T. S., Jr. Norton Co., Worcester 6, Mass. Gunness, R. C. Standard Oil Co. (Ind), Chicago 80, Ill. Haller, G. L. Dean, School of Chem. and Physics, Pennsylvania State College, State College, Pa. Hambraeus, Gunnar Tech. Adv., Swedish Embassy, 630 Fifth Ave., New York, N.Y. Hammer, C. F. Eng. Mgr., Westinghouse Air Brake Co., Wilmerding, Pa. Hartley, J. C. Winchester Repeating Arms Co. Div., Olin Indust., Inc., New Haven, Conn. Hawkins, G. A. Eng. Exp. Station, Purdue University, Lafayette, Ind. Hickox, G. H. Eng. Exp. Station, University of Tennessee, Knoxville, Tenn. Holcomb, W. F. Res. Dept., Parke, Davis and Co., Detroit 32, Mich. Holland, Maurice Ind. Res. Advisor, 40 W. 40th St., New York, N.Y. Holmes, L. C. Res. Dept., Stromberg-Carlson Co., Rochester 3, N.Y. Huebner, G. J., Jr. Res. Dept., Chrysler Corp., Box 1919, Detroit 31, Mich. Hughes, E. C. Standard Oil Co. (Ohio), 2127 Cornell Rd., Cleveland, Ohio Hyde, J. W. Coll. of Arch. &- Design, University of Michigan Ikehara, Shikao Tokyo Inst. of Technology, Meguroku, Tokyo, Japan Jacobsen, J. M. Office Govt. Spon. Res., University of Texas, Austin, Texas Jacobsen, Mrs. J. M. Office Govt. Spon. Res., University of Texas, Austin, Texas Jakkula, A. A. Res. Found., A. and M. College of Texas, College Station, Texas Jones, W. N. College of Eng. 8c Science, Carnegie Inst. of Technology, Pittsburgh 13, Pa. Jordan, Louis Div. Eng. 8c Ind. Res., Nat'l Res. Council, 2101 Constitution Ave., Wash., D.C. Keirn, D. J. Col. USA, Res. 8c Dev. Command, Dept. Air Force, Washington, D.C. Kelly, H. C. Sci. Sec., Office of Naval Research, Chicago, Ill. Keniston, Hayward Dean, College of Lit., Sci., g8 the Arts, University of Michigan Kenyon, R. L. Assoc. Ed., Chem. &c Eng. News, 25 E. Jackson Blvd., Chicago, Ill. Killian, T. J. Office of Naval Research, Washington, D.C. Kimball, C. N. Midwest Res. Inst., 4049 Pennsylvania Ave., Kansas City 2, Mo. Klopsteg, P. E. Technological Inst., Northwestern University, Evanston, Ill. Knapp, T. E. Res. Dept., Champion Paper 8c Fibre Co., Hamilton, Ohio Kolstad, G. C. Res. Div., U.S. Atomic Energy Comm., 1901 Constitution Ave., Washington, D.C. Kurt, O. E. Res. Labs., Ethyl Corp., 8 Mile Rd., Detroit, Mich. Larson, C. T. College of Architecture 8c Design, University of Michigan Lawrence, F. I. L. Res. and Dev., Kendall Refining Co., Bradford, Pa. Lehman, Paige Pillsbury Mills, Inc., Minneapolis 2, Minn. Lombard, A. E., Jr. Hdqrs., USAF, Directorate of Res. 8e Dev., 4E-322 Pentagon, Washington, D.C. Loughridge, D. H. U. S. Army, 3E-616 Pentagon, Washington, D.C. viii

Lowance, F. E. Naval C. E. Lab., Pt. Hueneme, Cal. Lyman, C. H. Capt. USA, Bur. Ord., Navy Dept., Washington, D. C. McKean, W. B. Col. MC, Naval Science, University of Michigan Mahin, W. E. Armour Res. Found., Ill. Inst. Tech., Chicago 18, Ill. Maier, O. C. Res. Div., Pullman Standard Car Mfg. Co., Hammond, Ind. Marchetti, John AF Cambridge Res. Lab., Cambridge, Mass. Marlowe, D. E. Naval Ordnance Lab., White Oak, Md. Mattill, J. I. Eng. Coill. Res. Council, Mass. Inst. Tech., Cambridge, Mass. Maxfield, F. A. Bur. Ord., Navy Dept., Washington, D.C. Meid, G. D. National Academy of Sciences, 2201 Constitution Ave., Washington, D.C. Miller, F. L. Esso Laboratories, Elizabeth 13, N.J. Montgomery, W. P. Booz, Allen 8e Hamilton, 135 S. LaSalle St., Chicago 3, Ill. Morgen, R. A. Florida Eng. 8 Ind. Exp. Sta., University of Florida, Gainesville, Fla. Morse, R. S. National Research Corp., Cambridge, Mass. Nelles, Maurice Eng. Exp. Sta., Pennsylvania State College, State College, Pa. Northrup, D. L. Hdqrs., USAF, Off. for Atomic Energy, DCF/O, Att: AFOAT-1, Washington, D.C. O'Brien, R. A. Amer. Soc. of Mech. Engrs., 29 W. 39th St., New York, N.Y. O'Bryan, H. M. Research and Development Board, Washington, D.C. O'Roke, E. C. School of Nat. Resources, University of Michigan Owens, J. S. Ohio State University Res. Found., Columbus 10, Ohio Parsons, W. S. Rear Adm. USN, Weapons Sys. Eval. Group, Off. of Sec. of Def., Washington, D.C. Pennington, J. V. Drilling Res., Inc., 1320 City Nat. Bank Bldg., Houston 2, Texas Perkins, J. A. Asst. Provost, University of Michigan Pettee, G. S. Ops. Research Off., ORD, Ft. McNair, Washington, D.C. Pierpont, W. K. Controller, University of Michigan Poehle, H. F. Engineering Research Institute, University of Michigan Quinby, E. J. Electronic Research & Dev., Monroe Caculating Machine Co., Orange, N. J. Quinsey, W. E. Engineering Research Institute, University of Michigan Ratcliff, R. U. Housing and Home Finance Agency, Washington, D.C. Reeves, E. D. Standard Oil Dev. Co., 15 W. 51st St., New York, N.Y. Reichl, E. H. Res. & Dev. Div., Pittsburgh Cons. Coal Co., Library, Pa. Robertson, R. M. Phys. Sci. Div., Off. of Naval Research., Washington, D. C. Rosselot, D. A. Eng. Exp. Sta., Georgia Inst. of Tech., Atlanta, Ga. Ruthven, A. G. President, University of Michigan Sawyer, R. A. Dean, Graduate School, University of Michigan Schade, H. A. Dept. of Mech. Eng., University of California, Berkeley, Cal. Scheumann, W. W. Cities Service Res. 8e Dev. Co., 70 Pine St., New York, N.Y. Schuh, A. E. Res. & Dev., U. S. Pipe & Foundry Co., Burlington, N.J. Seeger, R. J. Aeroballistic Res. Dept., Naval Ord. Lab., White Oak, Md. Seiler, F. J. Off. of Air Res., AMC, Wright-Patterson AF Base, Dayton, Ohio Selvidge, Harner Spec. Prod. Dev., Bendix Aviation Corp., Detroit, Mich. Seski, E. J. Cook Res. Labs., 1457 W. Diversey Pkwy., Chicago, Ill. Shepard, N. A. American Cyanamid Co., 30 Rockefeller Plaza, New York, N.Y. Shinn, C. E. Electronics Dept., Burroughs Add. Mach. Co., 511 N. Broad St., Philadelphia, Pa. Simon, L. E. Brig. Gen. USA, Ord. Dept., Aberdeen Proving Ground, Md. Smith, R. L. Sinclair Refining Co., Harvey, Ill. Smith, T. R. Res. and Dev., Maytag Co., Newton, Iowa Solberg, A. N. Res. Found., University of Toledo, Toledo 6, Ohio Spencer, R. G. Washington University Res. Found., Clayton 5, Mo. Stauffer, R. A. National Res. Corp., 70 Memorial Dr., Cambridge 42, Mass.

Steding, H. R. Chrysler Corp., Detroit 31, Mich. Steinle, J. V. Res. 8c Dev., S. C. Johnson 8c Son, Inc., Racine, Wis. Stern, Benjamin Col. USA, Signal Corps Eng. Labs., Fort Monmouth, N.J. Stevens, Raymond A. D. Little Co., Cambridge, Mass. Stevenson, R. A. Dean, School of Bus. Administration, University of Michigan Stewart, D. B. Res. Center, B. F. Goodrich Co., Brecksville, Ohio Stewart, Ross Cook Res. Labs., 1457 Diversey Pkwy., Chicago 14, Ill. Suits, C. G. General Electric Res. Labs., Schenectady, N.Y. Suter, C. M. Sterling-Winthrop Res. Inst., Rensselaer, N.Y. Tailman, O. Watson Labs., AMC, Red Bank, N.J. Thompson, A. P. Eagle-Picher Co., Box 290, Joplin, Mo. Tour, Sam Sam Tour and Co., 44 Trinity Pi., New York 6, N.Y. Townsend, J. C. Burroughs Adding Machine Co., 511 N. Broad St., Philadelphia, Pa. Travis, Irven Burroughs Adding Machine Co., 511 N. Broad St., Philadelphia, Pa. Trichel, G. W. Chrysler Corporation, Detroit, Mich. Upton, H. H. Double A Products Co., Manchester, Mich. Vaughan, H. F. School of Public Health, University of Michigan Vaughn, T. H. Wyandotte Chemicals Corp., Wyandotte, Mich. Vincent, H. B. American Structural Products Co., Toledo 1, Ohio Walker, E. A. Research and Development Board, Washington, D.C. Walker, H. S. Detroit Edison Co., Detroit 26, Mich. Waterman, A. T. Office of Naval Research, Washington, D.C. Way, Gordon T. Res. Div., Smith, Kline 8c French Labs., 1530 Spring Garden St., Philadelphia, Pa. Wendt, K. F. Eng. Exp. Sta., University of Wisconsin, Madison, Wis. Wernette, J. P. Dir., Bur. of Bus. Res., University of Michigan White, A. E. Dir., Engineering Research Institute, University of Michigan White, R. H. Maj. USA, Res. and Dev., Detroit Arsenal, Centerline, Mich. Williamson, M. A. Pullman Standard Car Mfg. Co., Hammond, Ind. Woods, Hubert Res. Dept., Portland Cement Association, 33 W. Grand Ave., Chicago 10, Ill. Woolrich, W. R. Dean, College of Eng., University of Texas, Austin 12, Texas Work, H. K. Chairman, College of Eng., New York University, New York 53, N.Y. Worthington, C. G. Ind. Res. Institute, 60 E. 42nd St., New York, N.Y. Young, G. H. Mellon Inst. of Ind. Res., Pittsburgh, Pa. Zarem, A. M. Stanford Res. Inst., Los Angeles 17, Cal. Zeder, J. C. Chrysler Corporation, Box 1919, Detroit 31, Mich. x

First Session Harold K: Work, presiding Director of Research Division College of Engineering, New York University CALCULATED RISK Its Place in the Selection, Control, and Termination of Research Projects -I by THOMAS H. VAUGHN Vice President - Research and Development Wyandotte Chemicals Corporation THERE ARE MANY FACTORS of risk which enter into the sions are increasingly interested in measuring the reselection, control, and termination of a research pro- turn from research. The time to start thinking about ject. Although many of these factors are common to the returns from a program of research is at the beall industrial research, there are a few which are ginning, when you start the various projects which peculiar to the type of research practiced by the comprise the program. chemical industry. Being in a chemical organization While research and development divisions in most and associated with other speakers, who are primarily chemical organizations have many responsibilities concerned with the type of research practiced in the which cannot properly be defined as research and petroleum industry and our military establishments, development, all, or certainly most, are primarily I shall confine myself to considerations of risk with charged with responsibility for the development of respect to research in the chemical industry. the future products of the corporation and, in most It is somewhat difficult to come to grips with a topic instances, are responsible for the maintenance of of this type in so short a time because there are many quality and superiority, under competitive conditions, variations in the types of risk factors and in the degrees of the products in which the corporation is presently of their application at various stages in the life of a interested. In other words, research has a broad reresearch project. Thus, certain factors which may be sponsibility for the future growth and success of the dominant and controlling in connection with the selec- corporation and in some instances has laid down for tion of the project may have little bearing on the con- it by the president or the board of directors the areas trol of the work but may again become important in into which its work must be channeled in order to connection with the termination of the project. be of maximum value. This places the executive in Accordingly, I wish to concentrate on one phase of charge of research in a position where he has some risk calculation which is of overriding importance in general background for the evaluation of ideas which all phases of a research project from its inception to its may be brought up for inclusion in the research profinal disposal, and that is the calculation or estimation gram. of the effect of the project on the health of the organi- There are many general criteria by which the suitazation supporting the work. I use the word "health" bility of a project can be judged. These range from primarily in a financial sense but also as including purely technical evaluations or guesses as to the posother factors, as will be apparent. sibility of completing the research satisfactorily from a All executive groups in industrial organizations technical point of view to very broad business queswho come in contact with the product of research divi- tions having to do with the future health of the child

Calculated Risk 2 of research, should it pass through adolescence and reach maturity. Such general business questions should ECAONOMIC DATA ON CHEMKIAL be of great concern to the man responsible for plan- AND ALLIED COMPANIES FOR 939 ning and conducting the research program. A few E T simple questions of this type are: Is the project one S which will lead to a product which is in our field of A business? Do we have competent manpower and equip- E 19,45, L9 - ment to do the job? Do we have a unique or good, PROFITONSALES 19% 1 raw-material situation? How much investment will it RRNNVEST. 7% E RETURN ON'INVEST. i. require? What kind of a return can we expect on that s Li_ investment? Will the product create new markets or u IATIO OF SALMENT.841 1.4 will it enter into a market which is currently of suffi- T cient stability and expansiveness to absorb the new 9804 1939 FIGURES BASED ON 94 CONFNIS' material? The answer to these and many other similar 207' 1949 FIGURES 8ASEDON 98 COMPANIES Nv 9 questions add up to a composite answer to the ques- T S ARIES ANo SPERCE INWSUTRAL ANO tion:TIs N ENGINEERING CHEMISTRY, 42,9360950) tion: Is it good for the health of our organization? s ENG At the very beginning of a project and frequently L T throughout its life, a project should be subjected to a S 0 searching economic evaluation. To do this requires a _1939 N9 " 19'39 1949 background or framework of reference with which Figure 2 the project may be compared. Fortunately, two arti- contained in this article for the years 1939 and 1949 cles have recently appeared dealing with the financial is indicated in Fig. 2. It will be noted that here too, facts of life of chemical and allied companies. The first when dealing with the allied industries as well as the article, by Pescatello, appeared in Chemical Industries chemical industries, the volume of sales for the year early this year and dealt with twenty-five selected 1949 is larger than the capital invested in the corporachemical companies. tions, with a ratio of roughly 1.4 to 1. The return on investment in 1949 agrees exactly with that indicated 3433 ECONOMIC DATA ON 25 CHEMICAL in the previous article for the chemical industry alone, COMPANIES namely, 10 per cent, and the percentage of profit on'R FOR 1949* 279 sales for the year 1949 was 8 per cent. It is interesting 0 S ALL FIGURES IN MILLIONS OF DOLLARS I to compare these data with those for 1939. Here net v sales are less than capital investment, the ratio being A E e S 0.84 to 1, and the net return after taxes expressed E T S M as return on investment was only 7 per cent as coml pared to 10 per cent in 1949, whereas the profit on S sales was 9 per cent as compared to 8 per cent for 1949. This financial information provides a general background against which research projects in the chemical and allied industries may be evaluated in terms of their 469 AFETER *ESCA possible worth to an organization should they prove NET t290 TAX CHEM. NDo to be technically successful. Of course, similar informaRoFI T I, % tion of interest to the research executive can be obtained for any selected group of companies by consultFigure I ing the annual statements of the companies or by using The economic data shown in Fig. 1 indicates the information available from Standard and Poors, Dunn overall state of health of these twenty-five companies and Bradstreet, and other financial reporting houses. for the year 1949. It will be noted that their gross For the purposes of this discussion, let us set up a sales were somewhat in excess of the invested capital hypothetical corporation which has approximately in the business with a ratio of approximately 1.15 to $100 million invested in its business and which at 1 and that the net profit after taxes (NAT) repre- present has a net sales volume of approximately $110 sented approximately 8.5 per cent of the gross sales million a year. Currently, this company is enjoying volume and a return of 10 per cent on the investment. a net return after taxes of only $8 million a year. This A similar study by Aries and Spence which appeared company is considerably behind the parade in several in the Facts and Figures issue of Industrial and respects: The ratio of sales to the capital invested Engineering Chemistry this year dealt with one hun- is lower than it should be, this ratio being only 1.1 dred chemical and allied companies and gave a rather to 1, whereas the average for the industry as a whole exhaustive survey of the financial facts regarding is 1.4 to 1. The net return on investment is only 8 these companies. A summary of the economic data per cent as compared to a 10 per cent average for the

3 Calculated Risk industry, and the profit based on sales is only 7.3 per research group into effect, assuming of course that they cent against an industry average of 8 per cent. Ob- are positive and turn out as he expects, stating when viously, then, this company has a sincere interest in a the changes in equipment could be made and when, research program which is aimed at projects yielding in his opinion, they should be made. He also provides an unusually high return and would probably be information of an economic nature, indicating the willing to put a large proportion of its research funds savings which would result if the contemplated change into such projects, even though some of them may were made. have considerable risk attached as compared to short- The savings in this particular case fall into three term, sure-payoff jobs which would not materially im- categories: first, savings due to the use of less catalyst, prove the relative position of the company. since the feed rate of the copper powder could be It is naturally the intention of all research admini- more accurately controlled than that of the cupric strators to do their part in having their company chloride and the excess quantities which are being operate with a profit picture that is considerably bet- presently used would not be necessary; second, savings ter than the average for the industry as a whole. It is due to the fact that copper is a lower-priced material obvious that the research administrator in this par- than cupric chloride; and third, savings which would ticular hypothetical company has quite a job on his result from the fact that there would be slightly inhands. creased yield of methyl esterate from the use of copper What I said earlier about the attitude toward short- versus copper chloride. term research in this particular company does not, of This information is reviewed by the director of course, preclude the possibility of its research director research and, after making some necessary changes, being very seriously interested in short-term projects deletions, or additions, is finally turned over to a which show a large return on investment. For ex- draftsman for the preparation of a chart indicating ample, let us consider a completely hypothetical case the salient features involved. This chart would look involving some hypothetical compounds. This corpora- something like that in Fig. 3. It will be noted that tion is manufacturing methyl esterate. The process has given them considerable difficulty and is not show- REPLACEMENT OF CUPRIC CHLORIDE WITH COPPER IN ing the kind of yields hoped for. It has been proposed METHYL ESTERATE PROCESS in one of the research laboratories that the process 100 ESTIMATED could be materially improved and that considerable I ECONOMIC FORECAST TIME TABLE savings could be made if metallic copper were to be Es ALL FIGURES INTHOUSAN | substituted for the present cupric chloride catalyst A ~ OF DOLLARS | used in the process. The research division has spent T. T A to date approximately $2,000 in preliminary investiga- o * M tion and feels that the chances of success are excellent. I F V 4JR Accordingly, the time has come for a decision to be AS D S made as to whether or not additional research effort | 7 NAis to be spent on this project. While the project has $ e 0 - had a rather cursory economic evaluation in the very ADD. -- rKKS. L A beginning, the time has now come for a somewhat S R RCOs A more thorough evaluation. _ t At this stage in our own organization, the supervisor PROJECT EVALUATED JULY 28, 50 in charge of the project would address a memorandum Figure 3 to the director of research of the division in which he would cover the bas i c o b ject ives o..he program, p ro- the estimated total savings resulting from this project would cover the basic objectives of the program, pro- are $87,000 per year, this being broken down into two are $87,000 per year, this being broken down into two vide background information (in a condensed form, parts, one dealing with the increase in yield and the since the director of research is assumed to be cognizant other with the catalyst savings. After income taxes of the main features of the methyl esterate process), (based on the taxes which exist now and not on those s based on the taxes which exist now and not on those provide a research forecast which states the amount which we are going to face in the future), the savings which we are going to face in the future), the savings w hof money whichp has db een spent in the mnonths in would be $48,000. It is estimated that the capital rewhich the program has had part-time attention, cover the present status of the project, estimate how much quired to put the new process in production would be money will be needed to conclude the research phase $100,000. The research supervisor in preparing this of the project, estimate the date on which the project estimate of capital investment has, of course, checked can be terminated, and outline the probability of tech- with the engineering department, with production nical success. He goes further and provides a produc- personnel, and with others who might be in a position tion forecast which indicates the probable investment to offer advice and counsel. Research cost to date, as which would be required to put the findings of the stated earlier, is $2,000, and the supervisor has esti

Calculated Risk 4 mated an additional $25,000 will be required to corn- tion departments is estimated as November, 1951, inplete the program. stead of August, 1951. The estimated time table indicates that the comple- The project, however, still looks like an extremely tion of the research phase will require until approxi- good project-in fact, almost as good as it did before. mately March 1, 1951, and that engineering and con- The $17,000 has been well spent and the project would struction could proceed very rapidly because of the naturally be continued. very minor character of the changes involved and be- Let us take another look at how the project might cause of the fact that most of the required equipment have turned out, however, as indicated in Fig. 5. We is already on hand, so that savings from this change could be realized by our hypothetical corporation as REPLACEMENT OF CUPRIC CHLORIDE WITH COPPER IN early as August 1, 1951. METHYL ESTERATE PROCESS This project costing only a comparatively small amount of money and promising a return on invest-,S ECONOMIC FORECAST ment of 48 per cent would appear to be a rather ALL FIGURES IN THOUSANDS ESTIMATED OF DOLLARS TIME TABLE favorable one and accordingly would be placed on F DLLA IME ABLE the considered program of the research and develop- ment division. Work now proceeds, and we move _ As CON. from July 28, 1950, to January 20, 1951. At this time, T according to Fig. 4, we have spent an additional $17,000, bringing our research costs to date up to M A I Ks'I[ RES.. REPLACEMENT OF CUPRIC CHLORIDE WITH COPPER IN TOTAL SAy. COST SAVINGS SAV. TO DATE ADD METHYL ESTERATE PROCESS 24 ATR 19 CES 122e SAVr l C; ECONOMIC FORECAST l C A ESTIMATED. PROJECT EVALUATED JAN. 20, T. T. t 100 ALL FIGURES IN THOUSANDS TIME TABLE Figure 5 sT OF DOLLARS AT A N ENs. L V notice that several things have happened. We have V 73 T S spent our $19,000 on research as we did in the previous S N. case, but the results of that research have indicated S I' | | - IJ e that the savings which we had in mind in the begin-. M E ning are much lower than our previous estimate and A A E that the investment has also risen because we have v ~RES. found that in order to handle copper, we must make I R COST TO DE 22 l F other changes in processing, which involve the use of It. A ADD. ES, x -. special alloy equipment instead of steel, and the in73 ________COST_. vestment figure is accordingly increased. The project RT:EALUIAT~fCD JAN. ~ now looks rather sick because the net savings after Figure 4 taxes would be only $14,000, necessitating an invest$19,000. We are approximately one month short of ment of $180,000. Now, despite the fact that the table the completion date which our supervisor estimated looks more optimistic than in the previous situation earlier and have not been able to spend money quite and that the added research costs are only $6,000, the as fast as he had anticipated. Also, you will note that research director would probably terminate the prothe project has undergone considerable change. It is ject since it does not tend to improve the overall finnow estimated on the basis of the results obtained by ancial picture of the company. research that the savings each year after taxes would I might say, however, that the information contained be approximately $73,000 and that the investment on this chart is not necessarily the whole story. It may figure has not changed and that therefore the return well be that the plant is having such difficulty in operafter taxes would be increased to approximately 73 ating the process with cupric chloride as a catalyst per cent on the investment. We find also that the re- that the corporation would be ahead if it were to make search group has not buttoned the project up as ade- the indicated change to copper even though it would quately as they would like and now estimate they cost them some money. This would be a reflection upon must spend an additional $22,000 to complete the pro- the research administrator and his sources of informaject. The time for completion of research has also tion because if that were the case, the estimated savadvanced from the end of February, 1951, to the end ings on the chart should reflect in some manner the of June, 1951, and the estimated completion of the difficulties which the production department is having project in the hands of the engineering and produc- with the process at the present time.

5 Calculated Risk Now let us look at some of the other research pro- the corporation, approves the expenditure of the estijects which our hypothetical research and development mated additional $87,000. He does this because he bedivision has under way. One project on which it has lieves that any project that will yield better than 25 been engaged for some time and on which it has spent per cent after taxes is a good one, provided all the $95,000 to date is the Melur-oil project involving the other factors seem favorable. In this particular case, synthesis of melurine-l, 5-disulfonic acid and its appli- the sales department is extremely anxious to have the cation as an oil-treating agent for automotive use. This product, and, accordingly, there seems to be every project looked good when an economic forecast was reason to push the project. made on it approximately two years ago, and our Another project which is brought before the director latest forecast, which was made on July 28, 1950, still of research in the form of an economic evaluation is looks exceedingly promising. It will be noted in Fig. 6 the synthesis of diisopropyl chloronitrilidine (Fig. 7). that this project, which calls for a capital investment of approximately $2,400,000, will show a gross sales SYNTHESIS OF DI-ISOPROPYL CHLDRONTR1LIDIENE of $2,100,000. This does not quite meet the needs of ANNUAL PLANT CAPACITY 220,O00 LBS. our hypothetical corporation because the ratio is still 0 EST. adverse, but, on the other hand, this project shows a ECONOMIC FORECAST TAMEE good gross profit and a net return after taxes of ap- ALL FIGURES INTHOUSANDS OF O DOLLARS proximately 31 per cent and will require the invest- X R s A ment of less future research money to bring it to frui- i tion than has previously been spent. In other words, _xL 9 it MELUR-OIL PROJECT 0 ANNUAL PLANT CAPACITY 37,000,000 LBS. 395 2400 A S EST RAW MATERIALCONSUMPTON ECONOMIC 5 TIME. ( PER ANNUM) 0 TIME P ORPORATIO P100 N, FORECAST _ ABL R 234 PRODUCTS OUTSIDE 20N FORECAST S TABLE R P4 PO IuC'r E'D JT 0 MET....... G S ALL FIGURES IN R. A RES.S 20 TONS |R T THOUSANDS OF DOLLARS l.T 13 COST ISOPROPANOL NITRILIDIENE O M S I TAX SODATE ADD. s 2AME2RSS IOITONS 9 TONS S N i8 CO HLORINE ACTIVATED ~ O,; 3, "-' ALUMdINA S ION.=A 9 I PROJECT EVALUATED JULY 28,50 L 5 S 1220 Figure 7 a D S s IA_ A This material is a new compound, but application s RAW MATERIAL CONSUMPTION work in the corporation's laboratories has indicated p ( PER ANNUM) I'732 )eCpOR,,OAPRATIONI OUTSIDE that the product will serve as an intermediate for highA CT NAF 000TONSE MELURI0NE temperature resistant, shockproof plastics suitable for RES. X BENZENE A TONS TAX TO DATE RESCOST EAD. LOOOUL. THYL ETHER many industrial and military applications. The pro31% r-95 8aI o LYSTO ject is new, only $3,000 having been spent on it to date, PROJECT EVALUATED JULY 28,50 but the prognostication of its future is extremely promFigure 6 ising. It looks like one of these projects which can be the investment of the corporation is increased by about brought into commercial production with an extremely 2.4 per cent to add approximately 2 per cent to its small investment, a project which will add a rather annual sales volume, while increasing its profit by large amount of gross sales to the corporation's total almost 10 per cent. This naturally appeals not only figures and which shows an extremely high net after to the research director but also to the other executive taxes based on investment, namely, 173 per cent. The officers of the company, and he, accordingly, will have gross profit of 42 per cent, based on sales, is also high, no trouble in pushing the project through. Equally and, accordingly, the project looks extremely promisimportant is the fact that the project would consume ing-even more promising when it is estimated that products presently produced by our hypothetical cor- only $25,000 need be spent on the project to bring it poration, namely, 3000 tons of benzene and 5400 tons to fruition and that the plant can be completed for of chlorsulfonic acid. It would also involve the pur- operation approximately October 1, 1951. The rechase on the outside of other raw materials as indi- search director would push this project for all it is cated in Fig. 6, which, it happens, are extremely easy worth since it would appear to be an ideal project for in supply at this time and which we will have no dif- the improvement of the corporation's picture, offering ficulty in obtaining in suitable quality and quantity a very large return on a small investment and providfor several years to come, according to our purchasing ing early profits for reinvestment in other projects. and market research departments. Consequently, our Still another and the last of the hypothetical proresearch director looks at this project in its present jects to be considered is the development of a continustatus and, after discussing it with other divisions of ous process for the production of sodium hexametasi

Calculated Risk 6 licate (Fig. 8). Sodium hexametasilicate has been ECONOMIC FORECAST found by workers on the Continent to be an extremely FOR ALL EVALUATED PROJECTS suitable alkaline salt for water purification. The cor- FOR YEARS 1951 M 1955 INCL. poration has both a captive use for this product and ALL FIGURES IN THOUSANDS OF the possibility of external sale. The product uses one DOLLARS PER ANNUM 4,070 of the corporation's products and requires the purchase on the outside of a relatively cheap and readily | INVESTMENT 3,115 I] GROSS SALES SODIUM HEXAMETASILICATE CONTINUOUS PROCESS i NET AFTER TAX ANNUAL PLANT CAPACITY OF 80,000,000 LBS. _ 3400 ECONOMIC_ sa c EST. l TIM l FORECAST T TAIME TABLE 1,l50 s ALL FIGURES IN A R 915 T THOUSANDS OF _7 63 680 8 DOLLARS I 525 N5~~~ o ~'9 ENe. 510 2050 T 5 2 274 204 183 120 R C o 1951 1952 1953 1954 1955 SI AC Figure 9 L | H 4 | ala tion may have to invest, what the corporation may S0 _' _ realize in the way of added gross sales, and what the 0 R.AW MAERIAL NUMnet after taxes may be. This type of information is P 46 E ADO. RES. CORPORATION OUTSIDE particularly helpful to other executives in the comRES. 2COST PRODUCTS LS TO AT SAW TONS 3,000 TONS pany since it shows them what their research program 14%T~~~ CAUSTIC SODA SAND PROJECT EVA-ATED JY 28, 5 may lead to. The people primarily concerned with the PROJECT EVALUATED JULY 2, 50 financial fortunes of the corporation are put on guard, Figure 8 because the program which the research division has available material. This again is a new project since under way will require the investment of certain sums only $6,000 has been spent on research to date. The of capital. They may then decide either that this capiprobable additional research costs are somewhat high tal requirement is exorbitant or that it can be readily -$250,000. The investment again is adverse, an in- met, and that decision will naturally have an effect vestment of $3.4 million being required for a gross upon the research program. Similarly, the added sales sale of $2,050,000. The gross picture is also not very which will result from year to year are of interest to good since only 39 per cent would be realized and the the executive in charge of sales. He knows whether he net after taxes is only 14 per cent. The research direc- needs to be training or hiring new salesmen and when; tor looks this picture over and may decide to go one he knows, from a consideration of the individual proway or the other. In this particular case, however, the jects, what kind of salesmen they should be, and thus sales department is demanding a product of this type, he is placed on notice as to exactly what may be exand he knows that many of the corporation's present pected of him in the future. The material is of equally products could be markedly improved if sodium hexa- great importance to the production department and metasilicate were available at a reasonable cost. There- other groups. fore, although the net after taxes does not measure up to the 25 per cent, which he has arbitrarily adopted Fig. 10 indicates the cumulativ e e ffect of the few research projects discussed on the future fortunes of as being the standard for research projects, he goes corporatio ahead wih gd wl ad rthe corporation. It shows that, if the projects were ahead with good will and retains the project on his successful through 1955, a total investment of $6,035,program. 000 would be required, which would result in gross Let us assume that these four projects constitute the sales of approximately $5,200,000, with a return after entire research program of the corporation under dis- taxes of $1,500,000. This would hold some appeal to cussion. This obviously is foolish, because the total the corporation as a whole because the return on the amount of research expenditure called for by these new projects after taxes is very much better than that projects is very much lower than that which a corpora- presently enjoyed by the corporation. Although this tion of the size indicated should be investing in re- portion of the corporation's total research program is search each year. However, time is limited and the decidedly deficient in that it does not improve the delineation of all the projects which such a corpora- ratio of total sales to investment, it does have a very tion should have would serve no useful purpose. appreciable effect on the over-all return-both on capiIn review, let us look at a summation of the economic tal investment and on sales. The net effect on corporate forecasts for these evaluated projects covering the finances would be to increase the return on capital years 1951 to 1955, inclusive, as shown in Fig. 9. Here from 8.0 to 9.0 per cent and from 7.3 on sales to 8.3 we find the picture, year to year, of what the corpora- per cent.

7 Calculated Risk duction capacity certain amounts of isopropanol and ECONOMIC FORECAST chlorine and that these amounts are not small but reFOR ALL EVALUATED PROJECTS present approximately one-third of the present unsold ONA CUMULATIVE BASIS capacity of the company. He therefore should be someALL FIGURES IN THOUSANDS what reluctant to enter into long-term contracts to sell OF DOLLARS a greater volume of isopropanol and chlorine or, alternately, he should insist that if he decides to enter into Q:3 INVESTMENT 6,035 such contracts, the production department be prepared I NEROSS SALES T5A87 4 I to expand their capacity so that by 1952 adequate C3 NEt AFTER TAX 4,677 4,070 A amounts will be available to supply the needs of the -? I~ research projects presently under way. Actually, in: the consideration of such figures, executives outside l1,562 l, 1 80 1 500 T the research department should pay very little atten1,0016,037 o,38 i~ tion to figures projected for more than one or two 235 274 465 680 X years. After all, the fact that our requirement for r1951 195 1953 0-r — n_ chlorsulfonic acid is going to take 54 per cent of the Figure 10 present unsold capacity in 1955 is not a very serious Another use of the information shown in the pre- matter as far as a sales manager is concerned. So many viother usres is toe in a form ecast of the demand things will happen before 1955 that he need not be on the corporation's present products to determine concerned about this requirement until it is more imwhether or not the present capacity of the corporation minent. is adequate to meet the demands which the research The estimated chances of success of the various prodivision is trying to set up by the expenditure of the jects have been passed over very lightly in the previous corporation's funds. Such a forecast is shown in Fig. discussion. In Fig. 12 is shown the estimate of the di11, which indicates for the years 1951 through 1955 the requirements for raw materials produced by the ESTIMATED CHANCE OF SUCCESS corporation for the research program presently under way. This is of tremendous aid to the sales and produc- PROJECT RETURN PROBABILITY PROBABILITY ON INVEST. OF TECHNICAL OF OVERALL tion executives of the company. In the case under % SUCCESS, % SUCCESS, % consideration we have no trouble because in every COPPER CATALYST 48 100 eS ANNUAL CONSUMPTION OF CORPORATION'S SODIUM HEXAMETASILICATE 14 100 95 PRODUCTS BY ALL PROJECTS MELUR-OIL 95 1 % OF PRE- PLAST 175 90 s o YEAR CONSUMPTION PRODUCT % OFANNUAL UNSOLD RLT30 TONS CAPACITY CAPACITY....__ 1950 Figure 12 1951 70 ISOPROPANOL 1.5 8.2 30 CHLORINE 0.15 iO.O rector of research as to the probability of success, both 1952 254 ISOPROPANOL 6.5 35.5- technical and overall, of the projects discussed. About 101 CHLORINE 0.5 34.0 the copper-catalyst project, which shows a 48 per cent 1953 1000 BENZENE 2.3 5.4 return on investment, he feels quite confident that 1800 CHLOROSULFONICACID 8.2 18.0 there is no possibility of falling down on the job and O54 CHSOROPANOL 0.5 345.5 that the copper catalyst will work. He does feel, how1954 6000 CAUSTIC SODA 1.6 34.0 ever, that the plant may have some difficulty in putting 3000 BENZENE 68 16.2 this catalyst into use and therefore he rates down his 5400 CHLOROSULFONIC ACI 25.4 54.0 254 ISOPROPANOL 6.5 35.5 probability of overall success to 95 per cent. For sodium 101 CHLORINE 0.5 34.0 hexametasilicate the same situation prevails. About 1955 8000 CAUSTIC SODA 2.2 195 00 BENZENE 6.8 2.2 Melur-oil he is not quite sure and feels only 95 per 5400 CHLOROSULFONIC AC1D 25.4 54.0 cent chance of success from the technical point of view, 254 ISOPROPANOL 6.5 35.5 I 01- I CHLORINE I. as 5 34.0 and overall, primarily because he is concerned about Figure II the reliability of his market information, he rates it instance the amount of product required as raw ma- at only 90 per cent. With Pre-plast, however, he has a terials for the production of products which the re- still different situation. He scales his probability of search division hopes to develop is less than the unsold technical success down to 90 per cent and the probacapacity of the company. They do, however, give the bility of overall success to only 80 per cent, but the executive in charge of sales some reason for concern. 173 per cent return on investment stands out foremost He is definitely put on notice that as early as 1952 in his mind, and he is perfectly willing to keep the this corporation may demand from its present pro- project active.

Calculated Risk 8 Today, in our troubled world situation, there is who have contact with the project are highly enthusione other type of calculated risk which any director astic about its probability of success. It would therefore of research must definitely consider, namely, how a seem to me to be an ideal project to push, even though project which he has under way fits into the overall the return on investment is considerably lower than planning of the nation, as he sees it. And here I might that which you would normally consider as good. On say parenthetically that "as he sees it" is a very im- the other hand, if you have a project which shows a portant phrase. An important part of the job of a 42 per cent return on investment, and production and director of research today is making every effort to be sales are against it for reasons which you feel quite sure that he sees it right. He must be informed on confident are wrong but which represent judgment both national and international affairs and must be in areas where sales and production people should be, in definite rapport with the changing events about us. and probably are, better informed than you, then you His ability to interpret may make the difference be- should turn it down because it is definitely a poor tween success and total failure in keeping his research risk. It is a poor risk probably because the bases of organization in business. Thus, a consideration of their judgments are better than yours. Moreover, if the data in Fig. 13 might well lead to some changes your facts are right and theirs are wrong, you will not have the kind of support from the organization as a MILITARY SIGNIFICANCE OF RATED whole that would be necessary to put the project over. PROJECTS The system discussed here has been used in our own organization for a short time only, but we are highly IMPORTANCE IN WAR gratified with the results. I do not say that this system HIGH MED. LOW is a perfect system or even that it is an extremely good COPPER CATALYST X system. I can say, however, that from our own experience it has been a very important tool in getting every SODIUM HEXAMETASILICATE X man on our research and development team as well as MELUR -OIL X the men in other divisions of the company to think in terms of calculated risks. We find that, after exposure PRE - PLAST X to this type of system for only a very short period of time, the man at the bench begins to think in terms Figure 13 of ultimate economics, and I know of no healthier in the research program to keep its military significance situation in an organization whose primary aim is the of a properly high order. commercialization of research. After one has made all these calculations and looked Obviously, most of what has been said would have at all these figures, one may believe to have. a basis no application whatsoever to fundamental research for estimating the risk involved in research projects. as conducted in our universities and some of our govOne is still faced, however, with the necessity of using ernmental laboratories. There, the primary aim is to judgment as a deciding factor. One knows that most push back the frontiers of knowledge rather than to of the figures he uses are not factual, that they are put to work scientific information and data for imguesses and that the area of certainty represented by mediate or at least early commercialization. each guess is open to question. One must also allow In conclusion, let me say that organized industrial the judgment of others on some of these points to in- research cannot achieve fullest success if the president fluence his own thinking, even when he is relatively and financial officer of the corporation, and for that sure that the judgment being expressed is incorrect. matter, the research administrator are the only ones After all, the modern industrial organization moves whose decisions are governed by economic principles. as a team, and teamwork needs the subordination of While the research worker must be primarily a person individual thinking and judgment to that of the group of scientific integrity, he should possess also an awareas'a whole. ness of economics as an attitude and a means to an For example, suppose that $370,000 has been spent end. Such a philosophy does not require an adding for research on a new project. Its technical success is machine for every chemist and a cash register in the predicted as 100 per cent, the return as only 12 per laboratory storeroom, but it does call for a wrecking cent, but the controller, sales manager, production crew for every ivory tower unless exceptional justifidepartment, engineering department, and all others cation can be found for its existence.

9 Calculated Risk DISCUSSION OF MR. VAUGHN'S PAPER MR. MARLOWE: I would like to raise one question, ploys an unduly high safety factor in order to insure and that is how on earth can you make the background that the actual costs of construction will always be necessary for judging markets available to a research comfortably below the estimated costs, you may come worker without spending a great deal of time on that up with an investment figure which turns the project necessary background information. down. MR. VAUGHN: The way that works out is this: The MR. VAUGHN: That is one of the toughest problems research supervisor generally has some general familiar- in this whole job. Our estimates are normally based ity with the field we are talking about, and we have a upon partial layout of a plant by an engineer who has market research department, whose figures are avail- had a lot of experience in "guesstimating," and we use able to him and to his group. The supervisor starts out his figures just as though we believed they were corwith some ideas on markets and by digging and by rect. The engineer shouldn't try to shade his figures talking his problem over with other informed people either high or low, but whenever he foresees trouble, in the company, he learns more and more about mar- he should take it into consideration and take the most kets and much faster than he otherwise would. Then, expensive way out. For example, if the question of after he arrives at an overall idea, it is finally subjected using steel or stainless steel comes up, he should figure to review by people who are usually still better in- on the use of stainless steel, thus putting in a possible formed. cushion. You would be surprised to know how often the su- I have seen plants, and everybody in this room has, pervisor's information holds up, and what a good that have been engineered, appropriated, and built effect it has on the entire staff to have them thinking which have cost 30, 40, and even 50 per cent more in such areas. than they were supposed to. If this happens on a proMR. BOYER: I was wondering to what extent you ject that shows 25 to 30 per cent return on investment, take into account the possibility that another company it doesn't cripple you too much, but if you get it on in a somewhat competitive field is working on about one that shows only 12 per cent, you are in serious the same type of product development and therefore trouble. is going to be competing with any possible market or So we have tried to keep our engineers from altering has patents which will dominate it. their guesses one way or another. We would rather get has.atent.hic will dominathem exactly as they see them. If they see them right, MR. VAUGHN: Well, we try to take those things into that is good, but if they see them wrong and pad them account insofar as we can. We had a project just re- too, you are in a bad way. cently, with which we were pretty far along, and we learned that one of our competitors was at the present. WALKER: You and I have talked about stimulating productivity in research personnel. I am wontime expanding their capacity and would have their lating productivity in research personnel. I am wonproject in operation over a year before we could pos- dering whether in handling these economic values, sibly have ours in operation. In that particular case, we decided that the market would expand considerably than they used to be and absorb not only their increase but ours, and for- MR. VAUGHN: There isn't any question about it, tunately, we were right. Had we thought their expan- definitely yes. sion program would more than eat up the existing MR. ABRAMS: Could I make one statement, just market, we certainly would have held our program in to throw in one more complicating situation? Do you abeyance. make any allowance for the purchasing value of the Generally, you do the best you can, based on the dollar? I think that ought to hit some of these things information before you, but there is no way to be that are three years off about as bad as anything. certain that you always reach the right conclusion. We make an estimate, and then we add enough to MR. DOAN: I would like to make one remark: Esti- overcome any inflationary effect. When the dollar can mates of investment costs are subject to a lot of varia- drop from $1.00 in 1940 down to 58~ now, you can see tion, depending on whether you are an optimist, a what is going on. pessimist, or a realist. I suppose most research super- MR. VAUGHN: I think there is some compensation visors have had the experience at one time or another built into the system you are talking about, that is, the of seeing an otherwise attractive research development sales dollar increases along with inflation, as a general go "out the window" because of engineering estimates sort of thing, so that the ratio between sales and inof high investment costs. If the estimates are realistic, vestment remains somewhat constant. as they should be, a reliable appraisal of the project The other thing is, don't worry too much about is obtained. On the other hand, if the engineering figures being valid three or four years from now, begroup responsible for making the plant estimates em- cause we are spending money trying to refine the

Calculated Risk 10 figures and find out what they actually are. The figures nothing, and you can't hold him too closely to his we like to have confidence in are those for 1950 and estimate. I think the uncertainty of his capital invest1951. We like to think that they are close to being ment figure is much greater than any effect inflation right. might have on the value of the plant. On the project that I showed you that ends up in And then, as I say, you have the overall compensa1954, an engineer has had to design a plant, and we tion of the value of the dollar reflected in both what have the research just started. He is working with you buy and what you sell. - II - by E. D. REEVES Executive Vice President Standard Oil Development Company Several years ago it became apparent to our research not primarily by its scientific achievements but rather people that the heavy-duty diesel oils our company by the usefulness of these achievements to the company was marketing at that time would not give entirely supporting the research. The fact that industrial resatisfactory performance in a new type of diesel en- search results are useful is no mere coincidence; it is gine that was about to appear on the market. This the result of endless planning of research programs meant that something had to be done immediately for this specific purpose. Each research project must to correct the situation, and, after a thorough study be carefully examined with this in mind, and only of the problem, the specialists in this field felt that those projects selected which fully meet the criteria one of our products could be improved by the time of ultimate usefulness. These decisions determine what the new engine got on the market. At the same time, research will be pushed and what will be abandoned, however, they expressed the opinion that what this and each decision represents a thoughtfully calculated engine really needed was an entirely new type of pro- risk on the part of research management. Willingness duct. Because of limitations of time and personnel, it to assume calculated risk is an inseparable part of the did not seem possible, however, to work on the im- management of industrial research and is a large factor provement of our current diesel oil and to develop in determining its success. Though you all know many an entirely new one at the same time. After some dis- examples in which real technical achievement and succussion of the matter it was finally decided that no cess have not gone hand in hand, I might mention work would be done on improving the existing pro- that if our research group concerned with the developduct but that all our efforts would be concentrated ment of oil burners were to come up with a truly reon the entirely new product. I am glad to say that volutionary coal stoker, our company would not be this work proved to be quite successful, and our com- particularly impressed with this scientific achievement. pany had a lubricant that gave entirely satisfactory While it is easy to talk about calculated risks and performance in the new diesel engine when it ap- their use in industrial research, it is not nearly so easy peared on the market. I mention this occurrence be- to put this discussion on the quantitative basis that the cause it illustrates a typical decision that must be made term implies. I think it can be said that what we are in planning industrial research almost every day. really trying to do when we take calculated risks is to I think you will easily recognize that we were forced secure maximum usefulness from a given effort. As to take a calculated risk here; that we had to choose applied to research, this means that we are trying to between bringing one of our existing products to a achieve some sort of a balance between the cost of point where it would get by or developing a sub- doing research and its probable value, i.e., we are trystantially better product. In the first case, the risk was ing to direct our research efforts into the most profitsmall and so were the benefits. In the second case, the able channels. This immediately raises a lot of quesrisk was quite a bit greater but so were the advantages tions about the value of research to a company that of a successful solution. are not easy to answer. I hope that this afternoon's When we think about industrial research, it seems discussion will throw considerable light on this probto me that there is one very important point that must lem and would only like at this time to mention some always be kept in mind. This point is that the suc- of the things that have to be considered when we try cess of an industrial research organization is measured to answer the question, "What good can research do?"

11 Calculated Risk Whenever a company decides to carry out research or he is going to do next. Failure on the part of research to support a given research effort, it takes the first of management to consider all these factors usually rea series of calculated risks in industrial research. The sults in research work being done on problems where fact that business managements have come up with the value of research can be determined by fairly so many different answers to this question clearly in- simple calculations, such as the direct replacement of dicates the complexity of the problem and the diffi- an existing process by another one making the same culty of putting down the answer in black and white. products but having a lower operating cost. In the It is, nevertheless, a very important factor in the long run, concentration on the more obvious needs evaluation of specific research projects because the of the company for research is bound to result in a lack usefulness of research to a business will depend on the of balance between the research program and the capacity of the business itself to put research results company's real need for research even though the to work. This is affected by the type of business, the research that is done appears to be paying handsome character of its management, the availability of capital returns. funds, etc. All of this must be kept in mind by re- The next thing that has to be considered in attemptsearch management in trying to evaluate its research ing to balance research effort against its value is the programs. cost of the research itself. This is a much less compliIf we now turn to the evaluation of specific research cated problem but still requires careful study, since projects, we will again find that the research director many research costs are often hidden among the comneeds an intimate knowledge of the business for which pany's operating expenses. Some of the things that he is doing research. The first thing he must deter- have to be considered are the costs of the laboratory mine is what phases of his company's operations are in work, pilot-plant operations, engineering, sales degreatest need of research. Should he work on raw- velopment, and possibly the cost of the plant that material supplies? process improvements? the develop- would be required to utilize research results. This last ment of new products? or even improvements in mar- item could be important since, other things being keting techniques? Although all these questions need equal, the research project that results in the least to be answered with reasonable accuracy, it is not capital requirements would have attractive advantages always easy for research management to make an ac- over the others. curate estimate of the importance to his own company The final factor in relating research costs and the of the successful solution to a given research project. value of research is concerned with the probability At the risk of further complicating the problem, I of success for the research project. In looking at our might even mention other factors that have to be con- own research costs for these purposes, we have come sidered. One of these is the consequence of failure up with a term that we call the equivalent cost of to work on certain projects and bring them to a suc- research; that is, if a research project is estimated to cessful conclusion. For example, if the company is cost $50,000 and we think it has a fifty-fifty chance of faced with a loss of its normal raw-material supplies, success, we then say that it has an equivalent cost of it is much more important to develop an alternate $100,000. In other words, we divide the estimated cost raw material than a new processing method. There whether successful or not by the probability of sucare also many corollary advantages associated with cess. This brings us to a relationship between the new products. In our business, the development of an value of a research project to a company and its equivaoutstanding grease not only often brings in new grease lent cost, which we call the desirability factor. I might business for our company, but also gives our marketers illustrate how this works. Suppose, for example, that an opportunity to sell fuel oil and many other pro- the successful solution of a given problem would be ducts to the same customers. Finally, there is the ques- worth $10,000,000 to a company and we have two postion of timing. It frequently happens that the value of sible research projects by which this problem could be a new process or a new product to the company de- solved. If one of these projects were estimated to cost pends on its availability at a certain time. For ex- $50,000 and its chance of success one in twenty, and ample, if the company is planning to expand its plant, the other estimated to cost $200,000 with an estimated a new and cheaper process would be very important chance of success of one in two, the equivalent cost of just before the expansion took place. Six months or a the second would be $400,000. On this basis, the secyear later it might be of no importance at all. ond method of attack represents the better risk, since In listing all the considerations I have mentioned, the desirability ratio in the first case is only ten while I am not just trying to make this sound completely con- it is twenty-five in the second. fusing. All I am trying to point out is that in deciding With the above in mind, we might see how these how much a successful research project might be worth various factors are applied in the selection, control, to his company, the research director is required to and termination of specific research projects. Starting be aware of all of the factors that the owner of the with the selection of projects, there are three things business himself has to consider when he decides what to consider. The first involves what might be termed

Calculated Risk 12 the use of categories in setting up an over-all frame- detailed technical programs. Here again, it is importwork to match research against the company's needs. ant to streamline the project as much as possible and This means that there must be some balance between refrain from obtaining information that is not esresearch on products, processes, raw materials, etc., sential to the final solution of the problem, even in accordance with the requirements for research of though it has a certain amount of scientific interest. the various phases of the company's operations. Within If the procedures that have been outlined above are each of these broad categories, it is also desirable to consistently followed, research projects will automatihave what an investment banker would call a balanced cally terminate as the need for them drops off or as portfolio. For example, let us say that we are consider- data obtained during the course of the project indiing research on processes. Part of the research ought cate that its chances of success have gone down. This to be equivalent to Government bonds; i.e., it ought does, however, require a constant and continuing to be work that is definitely needed by the company review of all projects, with particular emphasis on the and where the chances for a successful solution are distinction between technical interest and usefulness. very high even though the returns are not too great. I know that during the war our company was very In this class would come work on improving the de- anxious to develop processes for the manufacture of sign of fractionating towers, improvements in the de- triptane, which is a very high-quality component sign of process equipment, the determination of that would be useful for aviation gasoline. Having optimum operating conditions, etc. The next category started on the project, we kept right on working even might be what is called the blue-chip stocks. This though about halfway through, the importance of covers work on new processes for which the company making triptane had lost a great deal of its urgency. has a definite need and on which the return to the The reason we had so much trouble in stopping the company would be very high if the research were suc- project was that our research people did not want to cessful. In our company this might mean the develop- admit that they could not solve the problem and obment of an entirely new cracking process or a method jected to discontinuing the work without a suitable of making synthetic lubricating oils cheaply. Finally, answer. While there are many other phases of induswe come to the class of problems known as the old trial research that require the use of calculated risks, gold-mine stocks. These are projects which might take I think that the above discussion serves to illustrate the company into entirely new fields if successful and the fact that the taking of calculated risks is an integral could represent a real contribution to its expansion. part of industrial research. I think that I can sumThe development of Butyl rubber by our company marize our own feeling about this as follows: some years ago is probably a good example of work In carrying out industrial research, the success of along these lines. the work done is measured by its usefulness for a Once the company's research programs are set up specific purpose. These specific purposes must be to give a preliminary balance between what appears achieved by the research organization with a minito be the needs of the company for research, it is neces- mum amount of effort, and the over-all research prosary to take stock of the various individual projects. gram must be balanced against the needs of the comThis can be done by calculating the desirability fac- pany supporting it. For this reason, there are two tors for each project and then comparing them with things that are demanded of an industrial research others of the same general type. Comparison of the organization. The first of these is technical competence average desirability factors for research projects in in carrying out the research itself, and the second is different fields can also be used to check on the pre- wise planning or the use of calculated risks in meetliminary balance that has been set up and to earmark ing its responsibilities. the least desirable projects for replacement by new Calculated risks must be taken through all phases Iones. believe t is aparent fom the bove dicussion of the industrial research effort. This includes deterI believe it is apparent from the srabove discussion mination of the total research justified, its distribution that pretty much the same considerations would apply in different fields and the selection control and termto the control of the research projects themselves. In ination of specific projects. Calculated risks must also determining the magnitude of effort justified for a.... determining the magnitude of effort justified for a be taken in the direction of individual research progiven project, our normal feeling is that the most grams if these programs are to be of maximum effecimportant projects in each field should be given as these programs are to be of maximum effecmuch effort as they will stand. Our reason for think- tveness. ing so is that it is always desirable to complete a I feel sure that failure on the part of research project before the need for it disappears. Also, there management to take calculated risks would soon doom are certain psychological advantages to winding up that particular organization to mediocrity. On the each project as quickly as possible. In the control other hand, the evaluation of the calculated risks reof projects there arises a certain type of calculated quires all the skill and resources that research managerisk that is more concerned with the planning of the ment can bring to bear on it. It requires an intimate

13 Calculated Risk knowledge of the business for which the research is of calculated risks by research management is at best being done, an intimate knowledge of the research an intelligent effort to make the most of what is availorganization itself, a knowledge of the actual and prob- able. It is by no means a substitute for a good, hardable costs of carrying out research, and the ability to hitting research team with ideas for things to do. Its evaluate the real usefulness of the results of research. purpose is not to make poor research look good, but With all of its importance, however, the assumption only to make good research useful. DISCUSSION OF MR. REEVES' PAPER A VOICE: Both of the speakers so far have outlined have done more recently on synthetic fuels. In other systems for the calculation of risks. As I understand it, words, hydrogenation was of great interest to us in both have also outlined direct calculation of some of'28 and'30, because it was thought that there might the risks, followed by, let me call it, guessing at some be a shortage of oil, and hydrogenation was going to of the other risks. You figure out the cost of the new give us a much bigger yield of petroleum products from plant and the quantity of new sales you will have, the crude. and then you guess a factor of anything between one That did not turn out to be the case, so I think and ten as to the chances of success in the research, we might say that we made a mistake because we and you multiply or divide by that. thought there was going to be a shortage. The last That leads me to wonder how long these systems time we did work on synthetic fuels, we didn't think have been employed, where they have been employed, there was going to be a shortage. and whether any post mortem has been done on them. A VOICE: I can appreciate your taking a called third Can you cite the batting average of your system of strike, and that is one of the worst things we do in calculation? How many gross errors do you discover its effect on our batting average. I am still interested years afterwards? Can you discover your errors within and wondering whether you have calculated in any a year or within some years afterwards, and can you way, shape, or manner, the batting average over a by any means of investigation determine your own long time and how it works out by the calculation past batting average? you do. In other words, have you tested your method MR. REEVES: We are somewhat like the doctors; of calculating risk in the light of past experience? we bury most of our mistakes. I think that in most MR. REEVES: Yes, I think some of that is going to cases where you really make an error, you fail to do come up this afternoon, because we made quite a something which you should have done, because, as study of it about seven or eight years ago and tried you go through a particular project, you always have to calculate what we thought was the value for the the opportunity of stopping it if it doesn't look prof- company of the research we had done over a period itable. of years. We then tried to estimate the value for the Now, we have made mistakes in the past, and I next five years and recently recalculated it to check think the biggest mistake we ever made was in spend- on our estimates. ing as much money as we did on hydrogenation back The only trouble with the calculations is that it is in 1927 and 1928. That was a mistake because we hard to get the different groups to agree on the estinever got to use the process very much. mates. In other words, who makes the money on the On the other hand, the work that we did at that new product? the sales people, the manufacturing time was justified on the same basis as the work we people, or the research people?

Calculated Risk 14 - III by DONALD H. LOUGHRIDGE Senior Scientific Advisor Office of the Assistant Secretary, Department of the Army At the time Dr. Work first approached me on the believe we mean, namely, "likely to produce a certain question of speaking at this year's Conference on the effect." Administration of Research with the suggested title As far as "risk" is concerned in our present terminof "Calculated Risk," my first reaction was to wonder ology, it is clear that "exposure to loss" is the only as to whether the military had taken over our annual meaning given by Webster which can be applicable. meeting and as to whether we were to be subjected Hence, the combined meaning of our term in the to the so-called "military-mind" type of thinking. My present field of discourse must be "likely to produce reaction was due to the fact that I did not recall a certain exposure to loss." I should hasten to add, having heard the term "Calculated Risk" applied to however, that I would not restrict this loss solely to shrewd guessing before becoming closely associated financial loss. It could be equally well represented by with the military services. I replied to Dr. Work's in- loss of man hours, loss in competitive standing, either vitation stating that I did not believe there was such from a business or military standpoint, loss in the rate a thing as "Calculated Risk." However, since I am of pushing back the scientific frontiers of knowledge, an old member of the American Physical Society or the individual loss of the satisfaction of having and have long lived under the necessity of submitting given birth to a new, really fundamental concept. a short abstract about six weeks prior to a meeting of Now, since I believe I should present a discussion the Society to get placed on the program, I, along with of the application of this concept of calculated riska percentage of other members, have been known to or the likelihood of producing a certain exposure to stoop to the practice of sending in some preliminary loss-to the problem of military research and developresearch results with the hope that by the time the ment, let me take a few minutes to explain to those of meeting rolled around, we would have time to obtain you who may not be sufficiently familiar with the more convincing experimental data. On the basis of methods used within the Department of the Army how this principle I accepted Dr. Work's invitation to the R & D program is set up. speak. Your chairman apparently was also cognizant The Research and Development Board issues anof the above-mentioned general practice because he nually its Program Guidance to the three sister services, replied (to paraphrase his remarks) that it was not Army, Navy, and Air Force. This guidance assumes absolutely essential to talk on the subject nor to know the form of discussing in terms of eighteen catagories much about the subject but that it was his hope that of the RD B Master Plan those fields in which the by the time our Conference convened I would have RD B believes more, or less, or equal, support should had the opportunity to learn something about this be given in an approaching fiscal year as contrasted military figure of speech. with the present year. To mention a few of these cateHaving thus become involved in the implied tacit gories we find such titles as Land Combat, Air Defense, agreement to put some thought on the subject, I first Strategic Air, Sea Combat, Supply, Supporting Rewent to Webster to ascertain what was meant by search, and Basic Research. Each of these broad cate"calculated." I found the following: "Adapted by gories is again broken down into Technical Objectives, calculation, forethought, or contrivance to accomplish numbering from a very few up to 20 or 25 within each a purpose, hence loosely, likely to produce a certain of the broad categories. effect." The Department of the Army (and I presume a For "risk" I found: "Hazard, danger, peril, exposure somewhat similar procedure is followed in the Navy to loss, injury, disadvantage, or destruction." and Air Force) uses this Program Guidance of the Since to a physicist the interpretation of "adapted RD B to aid in the annual evaluation of its past R 8c by calculation" could mean nothing except the use D program and the formulation of its new one. It of precise numerical methods and since in the initial should be remembered that this formulation of proprocesses of selection of research projects it has been gram necessarily occurs approximately two years prior my experience that educated intuition plus mere ac- to the fiscal year in which it is to be executed. The cident has been the primary basis of outstanding procedure of Congress in granting yearly appropriascientific advances, it would appear that we must use tions requires this long-range planning. At about the Webster's last definition to represent truly what I same time as the RDB Program Guidance is fur

15 Calculated Risk nished to the Services, the Secretary of Defense gives name of research, whereas, in fact, most of the work to the Services through the RD B an interim R 8c D is development and consequently directed toward the ceiling which their proposed program must not ex- solution of specific problems. At the risk of being ceed. pedantic, I should like to establish clearly what I beNext begins the detailed breakdown of these bud- lieve to be the difference between research and deget ceilings into various classifications. Within the velopment. The basic objective of real research is to Army, the Office of the Deputy Director, G-4, for Re- add previously unknown knowledge to our stockpile search and Development proposes an interim break- of scientific facts and to explore fields of science withdown of the Army's R & D ceiling figure into an in- out any requirement as to a specific application. On terim ceiling for Ordnance, Signal Corps, Chemical the other hand, development or engineering work Corps, Engineers, Quartermaster, Surgeon General, consists in the careful and logical application of preTransportation Corps, and an Army-Wide category. viously known facts and conditions to the problem of Each of these Technical Services in turn submits a satisfactorily attaining a previously desired end-objecseries of projects, with estimated costs and tabulated tive or product. in terms of the Technical Objectives within each of Research problems of the kind above described can the Strategic Categories of the R D B Master Plan. only be recognized by those actively engaged in the The Joint Chiefs of Staff meanwhile assign priority work. It is the immediate result of the solution of classifications to the eighteen catagories, and the R D B very closely allied problems by the investigator, or assigns technical promise evaluations for the technical others working in the same field, and is conceptually objectives within each category. impossible of even formulation by one not closely Within the Department of the Army the next step allied to the particular field of research endeavor. To is for the seven Technical Services to submit their go even further, many of the most outstanding and proposed R & D program, properly budgeted within valuable discoveries are often the result of almost actheir respective ceilings, to the Deputy Director, G-4, cidental observations by the carefully trained investifor R 8 D, who now appoints a committee of three, gator of phenomena which came to his attention of late usually consisting of himself, a representative solely due to the fact that he was interested in a difof the Army Field Forces, and the present speaker, to ferent, though perhaps closely allied problem. To menhold a series of hearings with each Technical Service tion such a specific case, one needs only to recall the with a view of examining the proposed program in wave-like nature of electrons discovered by Davison detail, project by project, requiring justification where and Germer at the Bell Laboratories as a consequence desired, suggesting increases or decreases in the financ- of their initial studies on electron scattering by crysing of particular projects, elimination of some, and tals. The Nobel Prize was given for this industriallyaddition of others. In this review, there naturally re- financed piece of research. sults a shift in budgetary ceilings between the Techni- Although it is true that no one except a working cal Services. Since the Review Board is made up of scientist can intelligently select a real research probone member from the Army General Staff, one from lem, a good research director can support and encourthe Army Field Forces, and one representing the scien- age work in a particular field, and on the basis of his tific interest, the hope is that by this means a balanced usually greater experience and more mature judgment, Army R &c D program will result. he materially assists in the choice of fields in which An additional final check however is made by the research work needs emphasis. Services presenting their budgets to the RD B. Com- It has been my experience that many perfectly capamittees, panels, and the General Secretariat go over the ble research men, especially in their younger days, entire Department of Defense program project-wise become easily discouraged by being subjected to the and budgetarily, the theory being that a proper bal- request of management or superior administrative ance between the three Services and elimination of authority to justify budgetwise their needs for equipduplications and gaps will thus be achieved. ment and technical assistance. Quite frequently they After the above dry discussion of the channels pre- may lack someone with whom they can discuss their sently existing within the Army and the Department present state of progress and receive a fresh interpretaof Defense for the establishment of its R 8& D program, tion or suggestion which puts the troublesome problem it becomes desirable to examine in somewhat more in a new light so that a logical deduction or interpredetail the presence or absence of calculated risk (as tation becomes apparent. previously defined) in the procedure. Since the sub- Since we are talking about the place of calculated title of this morning's series of papers emphasizes re- risk in research projects, I believe I have said enough search projects, let us first be sure that we are all to show that is is my belief that the question can be interpreting the word in the same way. It has become answered only on a specific basis. Given, let us say, all too common practice within industry and the mili- one million dollars for research and development, what tary services to lump too much under the glamorous fraction of this sum should be used for real funda

Calculated Risk 16 mental or basic research? The question as stated is those critical times of financial expansion, there is alstill unanswerable until we specify the type and pur- ways a concomitant demand for rapid completion of pose of the organization administering the funds. If development items, a great need for money to purthe institution is a university, the answer is easy-80 chase service test items of hardware, and a belief that per cent should go to actual research with a 20 per in times of critical stress research must give way to cent allowance for overhead; but the consideration development. There is no doubt that the latter viewof calculated risk (meaning, likely to produce a cer- point must, in general, be supported, and hence we tain exposure to loss) should still be carefully evalu- find real research pushed out to a great extent at all ated in terms of the distribution between research times in the military R & D program. I must take this workers. The requirements of the scientifically more opportunity, however, to point out the exception mature, who have proven their ability to obtain re- which is always needed to prove the rule. The Office sults, must be balanced against the eager young minds, of Naval Research has done a brilliant piece of work who may be more imaginative and who have the extra during the past four years in its support of really basic incentive to produce in order to establish their scienti- research in the natural sciences. In the period since the fic reputation. To restrict the opportunities of an ex- war, when it became clear that this country must take ceptionally able but unproven young man can produce the lead (previously held by the European nations) a great likelihood of exposure to scientific loss. in fostering science, and before Congress succeeded in In the case of laboratories of large industrial or- establishing a National Science Foundation, the Navy ganizations, 5 per cent and sometimes as high as 20 grabbed the ball and has pioneered very successfully per cent of the total time can be spent on basic re- in fighting the tendencies to which I previously research, with the laboratory still yielding overall divi- ferred. An examination of the budgetary support dends to the industry which supports it. which the Navy has been able to provide for science in However, if the institution spending the million the last few years will readily show, however, that the dollars on research and development is a military or- ceiling provided in the bill lately passed by Congress ganization, we have a complicated problem. The or- creating a National Science Foundation will be inadeganization is not one primarily established for research quate unless the Office of Naval Research or other and development purposes. It is not an organization Service-supported activities are, at least in part, conwhich has to worry about paying dividends to its tinued. Very roughly, the Navy has lately supported stockholders. It is an organization which, with the a $20-million research program out of a $200-million exception of the Office of Naval Research (and this is R & D budget. Thus we arrive at a figure of 10 per only about four years old), has never been really cent for Navy support of research. This would seem a concerned with research projects. Even a very high reasonable ratio. percentage of its developmental accomplishments I have pointed out above that the selection of the have been done by simply hiring outside organizations most promising research projects can be intelligently under contract to do the job. Methods suitable for the accomplished only by the working scientist. His selecselection of programs for development are often ap- tion of projects must necessarily be limited to a great plied to the choice of research problems. Research ob- extent by the facilities and financial support which jectives are defined and a program is planned in terms it is possible to accord him by his employer. The of time and money. This practice is not only useless; field of his researches will naturally grow out of those it is downright harmful. The best way to conduct re- fields in which he has had experience. The research search is to decide on the fields of investigation, to director to whom he reports has, naturally, the reexplore these fields, and to feel one's way into the sponsibility to select those broad fields in which the unknown, allowing competent investigators to fol- laboratory's work should be concentrated. The decilow the paths which in their competent judgment seem sion by the scientist as to what research projects he most promising. would recommend being pursued involves the calcuBut this willingness to allow self-direction (by lated risk of his choices, involving such complicated either individual or group) in broad fields of research techniques, either known or unknown, that a satisis a very difficult point of view to gain adoption in factory solution of the problem may require too large the military services. The desirability and, I might a fraction of his active life and hence probably result even say, necessity, of its adoption is nevertheless re- in slow professional advancement for himself. The apcognized in many levels of the command channels. How- proval of suggested research projects by the research ever, there still exists a great desire to tie research director in turn involves the calculated risk that he projects to developmental end-items. As far as I can will be able to maintain proper support in facilities ascertain, this tendency is closely related to a justifiable and budget until the answer to the problem is found. belief that in times of retrenchment the research item This requires a continuous close following of the prowhich is standing on its own feet will be the first to be gress of work by the director, so that he may be in a cut off. How true this is! On the other hand, during position to defend the support of the work whenever

17 Calculated Risk called upon to do so. In large research organizations in the Army Staff in administrative or advisory capacithis is oftentimes done by project supervisors. These ties and particularly those in the Army laboratories continued reviews involve a control in the sense that and stations, be civilians experienced in specific proa calculated-risk type of decision is frequently required fessions. With rare exceptions-and they do existin the way of recommending an increase of effort, a officers should be excluded from participation in the slight change in emphasis, or a decision to alter the supply organization because they will not have the method of attack. necessary competence in nonmilitary professions. Finally, and usually, the toughest decision, is the The conflict and inefficiency which has sometimes calculated risk involved in terminating a project. In arisen within the Services with respect to this second the very few cases where a unique answer to the orig- task, that of providing the combat forces with the supinal problem has been achieved, which in turn sug- plies they need, arises principally because of the insisgests nothing needing further investigation, this de- tence of the officer corps on immediate direction and cision is easy. In most cases, however, the matter does control of this task. The difficulty can be avoided and not turn out this simply. After a reasonable effort has competence can be assured within the Services with been expended and the question has not been answered respect to this second task by separation of the military to the investigator's satisfaction, one is confronted with and civilian functions and by assignment of the two the question of the calculated risk involved in drop- tasks to the personnel that is most qualified to carry ping the work altogether. This means usually a com- them out. However, such a separation should not be plete loss in the effort already expended, although in a separation in the sense that there will be two teams, a few cases a negative result may mean progress. There one military and one civilian, but rather a separation is always the alternative that, with a continuation from the point of view that the most competent person of the work for some time, useful information will be should be chosen to direct and execute each specific obtained. There is, of course, no formula by which function. The choice of the director for each organito decide such questions. The inclination of the pri- zation, and of his subordinate personnel, should be mary investigator is usually to continue, since a suc- made on the basis of competence and experience for cessful research man is always an optimist. In a few the particular task. Military and civilian personnel cases he may elect to stop the investigation due to must work within the Services on a basis of equality, discouragement, but research is always such a continu- each according to his abilities. ous series of discouragements that the proven success- It is necessary that the eventual user of materiel conful researcher rarely fails to emulate the bull dog-he trol the detailed decisions which will lead to produchangs on. Consequently, the research director, having tion, since this is materiel which will be used in comfollowed closely the progress of the work and having bat. It is necessary, therefore, to provide means wherethe background of successful investigational experi- by the military users will control. those decisions in ences himself, must usually be the one to decide when research, development, engineering, and production the calculated risk would indicate that further work which affect the emphasis of effort and the nature of would probably result in waste of manpower and the final product. The control should be from a center money. directly connected with strategic and tactical planning Getting back a little more closely to the subject of which maintain close liaison with the field commanders military research and development, let us examine who are most capable of estimating the practicability the present mission of the Army. It is so wide in scope of specific materiel items. that it covers not only an enormous military effort but The philosophy underlying the above discussion is also an enormous supply effort. The military effort based upon the assumption that military and civilians of the Army requires above all competence in combat, alike have an equal duty with respect to the security including the military aspects of logistics. This part of their country and must work as partners, each acof the task can be carried out only under the leader- cepting, and being given authority within, those reship of professional military personnel. The problem gions in which each possesses competence. In this of supply includes the efforts of numerous nonmilitary partnership the civilians must accept the role of junior professions, beginning with research and ending with partners and must welcome guidance by the senior production. The required competence and experience partners, the users of weapons and weapon systems. It in the professions involved can in most cases be found should be pointed out that this same philosophy has in civilians who have made these professions their life- been used by the British in the organization of the time careers. Military personnel cannot and should Ministry of Supply. not be expected to develop high competence in such There appears to be a slowly growing appreciation professions, unrelated as they are to their own pri- within all the Services, particularly advanced in the mary field of specialization. In order to carry out the Navy, that the scientific administration of military resupply part of the Army task, it is essential that the search and development laboratories must be in the personnel engaged in these activities, both those with- hands of competent civilian directors. These directors

Calculated Risk 18 must be distinguished professional civilians capable hand, we must not be indecisive and delay to the point of commanding the respect and services of competent where we lose the advantage. scientists and engineers and of maintaining a working Only men of science, and a few others, realize fully partnership with the military organization. Such tech- the extent to which we depleted our research stockpile nical directors will then be able to assemble a group during the war years and to what an overwhelming of professional personnel as highly skilled in their extent that stockpile had been built up for us by profession as are the military officers in theirs. In European scientists. All that we had learned in the either case a lifetime of specialization is required. years between the wars and all that we had approWhen such research and development laboratories are priated from foreign research agencies went into the located, as is often the case, on a large military post or developments which we accelerated to such an extent test facility, there is obviously the need for a military during the past war. We must replenish that store post commander to direct the administrative and of proven research. In short, he who now discovers housekeeping facilities and to ensure proper planning basically new facts of nature, the storehouse the door for the enormous nontechnical requirements of the of which has only been cracked slightly ajar; he who research and development laboratory. But the au- now develops the skills and techniques for designing thority and responsibility of the technical direction and employing new weapon systems far advanced over must be entirely in the hands of the technical director, those that are currently being put into procurement; he who should report directly to the Service Headquar- who now refuses to limit his vision to further only ters. Organizations somewhat similar to that indicated short-range development of presently conceived weapabove now exist at the Michelson Laboratory at Inyo- oneering gadgets- will most certainly acquire a prekern, the Naval Ordnance Laboratory at White Oak, dominant equity in the role of fathering the security the Naval Research Laboratory, and the Camp Detrick of his country by positively insuring the military superLaboratories of the Chemical Corps. iority over any foreign aggressor. Calculated risk in military research and develop- Let us never forget that those really priceless proment does not mean taking long-shot gambles with ducts of human imagination, those strong skeletons our national security. It does mean, however, that we of basic scientific theory which firmly support the must "make haste slowly" in our planning and conduct branches and twigs of technological advance are never ourselves in the manner of the astute chess player. As the result of over-all planning, or coordination, or of mature players at the international board we must channelized administrative procedures. They are the think out each move carefully. We must take long result of freedom of thought, time for thought, and looks before we move our pieces and we must never the establishment of a natural tradition that science be in the embarrassing, and possibly disastrous pogi- has been proved essential in the continued fight for tion of moving in all directions at once. On the other national security. DISCUSSION OF MR. LOUGHRIDGE'S PAPER DR. MAXFIELD: Mr. Loughridge has discussed the An evaluation must be made to determine the most question of calculated risk in connection with research promising solution to satisfy the requirement for a in the way the military establishment looks at it. We future weapon or future device of value to the armed feel sure that there is practically no risk in true founda- services. tional research. However, when we speak about de- I am not going to discuss this subject in detail; velopment, which is the subject that the two preced- however, I want to point out some of the factors which, ing speakers have been discussing, there is a great deal it seems to me, we have to consider in evaluating a of risk involved. project for military purposes when we have already In attempting to attain a certain objective, military done the foundational research and want to determine agencies usually must consider several projects of- whether a particular proposal is worthy of developfering possibilities of success, and they must then ment. decide which one shall be supported. All possibilities First, we must consider the estimated cost of this cannot be supported because, as Mr. Loughridge development, because we must stay within the limitapointed out, the Research and Development Board tions of the Research and Development Board budget. fixes a ceiling within which research and development Second, we must consider the development time inmust be conducted. volved, since usually we have target dates which have The problem frequently arises, then, to decide which to be met. of several promising developments shall be pursued. Third, we must consider the facilities available to

19 Calculated Risk do this work, and whether this work can be done with for you to picture or create a picture from which you existing facilities or whether it will be necessary to can determine the degree of risk or the value of the initiate a program for new facilities either by govern- project. Can you go further into that? ment contract or through a government laboratory. MR. VAUGHN: Well, in our own organization, and I Finally, we must consider the cost to manufacture think this is true in many others, any man has some and maintain the ultimate product resulting from the opportunity to explore ideas that he believes may have development. Since usually the products of military some ultimate benefit to our business. I think I can development are only going to be required in large say categorically that we do not sponsor deliberately number in time of war or emergency, we must consider and formally any so-called fundamental research work what the product will cost under wartime conditions which has no possible application in our business, but and how its manufacture will compete or interfere we do some of that kind of research in the early periods with civilian activities. Then we must consider the of some projects. cost of maintaining this equipment once it has been The men in the laboratory should be able to spend procured and issued to the fleet. We must consider some time, if they so desire, on a project without it the cost of keeping the equipment up and keeping being officially on the program. personnel trained in its use. In general, however, we try to keep such work within All these factors have to be considered in calculat- the confines of the area in which we are interested, etc. ing and evaluating the risk involved in undertaking I know that some companies have adopted a pera given development. When a requirement exists and centage basis for the allocation of time to such work. limited funds are available, we must ask ourselves, I know one organization (and a representative of that shall we undertake a particular project which will cost company is in this room) which permits its research two or three million dollars and may produce the de- men to put in 20 per cent of their time on anything sired result, or shall we spend that two or three million they like, and there is no restriction whatsoever on dollars along an entirely different but promising line what they do with that time. which appears to yield a similar solution, or shall we We have adopted no particular percentage figure, split it up among several projects and have several but I guess it is running around 10 or 15 per cent of competing projects going along at the same time but our research effort. at a slower rate? DR. SPENCER: Mr. Loughridge spoke about depletI think this illustrates some of the important factors ing our supply of basic knowledge. Such a statement which must be considered in any calculation of the gives a wrong impression. We do not, in any way, derisk involved in military developments. plete our supply of basic knowledge by using it. By MR. HUGHES: I would like to inquire of the first two finding uses for basic knowledge, we increase our unspeakers what methods they use to allocate moneys derstanding of it and enhance its value. for the research end of the deal as expressed in the What Mr. Loughridge meant, I believe, is that more last man's definition of research? I think we in the fundamental research, aimed at the increase of basic petroleum industry are accustomed to calling that knowledge, should be conducted and that technologiexploratory research. Do you have any schemes for cal research is not a substitute for basic research. Certhat kind of research? tainly, additional fundamental research is urgently MR. REEVES: I think that the allocation is more needed. on the basis of necessity. In other words, we look at There is one risk which none of the speakers menthese projects from the standpoint of what information tioned-the risk of not doing research. When one is needed, and, if it is basic information, then the calculates the risk associated with conducting a remoney goes to basic research. If it is pilot-plant work, search project, he ought to calculate also the risk of it goes into that. In other words, we do not draw a not doing it. Suppose that a new project or a new prodistinction between basic research and some other cess is needed. The new product or process will not kind of research, such as pilot-plant work and develop- be found if the research is not done. The risk of conment work. ducting the research may be high, but the risk of We have an end to achieve, and we try and figure not conducting the research may be much greater. out what type of information we need to achieve it, MR. NORTHRUP: I have listened with interest to and that is where the money is finally allocated. the beautifully simple and apparently reliable methods COLONEL KEIRN: I would like to explore that ques- of calculation of risk presented by the first two speaktion a little further. I noticed in Mr. Vaughn's speech ers. I should like to raise a small voice in the wilderthat each project, before the research staff was able ness of regimented research administration to the to present it to management, had $3,000 or $4,000 or effect that very few nonhypothetical research projects something of that sort spent on it. There must be a ever follow such simplified charts. percentage of profits or something that can be ear- Really significant research results often obtain from marked for these explorations before it is possible projects initially started for entirely different purposes

Calculated Risk 20 or for no specific purpose at all. I am reminded of Mr. CAPTAIN FURTH: I would like to support Dr. Gerald Johnson's discussion of the projects started by Loughridge's statement regarding our having used up President Roosevelt in the New Deal era. In his book, the fundamental knowledge which was available prior The Incredible Tale, he cites many projects which to World War II. Franklin Roosevelt started for one purpose and from From the military point of view, he expresses his which, entirely by accident, he achieved more im- concern over the value of weapons which have been portant and wholly unpredicted results. developed and which make use of techniques which are Mr. Johnson compared this phenomenon to the well known. The value of the weapon is measured by tales of Hugh Walpole of the three Princes of Seren- its effectiveness; it remains effective only as long as dip (ancient name for the island of Ceylon). These the enemy is unable to deny us its use by countertales were always characterized by some entirely unex- measures. So, by having used up our fundamental pected and delightful experience encountered by the knowledge, the techniques which are now available princes during voyages originally planned for other and which were developed prior to and during the and usually less interesting objectives. last war are also available to probable enemies. Thus the term "serendipity" has come to mean To have effective weapons, we must keep exploring "getting more by accident than you do on purpose." new fields, looking for new techniques, with the hope MR. VAUGHN: Could I say that the principle you that it will take an enemy a long time to find proper just discussed is one of the factors of safety in your countermeasures and thereby deny us use of these calculation? It helps you a great deal sometimes. weapons. ROUND-TABLE DISCUSSION PAUL D. FOOTE, Chairman The use of charts illustrating calculated risk, the when economic prospects are uncertain. estimated expense, and the possible returns from a The tendency to avoid long-term projects because research project are of primary advantage in the sell- of the economic uncertainty existing today may, in the ing of research programs to management officials. A long run, seriously retard our scientific progress. It is director of research, in many cases, sells a research often possible that these long-range projects present program to his company, and since the results of the a field that logically might be most advantageously research program might well affect the future of his supported by the Government or by cooperative efforts company, the responsibility in properly evaluating in the laboratories of the universities and colleges. the factors involved in the performing of the research The risk involved in failure to conduct research and the utilization of the results is of great magnitude. was also discussed, with the petroleum and chemical One of the very important considerations in this day industries cited as examples of industries spending of rising costs is the financial capability of the company large sums for research, particularly in pilot-plant to exploit the results of the research project upon its construction and operation. Here the risk involved in completion, and with long-range research programs going directly from laboratory to full production is the evaluation of the financial element becomes in- considered too great. It was estimated that 75 per creasingly difficult. cent of petroleum research money gdes into pilot-plant Today's conditions necessitate a large part of the construction and operation. research budget going for those projects which tend The current world situation has injected a further to lower production costs and reduce capital expendi- element of risk into the calculations of the industrial tures. Such projects are largely short-term efforts research director, namely, the problem of Government quickly put into use; their effect may be estimated ac- research. In the general discussion, the following points curately. As a rule, the investment and cost is com- were raised with regard to Government research in paratively low, and the return on the research invest- industrial laboratories: ment is not very high. The very life blood of an in- 1. Large Government contracts disrupt industrial dustry, the projects that will produce the new pro- laboratories by forcing temporary staff expanducts which keep a company in business over the years, sions, with reductions in force upon terminasions, with reductions in force upon terminaare the long-term research projects involving many elements of risk. They require a careful analysis and ton of the contracts adversely affecting worker a strong selling effort on the part of research directors. morale. However, these are the projects which are neglected 2. When the Government projects require the

21 Calculated Risk industrial laboratory to go far afield from the It was estimated that in the next two years 75 per interests of the parent company, the termina- cent of every research dollar will be spent for military tion of the Government work often finds the purposes. It was recommended that if the Government company and its products in poor competitive were going to provide an increase in research and deposition due to lack of continual product im- velopment money of such magnitude, an accurate provement. survey of the men, facilities, and the effects of the 3. The present method of administering Gov- military draft upon the various laboratories should ernment contracts is so involved that some be made at once if this vast sum is to be expended laboratories prefer not to accept them. to the best advantage.

Second Session Henry M. O'Bryan, presiding Research and Development Board, Washington, D.C. MEASURING THE RETURN FROM RESEARCH — I by ALLEN ABRAMS Vice President, Marathon Corporation IN THIS DISCUSSION we are concerned primarily with value of science that by 1920 there were 300 industrial the evaluation of organized industrial research. Some research laboratories employing approximately 7000 directors of research may think it unnecessary to at- scientists and spending $30,000,000 per year. In this tempt the justification of a function which consumes year, 1950, there are ten times as many laboratories such a minor portion of a company's expenditures. as there were in 1920, about 20 times as many people Other directors will feel that in their own organiza- engaged in research, and the expenditures are 30 times tions there is quite full recognition of research as as much, or nearly one billion dollars. the basic insurance of the business. Industrialists do not spend money without good Yet there are good reasons why an evaluation of reason. It is, therefore, interesting to note some of their research is desirable. In an industrial concern research comments on research in the annual reports to stockis a parallel function to production and sales. The holders. One of the large chemical companies, successproduction man can point with pride to the number ful in introducing many new products, states: "The of units he has turned out. The salesman can present company believes that an aggressive and forward-lookan impressive report of his accomplishments. Through ing research program is the dominant factor influenchis operating statement and balance sheet the manager ing the sound growth of business." A large electric of a business is always cognizant of the money being company has an ambitious program for its research poured into research. Yet often he has no adequate department, "providing the means of turning more gauge of what flows out, so that the research man re- energy to man's advantage... is the end result of our ceives little or no recognition of his contribution. research." The pioneer company in photography says Industrial research laboratories are incubators of that its research laboratory is "responsible for the ideas-converters of thoughts to things. But without future of photography." An oil company headed by some measuring stick, research men, particularly in a chemical engineer believes that "In all our fields younger organizations, may develop a sense of frustra- of activity, research continues to bring greater efficiency tion which is good neither for their own morale nor to our operations and better products to our custofor the well-being of the company. Many directors mers." A paper manufacturer thinks that "constant of research would be happier, and more companies work in research and development is the best assurmight invest in research if they had means of evaluat- ance of the continuity of our business and hence the ing the results. For even among the blue-chip indus- security of employees and management, stockholders tries listed on the New York Stock Exchange there are and customers." notable omissions of research departments. Another measure of the value of research is in the There are a number of ways by which we may judge industries which have grown out of its findings and the value of research. The simplest of these is to note in the jobs which have been created by its application. the growth of research organizations in the United Consider a list of 1700 of our large industrial comStates. In 1900 formal research was practically non- panies. Nearly one-half of these were either nonexisexistent, but World War I demonstrated such potential tent or were infants in 1900-aircraft, automobiles,

23 Measuring Research Returns chemicals, electrical equipment, petroleum, and rub- or greater volume, the increment only is credited to ber. It is not pure coincidence that in half a century research. the following changes have come about. Jobs have While not entirely satisfied with its method, a cerbeen created 25 per cent more rapidly than our popu- tain lumber company classifies its research results into lation has grown; industrial production has increased three groups and rewards them accordingly: (1) Prothree times as rapidly as population; industrial output jects which have been concluded successfully and put per worker has multiplied by four times; today the into operation. It is assumed that the value of these United States consumes nearly one-half the energy discoveries is four times the cost of the research. (2) output of the world. To a large degree it is the findings Projects which have been completed successfully and and applications of research which have caused these are of definite value to the company in improving opchanges. erations but on which no well-defined action has been Another, and for our purpose the most pertinent taken. An example of this might be an investigation way of evaluating research is through the procedures which has kept the company from spending money on which have been worked out by various companies. a questionable process. Projects of this sort are assumed Our discussion here is based largely on information to be worth twice the cost of the research. (3) If the obtained from about one hundred industrial research project is unsuccessful, obviously, no credit is taken. directors. A few of these expressed lack of interest in One manufacturer has developed a detailed prothe subject and questioned the desirability of further cedure of evaluation. The research and development investigation. But one scientist expressed the view- department submits proposals on new products to the point of nearly all the respondents; he said: "Support factory manager. When the idea has been accepted of a research department is more an act of faith than and put to use, the accounting department studies it is a numerically calculated risk. However, any act the results and applies the "Index of Return." This of faith should be appraised sooner or later." index is composed of the savings for one year obtained Several research directors point out the desirability through an improved process; 3 per cent of the net of preliminary appraisal of projects in terms of ex- sales on a new product for a period of three years; 3 penditures required and the likelihood of successful per cent of the net sales on an improved product for returns. If such a study casts doubt on the project, then one year. The company believes that this index measit should probably be abandoned. Several companies ures not only the effectiveness of the research departclassify their programs primarily as defensive research, ment but the ability of the factory to accept and debelieving that the company would not be in business velop new and better products and the ability of the at all save for its research department. In the older sales department to sell these products. companies with well established research laboratories, A large oil company evaluates its research about management appears well satisfied with the effective- every five years. The principal benefits arise from: ness of research, judging in most cases by the general (1) elimination of royalty payments, which are easy health of the company. In one case this confidence in to estimate; (2) added capacity due to improvements the research department is so implicit that the phi- in existing products and processes, representing a savlosophy is to sell a product first and then hope that ing in capital investment and lowering of operating it can be made, rather than make the product and costs (another advantage comes from increased bushope that it can be sold. iness resulting from such improvements in quality); Some companies have developed formalized methods (3) additional profits from new products. for crediting research with its accomplishments. In Another oil company measures the results of research all cases the research directors point out the difficulties as follows: On new and improved products, research of such an appraisal and caution research not to claim is credited with the total profit in the best year of a too much credit. They believe that the evaluation can five-year period immediately following the initial marbe carried out most satisfactorily by having represent- keting of the product. On new manufacturing procesatives from production, sales, accounting, and research ses, research receives credit for the profits of one year. make the final appraisal. On improved manufacturing processes, savings in cost A large chemical company uses the following pro- are credited to research for a period of one year. On cedure: When the research department has been re- patents acquired through research, all royalties or sponsible for most of the original studies, including sales of such patents are credited to the department. application and process development, all future sales A manufacturer of paper products totals separately of the new product are credited to research. If, how- the annual sales and profits of products derived ever, the product comes to research partly worked out, through research and of products which are not so dethen the department completes the job and takes as rived. The net profit on research products is then credit to itself only some reasonable percentage of compared with that on nonresearch products, and the the sales. When a major improvement has been de- profit increment only is credited to research. This inveloped on an old process, resulting in higher yields crement may be divided by the cost of research for

Measuring Research Returns 24 the year, yielding a figure showing the dollar return of about $10.00 for each research dollar used in proper dollar expended for research. At this time credit ducing that amount. is not taken for process or product improvement nor These figures are significant, for a company is for royalties on patents acquired through research. thought to be doing well if it shows a profit of $0.20 The tangible returns from research expenditures for each dollar expended in producing that profit. Yet vary widely, dependent to a considerable extent on it should not be assumed that these are typical of all the effectiveness of the research department in push- industrial-research earnings. Nor must it be forgotten ing its ideas through to successful completion. Per- that the law of diminishing returns holds good in haps even more important is the difference in the evaluating research, else there would be no limit on formulas which are used for determining these returns. expenditures for this purpose. For example, one company has a committee which In the ultimate analysis posterity will judge research decides how much royalty can be charged on the cost not alone by the dollar sign, nor by the great indusof a new product and then credits this to the research tries research has created, nor by our improved madepartment. This company states that the successful terial standard of living. Profit is the immediate and developments more than pay for all the cost of the essential incentive for industrial research, but there research department. Another manufacturer indicates are far broader considerations than this. The findings that its research credits for a given year were about of research may be applied in making fiber to clothe equal to its expenditures. Since this is a very successful us, or to produce a deadly form of warfare; a drug company, it would appear that the formula is not may be used to heal the sick, or to inflict pain on the predicting the proper returns of the research de- healthy. Atomic fission, born in the stress of war and partment. first applied to the destruction of human life, may During a period of seven years, one oil company become the greatest benefactor of mankind. showed annual results ranging from a profit of $1.45 Research has extended the horizon of our knowledge to a loss of $0.25 on each research dollar, with a seven- until at last we have reached a no-man's land borderyear average profit of $1.35 per dollar expended on ing on the infinite. Yet, with all this great scientific research. Another oil company estimated a yield of advance, have we made the human race better or hap$15.40 for each dollar of research expense. This sum pier than it was fifty years or a thousand years ago? consisted of $3.70 savings in royalties which would It is just possible that we have been better scientists otherwise have been paid to outside groups, $9.60 for than citizens. Now more than ever we should labor to profits resulting from process and product improve- direct the results of research into constructive use. If ment, $2.10 for profits from new products. In one we succeed, that success itself will be the best measure typical year, a paper company estimated a net return of the return from our research. II - by W. S. PARSONS Rear Admiral, U. S. Navy It seems desirable to introduce the Weapons Systems probability considerations and the theory of error enter Evaluation Group (WSEG) to you first by setting... the scientist can encourage numerical thinking on forth the problem, then indicating the mechanism pro- operational matters, and so can help avoid running the posed for solving this problem, and finally by giving war by gusts of emotion..." a necessarily sketchy account of the manner in which A man from Mars might believe that several years the Group has actually operated in its one and one- after World War II, when the dust had settled and half years of existence. complete analyses could be made of what happened Concerning the scientific evaluation of weapons to all operators in that most extensive weapons laborasystems evolved in World War II, I shall quote the tory in all history, we would enter a period of full and following from Dr. P. M. S. Blackett of England, who "final" agreement on the question of how mid-20th is reputed to be the father of operational research: century wars are fought and won. In actuality, the "Many war operations involve considerations with which clouds of dust, atomic and otherwise, seemed to rise scientists are specially trained to compete, and in which higher as World War II receded into the past. From serving officers are in general not trained. This is especi- the technical standpoint this was mainly caused by ally the case with all those aspects of operations into which the fact that snorkel submarines, guided missiles, jet

25 Measuring Research Returns aircraft, and atomic weapons appeared in the last year techniques and choosing new ones. It is especially impor-some in the last days-of the war. So it is quite under- tant... that careful thought be given not only to atomic standable that by 1948 Dr. Bush, as Chairman of the and bacteriological warfare but to how far it is safe to Research and Development Board, received from his go in dispensing with pre-atomic tactics strategy. This is och d C tevelofement s, rced tom c the reasoning behind the creation by... Secretary ForAd Hoc Committee of eminent scientists, called to con- estal of the Weapons Systems Evaluation Group..." restal of the Weapons Systems Evaluation Group... " sider the problem of overall weapons systems evaluation, a report which included the following: tion, a report which included the following: The Christian Science Monitor declared that Secre"The problem of national security is so serious that tary Forrestal acted "boldly, constructively, and imagevery reasonable effort must be made to get the best an- inatively in creating the Weapons Systems Evaluation swers by the application of the scientific method. The speed Group." with which the technology of war has changed and will The Baltimore Sun observed: change; and the complex interactions of technology with "... The Joint Chiefs of Staff are weighed down with tactics, strategy, and logistics in total war are such that administrative duties which allow them little time for conmilitary judgment alone is not enough. If we are to ex- templation. Each of them also has certain complex instruploit fully the power of modern analysis, scientific methods ments at his disposal. Hence it follows that they think in for examining the nature of future war should be employed terms of doing a given job with the instruments already in the broadest possible sense." at hand instead of trying to find out what instruments The need for weapons systems evaluation was voiced would be best and how they are to be obtained.... If the beyond official documents. Foreign Affairs contains the problem were attacked first by determining the best instrufollowing paragraph in an article called "Science and ment for doing the job, it seems probable that the Politics in the Twentieth Century," by Dr. J. B. Con- differences... would be reduced in importance. That is ant: the idea behind the new Weapons Systems Evaluation "By what procedures are a free people to determine the Group answers to such complex questions as to whether a large The Kansas City Star noted: amount of the taxpayer's money is to be spent on the "... the announcement that a Weapons Systems development of a given weapon or its auxiliary? Granted Evaluation Group has been formed is.. one of the few the matter must be left to the people's elected representa- really important milestones in the nation's continuing effort tives and the President exercising through subordinates his to make itself secure... the army, the navy, and the air power as Commander-in-Chief, nevertheless the problem force in the past have given their attention to it sepastill remains, how are politicians to resolve conflicts of rately. The result has been much duplication and even opinions among scientists and engineers? Have we devised some working at cross-purposes.... Our people... will as yet even the first approximation to a satisfactory pro- be looking to the services for more of the solid assurance cedure for evaluating technical judgments on matters con- that can come from such actions as the centering of responnected with the national defense, including atomic energy? sibility in the Weapons selection unit." Some who have been close to the postwar scene in Washington and have followed some of the research and develop- Lieutenant General Hull, who was General Marment projects must be inclined to answer this question in shall's Assistant Chief of Staff, Operations Division, the negative." during World War II, and who in 1948 commanded James Forrestal, the Secretary of Defense, was in full Joint Task Force 7, which tested three atomic bombs agreement with Dr. Bush as to the need for weapons at Eniwetok Atoll, was a unanimous choice for the systems evaluation and concurred in the establishment first director of the Weapons Systems Evaluation of such a group by the Joint Chiefs of Staff and the Group. His deputy and research director for the group Research and Development Board. The document es- was Dr. Philip Morse of the Massachusetts Institute of tablishing the Weapons Systems Evaluation Group Technology, one of the founders of American antistates: submarine operational research in World War II and "The purpose of the Group is to provide rigorous, un- first director of the Brookhaven Laboratory of the prejudiced and independent analyses and evaluations of Atomic Energy Commission. Dr. Morse was relieved present and future weapons systems under probable fu- in June, 1950, by Dr. H. P. Robertson, eminent matheture combat conditions matical physicist from the California Institute of TechI think it will be apparent from what I have quoted nology, whose wartime scientific contributions earned so far that the concept of evaluation of major weapons h;m the admiration of all three services. My colleagues and systems was fully accepted in what might be as senior service members of the Group are: Major called the inner defense circles. The reaction in the General Gavin of the Army, of World War II parapress was almost embarrassingly good. As samples of troop fame, and Major General Barnes, with a distinthe press reaction, I give the following: guished record in the Air Force. Dr. Robertson has a The Washington Star commented: civilian scientist, Dr. Welch, as his deputy and, as a "With methods of warfare undergoing revolutionary rule, the project leader in each evaluation is a civilian changes, it has become necessary to use the greatest pos- scentst. sible wisdom in discarding conventional weapons and The operational research groups of the several serv

Measuring Research Returns 26 ices and such groups as the Air Force Project RAND place rocks as weapons in primitive warfare. In those at Santa Monica have been of great assistance to prehistoric times such a decision was evidently a "horse WSEG in its analyses and evaluations. The same ap- back" decision involving some energetic operational plies to university laboratories engaged in weapons de- tests but probably no operational research as we use velopment, such as the Los Alamos Scientific Labora- it in the Weapons Systems Evaluation Group. In the tory in the case of atomic weapons, and the Johns past there has always been a reluctance on the part Hopkins Applied Physics Laboratory, in the case of of the military to accept new weapons until they have guided missiles. WSEG borrows key personnel from been tested in battle, and even then new weapons had these research activities, and the exchange has been to be "sold," so to speak, before they were accepted. found beneficial both to the parent group and WSEG. Lieutenant General Hull is my authority for this As WSEG was being formed, sage advisors urged history of introduction of the machine gun in the U.S. that we cut our teeth on one or two fairly easy prob- and British armies. The first machine gun was inlems. This was not to be our luck. One of the first vented by the Gattling brothers in 1859. They tried problems handed to us related to strategic air warfare. to get the Army to adopt their gun but without sucBy August 24, 1949, General Bradley, Chairman of cess. All through the Civil War the Gattling brothers the Joint Chiefs of Staff, testified as follows before the followed the Union Army around trying to anticipate Armed Services Committee investigating the B-36 when a battle would occur where they could demonbomber program: strate their gun by firing at Confederates. It wasn't "Insofar as we could, we believe that all weapons of a until 1866, after the war was over, that the Army finally major type and of a major importance should be evaluated adopted it. They did not make use of it in their by an evaluation board, consisting of military people and military forces until some time later. As a matter of scientists and there has been a board set up for this pur- fact, the British Navy adopted it before the British pose." Army did. The problem of evaluating weapons systems in The realistic evaluation of ground weapons and peacetime is considerably different from wartime opera- close air support now going on in Korea naturally tional research. In war the time scale is fairly obvious modifies and simplifies some of our problems. Howand if combat reports are accurate there is not much ever, in several of the major weapons fields the effect risky extrapolation to be done, particularly if the of the United Nations reaction to Korea has been to enemy is sufficiently hard pressed so that he may be intensify the rearmament effort, with a corresponding assumed to have laid his available weapons cards on need for evaluation before committing ourselves to the table. By comparison, weapons systems evaluation billion-dollar programs. In other words, some of the in peacetime is much less definite. Even to get reliable peacetime budget brakes are now off, and the responsiinformation on our own weapons and systems requires bility to be right is correspondingly greater. extensive travel to establish contacts and get first-hand One field in which development is extremely imtechnical and operational knowledge. In our war portant and expensive is that of guided missiles. These games we must pit our known weapons against as- differ from manned aircraft, radar, etc., in that the sumed counterweapons which may be better or worse more successful the missile the sooner it will reach the than our best but will certainly be different from ours. full-scale hardware stage in flight tests, and from then From our vantage point close to the center of techni- on expensive crashes will pace progress. While the cal weapon development and operational analysis, magnitude of the missile development absolutely reWSEG is in a position to assess the value of applied quires as much simulation and calculation as possible, research in many technical and operational fields. The it also brings in a hazard which may be expressed as need to see through and around each military weapons follows: We thoroughly simulate and impose all of problem is more urgent certaintly in peacetime than the diffiiculties we know about, but we pamper the any similar need in industry. The reason is that most model in the calulator by neglecting to impose any industry is in "combat" in peacetime and therefore of the difficulties we do not know about. These are has the advantage of knowing how the game came out some of the reasons why we must have focused on the each year. A further advantage in industry lies in the missile problem many minds which have been kept fact that the scale of the budget can be determined on sharp and resourceful by basic and applied research. a fairly rational basis, whereas in defense planning the Measuring the return from research from the standvalue of victory and cost of defeat are not predictable point of the Weapons Systems Evaluation Group is on a quantitative basis. hardly a quantitative operation. Qualitatively, we Need for formal evaluation of weapons has develop- can say that the growing fund of technical and operaed with the machine age in the 19th and 20th centuries. tional knowledge bearing on weapons and systems is a It is doubtful that operations analysis or higher mathe- direct product of applied research in many fields. Also, matics of any kind were used before the decision was those scientists and engineers who have become vetmade to enter production of bows and arrows to re- erans in these new fields of applied research have

27 Measuring Research Returns thereby qualified themselves to work in the more com- Our experience in WSEG to date has been highly plicated game played with a crystal ball in one hand encouraging. We have had the utmost cooperation and fairly accurate proving ground and operational from the military services and from the civilian groups reports in the other. we have visited and with whom we have dealt. They As in all advanced fields of endeavor, in order to get even failed to make the usual screams to high heaven results it is necessary first to have good men and then when we gently suggested that we needed to borrow to organize around the problem. The organizational some of their best people. Also, in spite of the difficulpart is not difficult. Inside WSEG the color of the ties of wrenching a problem and concept into a form uniform and the differences between military and amenable to evaluation, we have normally succeeded civilian points of view tend to resolve themselves in doing this by the process of what might be called easily. But the problem of getting first-class civilian "approximation by argument." I might add that these talent is complicated by the idea held by many scien- arguments have very seldom been settled by Blackett's tists, and to a less extent by engineers, that the closer "gusts of emotion." they get to the center of the Pentagon, the more stifl- To summarize and conclude: From the standpoint ing will be the atmosphere dominated by the so-called of weapons systems evaluation, the measure of return "high brass" and "military minds." I think that one from research in this anxious decade will be the degree of the returns from our research to date has been the to which it increases our national stockpile of flexidemonstration that civilian scientists can operate ef- bility, resourcefulness, and alertness, both in the hufectively in such an atmosphere. man and material fields. - III by C. G. SUITS Vice President and Director of Research General Electric Company Last month I attended a series of company meetings turing, engineering, or research. Although it is true at Association Island on Lake Ontario in which each that most successful products draw fully upon all these functional activity of our company was reviewed, start- industrial functions, the problem of measuring the ing with research and proceeding through engineering, contribution of each of them separately is not easy. Yet manufacturing, marketing, as well as the principal this problem has been studied many times in the course administrative activities. These meetings provide, in of the history of our laboratory, and I can summarize our company, an indispensable source of integration my talk at once by saying that in our case we have and balance, and, incidentally, illustrate to an excel- never found an objective and complete solution of the lent degree the problems inherent in measuring the problem of measuring the return from research. My return from research. The proponents of each of these comments will, of course, apply to a research laboraactivities of an industrial enterprise feel, with proper tory in industry and, more particularly, to the laboraenthusiasm, that theirs is the most important com- tory in my company with which I am most familiar, ponent of the whole. For example, we in the labora- which has long-range exploratory research as its pritories feel that without our pioneering scientific work mary function. As a necessary consequence of this the corporation's future would be in doubt. We must specialization in long-range research, we depend admit, however, that some of the output of our labora- heavily for the practical utilization of our work upon tory is hardly ready for market in the form in which development laboratories and engineering-design diviit leaves our hands and that a great deal of rationaliz- sions, where the application of science is the primary ing and reducing to practice must be accomplished task. The contributions of many groups of specialists by engineers as the essential next step. is thus generally involved in the successful life history The manufacturing people, in turn, are quick to of a new idea. point out that no matter how well conceived and en- I propose to discuss this question in terms of a gineered, unless the highest degree of manufacturing variety of case histories of actual developments, which skill is applied to the product, it will never reach a will serve to illustrate the complexity of the problem, satisfactory market position. Finally, the marketing without, however, leading to any very workable means groups say, with some justification, that without skill- of measuring the return from research expenditures ful selling there will be little or no need for manufac- in a specialized industrial environment. I hope, how

Measuring Research Returns 28 ever, that we may obtain a better understanding of the education and career of a new baby and is fraught matter from a consideration of actual examples. Ob- with as many uncertainties. It would be better to wait viously, the character of the industry in which the until something is known about the child's personality, company operates, the size and kind of business in intelligence, and character traits as basic factors upon which the company engages, and, finally, the kind of which to plan. laboratory operation-all have a great deal to do with This is not to imply tlhat planning of this type of the question. Hence the conclusions drawn from case research is not possible. One can plan the scope of histories may not suggest generalizations at all. In fundamental and exploratory research and can relate specific cases the return from research may bear little the scope of the work to the scientific opportunities relation to the cost of research; the return may in some presumed to exist in the field of work and to the plans cases be very much greater than any properly allocated and aspirations of the company in relevent industrial cost, or, in other cases on which I would not like to areas. Although it is frequently true that the detailed dwell, very much less. results of exploratory research are unpredictable, some Now, measuring the return from research is gener- of the gross results are not subject to the same uncerally desired for reasons of good business, for, if we tainty. For example, some work in our laboratory can accomplish this result, we can judge whether our pointing toward new methods of interrupting electric expenditures have been made wisely or whether they arcs is more likely to lead to results of importance in should be greater or less than at present. Another the field of switchgear than in polymer chemistry, alequally important, related question is the prediction though I know of a case in which the opposite was of the future value of the results of research, so that true. Planning for the future of such work must be this may, in turn, be compared to the cost of research done on a much more expedient basis than in the case in progress. of engineering development work. The director must Everyone would, of course, like to be able to ac- expect to do a good deal of playing by ear, bearing complish this admirable result with precision, for it in mind that new experimental results may, without would provide the final answer to many important warning, completely alter the complexion of the work questions of research planning, effort, and expendi- for better or for worse. tures. Because the term "research" covers a great variety At the other end of the scale of research is the purely of technical activities, it must be admitted that some exploratory work in the frontier areas of modern types of research activity may be susceptible to detailed science: research at temperatures near absolute zero planning at every stage, including the prediction of involving superconductivity and superflow, investigareturn. tions of the phenomena of semi-conduction in solids, Consider, for example, the hypothetical case of the exploration of nuclear reactions, to mention some curlaboratory activity, principally of an applied engi- rent examples in progress in our laboratory. Since the neering character, involved in the substitution of a results of exploratory research are thoroughly unprepressure die casting in a product which is now being dictable, the application of planning methods which manufactured with a fabricated part. The circum- require a detailed knowledge of the time and cost of stances may be such as to permit the anticipation of a doing the work and the possible application of the reconsiderable saving in manufacturing cost if a die sults in terms of products and markets involve uncasting can be employed. The cost of doing the neces- knowns at every point. Planning the career of the resary laboratory, engineering, and pilot production search project in the case of truly exploratory work cannot be done in the definite terms with which one work may be estimated with satisfactory accuracy, as may also be the expected manufacturing cost savings. may deal in the case of applied reasearch or engineerSome of the expected savings may permit price reduc- ing development. In spite of the fact that principal tions and a consequent expansion in volume, which information on which planning and forecasting demay be determined by a market survey. In other words, pends is generally not available in the early stages of enough information is available or obtainable to make exploratory research, engineers frequently attempt to possible an estimate of the cost of doing the work and plan for the future results of the work. I feel that the return which may be anticipated from this particu- forecasting in such cases, for example, estimating the lar example of applied research. With such facts, or cost of a new chemical product from very incomplete reasonable facsimiles, fairly complete planning of re- early research data, frequently does far more harm search effort may be feasible. than good. Once numbers are obtained, one is likely Everyone is familiar with the way in which pre- to associate with them a degree of reliability that is by liminary estimates, particularly cost estimates, acquire no means justified by the assumptions upon which an aura of exactitude with the passage of time, fre- they are made. quently to the embarrassment of the estimator. Plan- THE CASE OF THE SILICONES ning the commercial future of an exploratory research I would now like to consider a relatively simple case project in its early stages is analogous to planning the of a research project on which a priori one ought to

29 Measuring Research Returns be able to measure the return from research. I refer to are difficult to make. We use silicones in many types the silicones on which our laboratory has been work- of equipment with beneficial results. ing for quite a number of years. The initial exploratory Let me give you one example. Silicone rubber is work was instigated largely by a consideration of one compatible with pyranol, an important insulating of the primary weaknesses of organic materials, the dielectric liquid, so that we have, for the first time, an carbon-to-carbon linkage. It is the failure of this bond elastic gasket for pyranol-filled equipment. It has diswhich accounts for the long-time thermal decompo- placed all other material for sealing pyranol capacisition of organic materials at a temperature of about tors at an attractive annual saving in manufacturing 105~C; the carbon-hydrogen bond is much stronger. cost. Although the saving in cost is important and The silicon atom is similar to carbon in much of its readily calculable, it is probably not as important as chemistry and is not too much larger, so that silicon- the improvement in the product which it permits and substituted organic compounds were possible, and a the wider market which may thereby be served. few such materials had been synthesized. Although it Considering now the rapidly growing silicone chemiwas recognized by analogy with the numerous organic cal business, it is probably too early to calculate the family that a very large number of compounds of sili- return in this direction. We have made large capital con, carbon, and hydrogen must be possible, it was investments in this business and will presumably make the potentialities of high-temperature insulating ma- more as it grows, all at an uncertain but ever present terials which were sufficiently attractive to justify an business risk. In any event, the return from this reextensive investigation. It is important to point out search will probably never be known. Without rethat it was not the sale of silicone materials as such search there would have been no silicone business. which provided the incentive for this work. It was But the same statement applies to the other industrial rather the prospect of very greatly improved perform- functions, without which the project could not sucance of electrical machinery which silicone materials ceed. might make possible. This project was carried through Finally, the prediction of the return from this reall of the stages of laboratory and pilot-plant research, search, during its early progress, would obviously have and it finally led to the building of a silicone manu- been subject to gross error because of the many reasons facturing plant at Waterford, New York. Concerned which are readily apparent. with other developments, such as formex and glyptal, it became the basis for the Chemical Department of THE FORMEX DEVELOPMENT our company, which was set up in January, 1945. Since we both manufacture and sell new materials Although the early expectations concerning thermal as well as apparatus which uses new materials, we stability have been well borne out by experimental have a double reason to be interested in the synthesis results-the Si-C bond is stable at temperatures in the of new materials. I would like to describe a developrange 170C to 200~C-the silicon analogues of the ment in which the effect of a new material on our organics have been full of surprises. For example, products is far more important than the sale of the nearly all members of the family have small thermal co- material as such. I refer to formex, an organic product efficients of mechanical properties. The oils have a which satisfies to a superlative degree the requirements small thermal coefficient of viscosity, and the elastomers of an insulating enamel for copper wire. Prior to the a small thermal coefficient of elasticity. Many silicones development of formex, oil-base enamel was used for have astounding antiwetting properties. Others have this purpose, primarily because it was priced at 150 a remarkable antifoaming performance. As these un- pound; it was not and is not now a very satisfactory usual properties have come to light, they have been material for this application. The technical requireput to use so that at the present time, well over half ments of an insulating enamel are very elaborate, but of the output of our silicone plant is going into uses it is principally in mechanical strength that formex that were not foreseen by early research workers on is outstanding; it is undoubtedly one of the toughest the project. The largest surprise of all has been, of organic coatings known. Although formex is much course, the fact that we had in hand not only im- more expensive than oil-base enamel and is more diffiproved high-temperature materials, but that we had cult to apply to wire, it has practically displaced the a very large new family of chemicals with unusual latter. Ninety per cent of all wire enamel used in this and versatile properties. Although the silicone bus- country is formex. The reason is that the toughness of iness is still in its infancy, it seems abundantly clear the formex film makes its use possible where ordinary to us that it is destined to become an important seg- enamel will fail. It can be wound by automatic mament of the chemical industry. chines at high speed without damage to the integrity of The return from this research must be sought prin- the insulation film. cipally (1) in improvements in products which use Because its electrical strength is also excellent, it can the new materials or (2) from the sale of the materials be used in a thinner film for a given electrical stress. themselves. In each of these directions precise estimates To provide a given number of turns in a motor or

Measuring Research Returns 30 transformer, therefore, less space needs to be allocated experience, were far more upset if there was to be any to insulation. This reduces the size, weight, and cost appreciable delay in repairing or replacing their blanof the magnetic core. Thus, the ultimate gain in the ket. The obvious enthusiasm of users of electric blanuse of this material is in the design of electrical equip- kets, even under the somewhat difficult circumstances ment. Probably the most important single factor in I have described, convinced us in the laboratory that the small size and inexpensive construction of modern there was much merit in the idea, and we promptly small motors and transformers is formex. The poten- set about to interest our engineers in the results we tial market for wire enamel is not large; it is a specialty had obtained. The danger of the rolled-up blanket chemical product of limited application. However, was solved by placing a number of tiny safety thermoimprovements in wire insulation have a far-reaching stats in the blanket, which open the electric circuit effect on the cost of electrical machinery, and it is in if the temperature rises excessively. The latest design this area that the important results of the research of electric blanket has a far more elegant solution of appear. If the costs of the research which led to formex this problem. are allocated against the end product, which is the It did not take long to interest our appliance engiwire enamel itself, the sale of the improved material neers in the electric blanket; but everyone recognized readily permits a liquidation of the cost of the research. that the very novelty of the idea would necessitate a However, no accounting system we have devised considerable period of public education before one measures the great importance of this development to could hope for any general acceptance of the idea. the design, manufacturing methods, and hence the The development of a new electrical appliance is frecost of electrical machinery made possible by the for- quently preceded by market surveys and market testmex development. Hence the return from the research ing. in this case is apparent and satisfying, but its measure- But a market survey in the case of a product as new ment, as that term is generally understood, is a matter as was the electric blanket in 1936 would have been of great complexity. useless; any housewife confronted by the question, THE ELECTRIC BLANKET "If an electric blanket were put on the market at a I would now like to describe a project of our labora- certain price, would you buy one?"-could only have tory which I hesitate to call research. In this case the given the questioner a puzzled expression. She had difficulties of judging the value of the result and de- never seen an electric blanket and could hardly be veloping a market were prominent. Dr. Whitney be- expected to imagine what they were like. The first came interested in the idea of the electric blanket in regular production of electric blankets took place in the early nineteen thirties, more or less for the fun of 1940. Production increased every year, with a hiatus it. I don't think he was consciously trying to develop during the war, but it was not until 1947 that a profit a new electrical appliance; he rather was motivated by position was achieved. Meanwhile a great deal of ena curiosity to see how an electrically heated comforter gineering development went into progressive refinewould work. He got Mr. William Kearsley to build a ment of this basically simple device. At present, ten first sample, and it worked well from the beginning. years after the blanket was first marketed, the national A type of control was devised which automatically market is estimated to be some $12 million compensated for changes in room temperature, and a It is easy to measure the modest research expendinumber of experimental blankets were built and used tures which went into this development, but for by willing laboratory collaborators. One distressing reasons similar to those found in the previously mendifficulty was discovered. When the blanket is prop- tioned cases, the determination of a return from the erly spread on the bed, a satisfactory temperature development which is assignable to research is quite requires a heat input of about 150 watts. If the blan- another matter. The contribution of many groups was ket should be rolled up without being turned off, this essential to the success of the project, but special credit 150 watts of heat, well insulated, will cause the center is due to the marketing people, who must be particuof the rolled-up blanket to become very hot and even- larly skillful in the case of any basically new product tually to burn. Before we had a solution to this prob- such as this. lem, a number of these early experimental electric PROJECT CIRRUS blankets caught fire. A peculiarly intricate example of the difficulty of When this happened, Mr. Kearsley was invariably measuring the return from research is presented by confronted the next morning, if not earlier, by the un- Project Cirrus, the work on weather modification, in fortunate individual who, clutching the charred re- which our laboratory is participating in an advisory mains in one hand, described his misfortune with the and analytical capacity. other. Happily, there were no serious accidents. Mr. During World War II, Langmuir and Schaefer unKearsley began to notice that people who burned up dertook a study of aircraft-wing icing, which focused blankets, although naturally somewhat upset by the attention on the problem of supercooled clouds. Ice

31 Measuring Research Returns forms on aircraft structures when they come in con- means for influencing weather. These are the bare outtact with clouds in which the liquid water particles lines of the case history of the project. Now let us try are at temperatures substantially below the freezing to measure the return from this research! I know of point. The meteorologists were familiar with super- no single application of this work to the products cooled clouds, but they didn't know very much about which we manufacture. Yet it was decided in the them, and the phenomena held a great fascination for highest councils of the company that the importance Langmuir and Schaefer. So, at the end of the war, of the work to the economy as well as the national they set out to learn how the water in supercooled security was such that we should continue the work. clouds can be made to freeze. A home freezing cabinet This case illustrates the necessity and obligation of was the principal laboratory equipment required for considering many factors, some of which are nonthe experiments, and in it Schaefer formed supercooled economic, in measuring the results of research of this clouds and studied their properties. He discovered that character. a particle of dry ice would quickly nucleate the entire The problem of measuring the return from research cloud, causing it to settle out as snow in the bottom of varies greatly with different types of work. In some the chamber. Later Dr. Vonnegut determined that cases the problem reduces to that of separating the nucleation of a supercooled cloud could be accom- contribution of research from the contributions of plished by silver iodide particles. Both methods were the many skills-engineering, manufacturing, and marsuccessfully applied to actual clouds: by "seeding" the keting, etc.-which have by their collective efforts made cloud from aircraft, dispensing dry ice or silver iodide, possible a new development. In other, less simple cases or by silver iodide generators operated on the ground. the effect of a new material, for example, formex, has At this point we asked the armed services (and they so many ramifications in so many different areas of agreed) to undertake a contract in partial support of an industrial economy that it seems hopeless to measthe work, primarily because of the legal liability aspect ure it adequately. In the case of still other research, which is inherent in parts of the task. for example, Project Cirrus, the results of research With this pioneering work has come into being a extend so far beyond the boundaries of a given indusnew science of experimental meteorology. Active ex- trial enterprise-in this case they may well extend perimenting on weather modification is known to be beyond the boundaries of the nation-that one cannot going on in Australia, Canada, Mexico, Hawaii, South hope to measure the return from the research. Africa, in many South and Central American coun- I have developed the thesis that measuring the retries, and probably at many unannounced points. In turn from research is difficult because of the manifold the United States, in addition to the well-known re- ways in which the beneficial results of research maniservoir-filling operation carried out by New York City, fest themselves and because of the important and there are many local cloud-seeding projects going on inter-related contributions of engineering, manufacin various parts of the country. It is somewhat astonish- turing, and marketing to a research result which attains ing to behold the rather casual and routine manner a market. I would like to leave no reasonable doubt, in which rain-making contractors carry out their opera- however, on one point. Although the contributions tion in the southwestern part of the country. Probably of many groups are vital to the practical utilization of far more important than local rain making is the a research result, it is the scientist himself who progrowing evidence that continent-wide weather modi- duces the vital foundation on which the whole strucfication is feasible. Where this work will lead eventu- ture rests-the new fact of nature. The scientists and ally is impossible to predict, but its potentialities are the laboratories which have the freedom to follow of the very greatest importance. It is certain that for their curiosity in search of new facts of nature are the first time in history man has at hand a powerful among our most important national assets. DISCUSSION OF DR. ABRAMS', ADMIRAL PARSONS', AND DR. SUITS' PAPERS DR. STEVENS: In the various papers that have been research. I would like to know what would have happresented for the quantitative measurement of the re- pened if you had not done the research. sults of research, I see no base line, no measure of what I do not know whether there is any particular base. would have happened had the research not been con- I know that in our own company, we can segregate ducted, and I would like to ask Allen Abrams how he the developments that have come through research, would cover that. the new products we are now making, and the profit It seems to me there is a little hazard in telling that comes from those products. manufacturers that you can measure the return from DR. ABRAMS: I have to refer to the various men

Measuring Research Returns 32 who responded to our questionnaire. A number of with an index of returns if you had one? How much companies said that if the research department were research would industry support? not there, they would no longer be in business. MR. HUGHES: I was wondering if there has been A director of research who has been at the job for set up in any industry a body or division, an independsome time acquires an innate feeling that he has or ent body, that is analagous to the weapons evaluation has not accomplished certain things. group. If so, what effect has it had on industry? Research has developed something like 30 per cent DR. O'BRYAN: I don't know whether there are any of our present production, on which we show about industry-wide evaluation groups. I am sure that every 45 per cent of our total profit. company has its evaluation, with its quality controls COLONEL SEILER: I was wondering whether the and market analysis. Weapons Systems Evaluation Group has any method A VOICE: We have set one up on a fair-sized scale for apprizing the research and development people in and have men that have had experience in military the defense department of the shortcomings as they ap- work. We have applied it to an area that five years pear to be evident from their evaluation? ago would not have been open to research. ADMIRAL PARSONS: That brings up quite a ques- DR. O'BRYAN: I think this sort of thing has been tion. Of course, in our travels and cruises, in B-36's going on since the beginning of time, but it has never and in submarines, and so on, we run into a lot of been so formal, or was ever called "operational analyequipment which is apparently not operating right. sis." The question is how to handle such reports. If we DR. KILLIAN: I might also mention that the handle it on a formal basis and blacken the name of National Research Council has a committee studying the organization by reporting it after we get back to the application of operational research to industrial Washington, that would probably be the last trip and civilian needs. It is quite an active committee. It on which we would get any information of value to is a new one and it is just making its first study. us. DR. HOLLAND: It seems to me that for a long time We would become confused with the Gestapo, or there has been a need for something along that line, the Inspector General, and we would kill our sources an evaluations group in industry, and to a limited of information. extent there has been something of that kind going on. We find that the best way to handle situations like For instance, at one time before the practice became that, when we find something hot, either in a good illegal, testing laboratories used to test the products or bad sense, we come back and deal on a completely of the various laboratories in the country. But there informal basis, at the points in Washington where it is one fundamental difference in the job of the miliis likely to result in the most good. For instance, tary and the corresponding job in industry. Military we actually went to the Naval Research Laboratory, people in the time of peace cannot get operational and General Barnes of the Air Force was very struck data. We can. with some new things of a classified nature going on. If the products fail, we find it out immediately, beHe proceeded to help the Air Force Major who was cause the customers cast their vote. So, in a sense, our the liaison officer there by becoming a major general weapons evaluation is carried on as a current activity. liaison officer for that particular project and seeing to Perhaps industry could go farther in anticipating it that it got pushed ahead. customer reaction, and I suspect that there are some I think that is the answer to your question. We do opportunities there that are being overlooked right approach these things, but we very seldom write let- now. ters, unless it is a question of, well, almost international MR. FISHER: First, I would like to make a comment importance, such as the logistics of bases or something and then ask a question. like that. We do not have an opportunity to test our product If we become aware of a serious problem affecting without very large capital investment, and I have our ability to perform strategic bombing, for instance, somewhat of a personal horror of industry, where we would perhaps write a letter to the joint chiefs in everyone agrees that the indices have to be modified that particular case. That does not reflect on any group by judgment, but many fail to do it. or organization that has been our host in the United I would like to ask whether the people who are usStates. ing these indices find that they are misused to any DR. O'BRYAN: That has been our experience. They great extent, to any damaging extent? not only have General Barnes, they have General MR. REEVES:'I would like to comment on that Gavin and Admiral Parsons, and also other sources of question. I agree that it is very easy to misuse them, information, and they very quickly get it to the per- and one reason I was a little evasive this morning, sons who are concerned, and very rarely write letters. when somebody asked what we had done with them, DR. HOLLAND: I would like to make an observa- was that we have found that they can have some tion about this index of returns. What would you do fairly surprising results. In other words, if you try

33 Measuring Research Returns to claim too much for research, other people start is made for native American genius to flow into such arguing in the other direction, and there is quite a an evaluation system? battle on as to whether the research is doing any good ADMIRAL PARSONS: In the Weapons Systems Evaluaor not. tion Group, we really get the fruit of research and deIn answer to the question of why you need these velopment after it has reached or at least approached things, it seems to me that it is very important for a the prototype stage. company to know what its different operations contri- Now, in the case of the various departments, for exbute to the overall operations. I know that in our ample, the Air Force, the Army, and the Navy, the industry we have had a very clear illustration of this Research and Development Board may get a certain because we make a great many products out of one invention, and it will eventually end up in a contract raw material. As a result, it was extremely difficult for the development of that particular article. I bein our early operations to try and find out how lieve that we would come in quite a little time after its much it cost to make each product. inception, in that particular case. In other words, we needed to find out how much We might come in as midwives, but we would not gasoline costs, how much heating oil costs, how much decide how the infant is to be reared. I believe that it different lubricants cost, etc. At one time all products is a function of the Research and Development Board were sold for what we could get for them, and if the and the department concerned at least to bear the company made money, all thought they had done a child. pretty good job. We then would look at it and be very interested, Later they set up a system whereby they could or perhaps not so interested in a particular case. Let analyze the actual costs that went into the production me take the case of helicopters. I can think of several of different products, and that revealed that there new systems in which the development of the helicopwere many weak points in the operations. There were ter is the principal bottleneck. The helicopter will some products that were losing quite a bit of money have to become about four times as good, e.g., flyable while others were big money-makers. From this study in the dark, and on instruments, etc., before any parit was possible for the company to readjust its opera- ticular hot idea will bear fruit; so we go around and tions so as to make the products which were the most talk up helicopters for the purpose of putting across successful from a financial standpoint and would in- some ideas which can only be used for helicopters. But crease the overall return. I would like to say that we do not receive the invenI think the same applies to research. I believe that tion at the earliest stage, which might be called the it is fairly certain and accepted by most companies glint-in-the-eye stage. that research is desirable and that it is worth-while. A VOICE: Now offhand, if only to give the dogs a But the question is, how worth-while is it? Are we rag to chew on, so to speak, I would like to submit a doing too much of it or too little, and are the research suggestion for the agenda at the next meeting. Perefforts in the right direction? If we had the answers to haps it might be possible not to have so many rethose questions, I suppose research would be much search directors but rather some of the people that more effective than it is now. handle the money, to discuss how they decide whether DR. ABRAMS: I think that is a very good statement. their research is any good. I believe we have been After investigating a good many of the results of these discussing it all from the point of view of the reresearch directors, my own feeling is that any kind of searcher, figuring out how to prove to the man with index you may set up is perhaps a morale builder, and the money that our product is worth-while. as long as nobody in the company objects to it, the DR. O'BRYAN: I think that Dr. Abrams, with years index may give someone a lift. with the Marathon Company, knows something about MR. HOLLAND: My question is addressed to Ad- that. miral Parsons. I was very much impressed with the DR. ABRAMS: The question you raise bothers the Weapons Systems Evaluation Group. I followed your average research director. He wonders how good he is. talk with a great deal of interest and I feel a lot more I suppose that to be one reason you find so many secure as a citizen, knowing that such work is going on. people interested in getting some basis for a determinaMy question has to do with this thought. Long be- tion of what they are accomplishing. fore we had organized research, we had in this country It is interesting that such a minor part of a coma great talent for developing individual inventors. In pany's expense should be subject to so much discussion. fact, an ancestor of mine happens to have been one of In the average company today it is only one or them; his name is John Holland, who invented the two per cent that is being spent on research, and the submarine. insurance on the building and equipment is more What part of the flow of material that comes in than the average research budget. comes from organized research groups and what part Management requires a feeling that the research comes from independent inventors? What provision department is producing the goods, and the general

Measuring Research Returns 34 health of the company is good because of that. Since cooperation and teamwork all along the line A VOICE: To clarify what I was trying to say be- are required on large and technically complex developfore, part of the difficulty arises out of a full formula- ments, it seems to me that it is desirable tactically for tion of the problem. When you try to find the linear the research laboratory to show restraint in claiming relationship and can't find any, do you leave it there credit for its part of the work. or do you reformulate the problem? DR. FOOTE: I would like to enlarge upon that a It seems to me it is a fairly common occurrence to bit. Some of the reluctance of corporate officers to exseek a solution along a certain line, and the discovery pand research and research facilities is due to the fact that you cannot find it along that line is frequently the that, except for routine improvement of operating precious discovery that permits you to reformulate it problems, the results of research usually necessitate and find it in another direction. capital assets expansion. Research either renders exIs that perhaps true in this case? That is the ques- isting processes obsolete or it introduces a new prodtion. uct or process, both results requiring further capital DR. O'BRYAN: It is not necessarily inverse, is it? investment for plant equipment. The industrialists in A VOICE: No, not necessarily inverse at all. general would like to encourage every development of DR. O'BRYAN: Dr. Suits will be unable to be here the research laboratory but are limited by funds that tomorrow, so he won't be in the round-table confer- are necessary to capitalize upon these developments. ences, and he has agreed to be a special target for the In fact, it may be possible that the quantity of renext few seconds. search that a corporation can finance should be deDR. SUITS: If I may comment on the last few re- termined by the amount of capital asset expansion marks, I agree with Dr. Abrams and the other gentle- that should be diverted to putting the results of remen, that an important problem is that of -determin- search into commercial operation. This is many times ing the over-all matter of how much research the com- greater than the research budget itself and constitutes pany should do, as well as the detailed evaluation of a major problem of finance, especially in our present what has come from the individual projects. As Dr. inflationary spiral, where funds arising from amortiReeves pointed out, there is an unfortunate aspect of zation are so totally inadequate. the attempt to evaluate separately the contribution A VOICE: I don't think you can place too much of research to the success of the project. If you at- stress on the inter-relationship between departments. tempt to ascribe to the research contribution an un- I know a man who is running a research organization realistic share in the success of the over-all project, for a company selling about ten million dollars' worth you may endanger the continuing cooperation of all of products a year in a highly competitive industry, the groups which by their joint efforts have done the and this organization during the past ten years has job. developed about five per cent of all its products that There is a well-known technique in our company the corporation is currently selling. The products are of ensuring that the step following research will be good as a whole, and so on, and yet the corporation is taken, the next step being an engineering step. When losing half a million dollars a year. What kind of an a project leaves the laboratory, there is quite a bit index can you dream up that will give that research left undone; otherwise it would be difficult to interest department the break it needs? There is another part the engineers in the job. At least that is what we claim. of the business in this case, which, I am sure, would The fact is that we usually don't know how to take fall down under any index system that I have ever the next step, so that a great deal is left to the ingen- heard discussed. You cannot get a break for the reuity of the engineers in the development laboratories. search department under such conditions. ROUND-TABLE DISCUSSIONS * -ILEROY A. BROTHERS, Chairman The discussion centered around indices as a means of Concern was expressed that the misuse of indices measuring the return from research. Agreement was might be more dangerous than if no use were made reached on the view that the use of indices generally of indices. This was countered strongly by those who should be restricted to development, to operations re- held that intelligent use of indices was so valuable search, and to applied research, and that they should * Two round-table discussions were necessary because of the not be used for fundamental and basic research. large attendance.

35 Measuring Research Returns as to warrant the minor risk that some harm might be the vitality of a research organization. done through misuse. It was stated that a research Indices of the return from research may be of many worker who could recognize how his efforts might kinds, some capable of being assigned numerical values create a return for his organization might easily be and others which are relatively indefinite. Some of influenced to change his efforts from a research proj- these enumerated were: dollar return, capital requireect of small potential return to an area of greater ments, manpower costs, scheduled time of completion, apparent return. The importance of a set of convinc- casualties in warfare, consumption of strategic maing figures to be presented to a board of directors was terials, patent royalties, morale, and public relations. also emphasized. Emphasis was placed on long-term evaluation, ending For fundamental research it was recognized that with more rephrasing than occurred in the talks beindices were probably of very little value. Other meas- fore the Conference. These included: vice president in urements of the return from research were stated to charge of "change," vice president in charge of "no be the personal satisfaction to the research worker and change," and president in charge of "profit." — II - ALLEN ABRAMS and W. T. BLAKE, Co-Chairmen Evaluating the results of research is a subject on which Research has been likened to insurance, yet, since it there is a wealth of interest but a dearth of knowledge. doesn't guarantee protection from techno-economic Evaluating research and development is done to some hazards, it requires salesmanship. extent by all research directors and their superiors. Evaluating the results of research to the research Occasionally one finds formality in evaluating re- director is perhaps the most important reason why search, including the techniques of the statistician and the interest in this topic is so great. Precise facts on chartist. More often, however, one finds the procedures the results of past research can help the research diinformal, even obscure. rector to judge the potential value of new programs The extent of interest in this subject is based on the of research. Records of past expenditures of time and realization that research evaluation is vital to research money and the effect of these expenditures on the and development growth. There are two recognized profit and loss statement of the company can be of valuable elements in evaluating the results of research, great service in appraising the calculated risks of new whether formal or informal, namely, the element of research programs. good salesmanship to those who hold the purse strings The need for formal procedures of evaluating reand the element of better control of the research search results depends to some extent on the need for program, both present and future. research salesmanship. A research director who is a Some means of evaluation is important in order to good salesman puts the story of research results in the put research and development on a comparable basis language of the businessman, which often means that with other departments which have tangible products. it must be related to the dollar sign. In contrast, some It is easy for a manufacturing department to point companies are old and experienced in research and to the number of units which they produced during have managements that understand and have faith a certain period. The research and development de- in the value of research. Therefore, evaluation propartment also needs to be able to count the quality cedures for purposes of salesmanship are less imporand quantity of its "products" during any time span. tant. These are often companies that can point with The need for this is great at all times for the further- pride to the fact that their whole existence now deing of the research worker's feeling that he is making pends on products which came up the "research way." a definite contribution, particularly great in times of There are but a few organizations with specific, preausterity, when research and development is often cise methods for evaluating the results of research beconsidered a luxury that can be retrenched. Data on cause the problem is fraught with great difficulties. research results is helpful, not only to the research Precision in method is easier for a research and dedirector himself in selling the value of his services velopment department which devotes almost all its to his superiors, but to each successive echelon above energies to product and process improvements than for him, in their salesmanship to those to whom they are a department that devotes a great share of its energy responsible. Thus, even the board of directors in in- to research of a more fundamental nature. It is easier dustry must be armed with facts on research results to evaluate the results of a process which improves when subjected to the interrogations of stockholders. yield than the results of some fundamental research

Measuring Research Returns 36 which is applied in varying degrees over a time span greatest hazard is the resistance from other departin a number of company operations. ments of the company that will result from lack of One research department makes periodic checks modesty and over-salesmanship. Any formal procedure of its profits resulting from products that came up for evaluating research should have the blessing of the "research way" vs. those obtained otherwise. This other departments, and savings ascribed to research research department evaluated its usefulness by point- should be agreed upon by representatives from other ing to the profit increment on "research items" vs. departments concerned. Failure to recognize that re"nonresearch items" and divided this increment by the search is but one member of the team can cause more annual cost of research. In this case, the ratio of profit harm than good. Research and development results increment to research costs on an annual basis was often cause the production man the "trouble" of 10:1. Another company divided the value for one abandoning the status quo and changing to a process year of projects completed during a fiscal year result- that expects more of him in terms of control, watchfuling in reduced product or process costs by the annual ness, and direction. cost of research. The ratio in this case was 14: 10. The There is always a possibility also that formal protechniques used by Olin Industries can be referred cedures for evaluating research will result in building to as perhaps the most elaborate effort at precision a research program around those types of projects in the area of evaluating the results of research. (See where research results can most easily and most quickly Business Week, December 3, 1949, p. 26). be recognized. This could act as a substantial deterSome companies have minor techniques that serve rent to the undertaking of long-range research of a the purpose of evaluating research to some extent, one more fundamental nature-research which in the final of these being the maintenance of a log book as to analysis could be vital to the company's existence. time and place that research and development findings Therefore, a research director must guard against alare adopted by the production departments of the lowing formal procedures for evaluating research to company. Another company lists the annual sales vol- unbalance his research program, However, a realistic ume of products that resulted from research and leaves research director, especially in a company new in rethe precise dollar credit due research an open ques- search, will have in his portfolio of projects a sufficient tion. Some companies "sell" research to some extent number of the short-term tangible-results nature to not by evaluation but rather to "keep up with the provide him with facts for his salesmanship. Jones'," either for public relations or out of fear of Problems of the evaluating of results of Government being left behind in the techno-economic race. Other research are somewhat unique, yet the need for salesresearch directors, not having a system for evaluating ae mewh research, admit to the real need for their keeping their manship at the time when research appropriations are "powder dry" to protect them against all who assail being considered is often great. There are many Govresearch. ernment research projects costing substantial sums of There are many precautions to be taken in tech- money that can never be properly evaluated until time niques that place a dollar value on research results. The of war.

Dinner Meeting James P. Adams, presiding Provost, University of Michigan ENGINEERING RESEARCH IN MODERN CORPORATIONS by JAMES C. ZEDER Director of Engineering and Research Chrysler Corporation I CAN'T BEGIN to tell you how honored I am to be in- believe that this same method of classification will apvited to address this distinguished audience. As a ply to most modern corporations, even though differgraduate of this University, I am proud that it can ent terms may be used. These classifications, according serve as host to a gathering of the leading minds of to the names which I prefer for them, are: production this country on a subject so fundamental to progress testing, product engineering and development, engias research. As an engineer from industry, I am pleased neering research, and academic research. to share in this joint effort by universities, Govern- The terms "production testing" and "product enment, and industry to reach solutions to problems gineering and development" are self-explanatory. which we have in common. As an individual, however, Production testing is the devising and performing of I would feel much more comfortable sitting out there tests on production samples to insure that the product listening to some one of you speak from this platform meets the standards of performance and life desired in my place. for it. Product engineering and development is the Research is a controversial subject. Nearly everyone activity of bringing out the annual new models-inhas strong opinions about it, and yet even among cluding the initial establishment of all specifications, research men opinions differ widely as to what it is and the preparation of drawings which enable the product how it is done. Outside the research laboratory the to be manufactured, and the constant refining and opinions get wilder and wilder. In some quarters, the improvement of the product after it is in production. belief obtains that all good-political and social har- "Engineering research" is a little farther upstream mony, health, and complete freedom from economic from the product than production engineering and dedrudgery-in short, heaven itself, will eventually velopment and is therefore harder to define. What we emerge from the laboratories if we will wait just a little mean at Chrysler Corporation by this term is the longer and spend just a few more dollars. organized effort to apply new materials, new processes, Much of this is generated by our commercial attitude and new principles to the improvement of our prodtoward research. Research is a word which has become ucts by basic revisions in design, the introduction of a symbol for progress, and, in the highly competitive entirely new features, or even the addition of new atmosphere of modern merchandising, industry tends products to the line. to divert attention from the heardheaded pick-and- The fourth category, "academic research," we idenshovel job that a research organization has to do to its tify as a search for knowledge without regard to its more romantic but specious aspects. possible application. It is sometimes referred to as I cannot tell this group anything new about re- "fundamental" or "basic" and has been called research. However, I can explain the thinking and atti- search for research's sake. For reasons which I will tudes on this subject which derive from the experience explain in a few moments, we consider it to be outof one industrial engineering and research organiza- side of the scope of an industrial organization. tion. At Chrysler Corporation, the sequence of these I have reviewed these classifications only as a matter activities from the uncovering of new knowledge in of orientation, for the only one I intend to talk about the engineering sciences to the production of finished is engineering research. Obviously, both production goods may be classified under four headings, and I testing and product engineering and development

Engineering Research 38 are essential to a successful corporation. Unless a prod- some share of the blame because selling research is a uct is being manufactured which has no unknown substantial part of his job. quantities and unless that product can hold public Sometimes, you will find a research engineer who acceptance indefinitely without change, no industrial will ride a project he likes for as long as you will let organization can long continue without both these him. Give him one job that interests him and he is functions. fixed for life. You can never get him to wind it up. In regard to academic research, on the other hand, Then there is another type who will stick to a project our experience at Chrysler Corporation has definitely until he loses interest or the going gets tough, and indicated that it can be carried on more economically then he will switch to another project. This kind in the long view by universities and research founda- never gets anything done, either. A good research ditions. Academic research is the fountainhead of all rector must constantly be on his guard against people future progress; but to carry on such research in an of this kind in his organization. industrial corporation imposes too much of a strain Finally, a good research head must keep his eye on normal viewpoints and attitudes. After all, a cor- on company objectives and not be easily lured away poration is run for the express purpose of turning out from what is best for the corporation to a pursuit of a product at a profit, and setting up an academic re- his own special interests. Sometimes this means havsearch group within corporate confines and then try- ing the courage to stop a project, even though some ing to wall it off from the motivating philosophy of the sacrifice of personal prestige is involved, when it becompany is almost impossible. Too soon you will find comes obvious that the results to be attained will never your academic research group joyously-and profitably justify the expenditures. -engaged in engineering research or product develop- Successful operation of a research group demands a ment. continuous flow of new projects, as well as an output The most important element in a successful engi- of successful and valuable developments. As a matter neering research group is the quality of the man who of fact, the quality of research output depends to a heads it. Research must be an organized effort to be very large extent on the selection of the projects. effective; hence, a research head must be primarily a Research in an industrial organization must have a good organizer and administrator. We have had the definite objective; it cannot move in the line of least experience of taking a research department out of resistance, or necessarily that of greatest scientific inthe hands of a highly skilled researcher with little ad- terest, but must be acknowledged from the beginning ministrative ability, turning it over to a good technical as a step in product improvement. In the selection man who was not a researcher but who knew how to of these projects at Chrysler Corporation we ask perorganize, and seeing the effective research output of tinent questions: that department increase enormously with the same 1) If the work is successful, will the results be used facilities and staff. in the product or for the direct benefit of other corpWhile a research head need not be the top scientist orat Wi ll the probable cost of doing othe researchp in the organization, he must certainly have a thorough 2) Will the probable cost of doing the research knowledge of his field, a good understanding of what justify the results obtained and be within the financial knowledge of his field, a good understanding of what J research is, and an overwhelming desire to do it. Too many research directors tend to fool both themselves 3) Will it be logical to use the results at the time many research directors tend to fool both themselvesv and their managements with dramatic demonstrations they become available or will they be so far in advance and their managements with dramatic demonstrations 8.. of man-made lightning, frying eggs on newspapers, and of manufacturing techniques and market requirements..', gthat public acceptance will be questionable? the like. They confuse exhibitionism with research and There are many other questions such as: Is manas a result accomplish little that is good and lasting. y n On the other hand, a little showmanship is some- power available and capable of handling the job? Are On the other hand, a little showmanship is some- s... Is times necessary to get an idea across. A research head adequate facilities and instrumentation available? Is has one of the most frustrating jobs in the world its scope too broad? Will it interfere with the everyhas one of the most frustrating jobs in the world- dt day business of the corporation? selling new ideas to people who are operating success- bu fully with old ideas-and he must have not only the We do not lay the burden of answering these quescourage of his convictions but also the salesmanship tions solely upon our research group. After a prelimcourage of his convictions but also the salesmanship inary investigation by this group has convinced them Bsneaesmsary to get them nshipIdon'tjustmeanag vene that the project has value, the project and its proBy salesmanship I don't just mean aggressiveness, aposed scope is discussed with the heads of the produceither, but rather the skillful use of all the refined tion engineering and development groups in whose techniques which have been developed by modern field of activity the project may eventually lie. This business to gain acceptance for ideas and products. If not only aids the research group in a determination of management is consistently turning down ideas from whether the project should be undertaken but also its research department, the research director must take establishes a close liaison between research and devel

39 Engineering Research opinent from the very inception of the project, which to reeducate a development group to the point of aids greatly in the utilization of the results if the proj- progress already achieved by research, and it is faster ect proves successful. and better to allow research to complete the job alBefore we had this arrangement at Chrysler Corpora- most to the point of commercialization. This condition, the development laboratories were inclined to be tion, however, is unusual and it should not be alantagonistic toward any projects carried on by the lowed to interfere with the normal transfer of research research group. The complaints were that they should projects into the development and production enhave been assigned the job, or that they could do it gineering groups. better, or that the work should not be done at all. We have learned one thing at Chrysler, however, By acquainting the development laboratories with the and that is to allow the research idea to continue to research project at its very start and making them a the point where its acceptance by development or proparty to the decision to start it, we eliminated this duction engineering groups is assured on at least an friction and made the production engineering and de- equal basis with their own developments, since only velopment groups much more receptive to the end too often we attempt to pour the new wine of reresults of the project. search into the old bottles of conventionalism and lose Despite the care with which these early stages are both the wine and the bottles. considered, it does not always preclude the possibility The direct transmission of projects and ideas from that engineering compromises or unforeseen difficul- research to the production engineering and developties will make the project completely unattractive be- ment group is not the only way in which a research fore it is finished; and, consequently, progress reviews department serves its corporation. Frequently, a new are held with the original group, in order that they design which comes out of research is rejected on the may be acquainted with any change which might grounds that it is too far in advance of its time, or affect their interest or later use of the results. that immediate application of it cannot be made, or We have learned through sad experience that noth- simply that it is "no good." But the idea nevertheless ing is quite as disorganizing to a research group or as has had its influence on the thinking of the organizadisturbing to its relations to the organization as a tion. It may just help pave the way for the next idea whole, as a project which has been approved merely which comes along, which by contrast may not seem because it was a hobby horse for someone to ride or quite so radical. Or it may simply act like a seed which was oversold originally by its sponsoring group. The lies dormant for a time and then germinates into a unwillingness to give sufficient time to the initial analy- new plant on the development side of the fence, with sis of the problem or to spend enough money on its the fact forgotten that it originated in the research early exploration inevitably results in a high percent- department. In these ways, a good research program age of rejected effort and the quiet burial of the work is constantly having its impact on the development of capable men in the chilly dawn of realization that operations of the company, even though there does the work should never have been started in the first not seem to be any direct mechanical linkage between place. the two. At the very least, research is a stimulating The operation of a research department has no influence which a progressive and growing company royal road to success. To stretch the metaphor a little cannot afford to be without. further, we might say that it is an extremely rocky It is an oversimplification, however, simply to assert path, if there is any path at all, since, of course, the that research is a good thing and that industrial corjob of research is that of the pathfinder and the porations should do it. There are a lot of companies pioneer. Just as the pioneer will find a place to settle which are rich enough to have extensive research down and take root, so the researcher must occasion- facilities but not smart enough to direct them propally stop long enough to consolidate his gains into us- erly. I have had the experience of having to exhibit able form. If this work is successful, it becomes the polite admiration when being shown around such labproblem of the corporation management to decide oratories while being privately amazed at how poorly when it shall be taken out of research, or, expressing organized they were and how little value was coming it from the research man's point of view, torn from out of them. the sympathetic arms of its fond parents and thrown The mere fact of having a research department is to the development wolves. no justification for a feeling of smugness. No one will From the viewpoint of the research department, the attest more vigorously to the value of research than new idea is seldom old enough to stand on its own we at Chrysler, but I don't mind admitting that we feet when it is taken from them and passed over to a made several false starts over a period of years before development group. It will probably always be that we developed the type of research organization which way, and in extreme cases it may be necessary to allow was right for us. We found through experience that the research group to continue to do the developmental a research department must be tailored to the particuwork. This may arise because it would take too long lar job it has to do. There is no general prescription

Engineering Research 40 that can be followed. What is right for one corpora- When it does not, it is probably not because the retion with a certain product selling in a given market search department is too big or because it is spending may be all wrong for another combination of condi- too much, but rather because it has not been properly tions; so each company must devise its own formula. organized and directed to meet the specific needs of The most productive results are obtained when the the company. In such cases, the trouble is not with the research department is integrated into the engineering job which they are doing, but with the management organization instead of separated from it. There must policies under which they are operating. With intelbe a balance, a harmony; one must not be tied to the ligent management, good direction, and sound organiother. Isolating a research facility from the rest of the zation, industrial research is sure to pay off. corporation-either geographically or organizationally I have left the sixty-four dollar question to the last. -has definite disadvantages. The only outlet which Why must industrial organizations carry on long exresearch has for its products is the production engi- pensive research programs and spend millions of dolneering and development group, and you can't sell to lars for permanent research facilities? There are una customer without maintaining a good contact with doubtedly a good many reasons that I have not thought him. Also, research needs a reasonably close association of, but, fundamentally, it is true that at no time in with the rest of the engineering department in order history was there any indication that man was satisfied to keep informed on the problems which exist and the with his lot. He has experimented with almost every directions which future research should take. phase of his experience and has continuously sought How much should be spent on research? If enthusi- for improved transportation, housing, food producasm is allowed to run away with good business judg- tion, distribution, social organization, defense, and ment, it is possible to spend too much. Research it- health. This desire to achieve a closer approach to self has become more and more costly. The instru- what man considers perfection in all fields of activity ments and equipment required are much more exten- is so fundamental that we can almost conclude that sive and complex than they were in the past. Also, in it is a law of human behavior. the products of today we have run out of the simple, Viewed from this standpoint, research is good busiobvious improvements which could be made, and now ness, and by an organized effort in this direction we it takes much more research manpower and facilities are making use of a fundamental law of human nato get even a little new information. So research is ture in satisfying the desires of individuals for constant an expensive commodity, and in some ways it sug- progress. Competition certainly spurs industrial orgests a parallel to the prospective buyer who asked ganizations into this activity. As an example, no one the yacht salesman how much a week it would cost could compete in today's automobile market with a to operate one of his larger models. The answer was, 1930 automobile, but competition itself is merely an "Mister, if you have to ask that, then you shouldn't exemplification of the way research works to gratify be buying a boat like this." the human desire for better things. Staying in business in a competitive economy means In some companies the idea prevails that a fixed that research must be a business venture, and that as percentage of each sales dollar should be set aside for such it has to be done with good business judgment. research. At Chrysler Corporation we are definitely op- The good business judgment demanded by research posed to any such rule. A good research program de- is like the business judgment that would be exercised pends on an intelligent understanding of the job to with any other capital investment. Patience is required, be done, not on a knowledge of how much is available because research cannot be turned on and off like a to be spent. When you have more money than proj- machine. Once having begun it, there is no turning ects, the research director has to look around for ad- back because you have committed yourself to progress ditional ways to spend it, and you have robbed him and your customers will demand it. Tomorrow's cusof the stimulation of having to compete with other tomers will have grown to expect even more from all divisions of the corporation for his budget allocations. of us in industry, and if we wish their approval the On the other hand, during periods of low sales vol- output of our research efforts will have to be even ume, the fixed percentage system may result in drastic better than before. reductions in the research program at the very time Research has sometimes been characterized as a when research should be expanding instead of con- gamble. In the sense that it is an investment in an tracting. It is far better for the corporation if the re- enterprise with an unpredictable outcome, it is a gamsearch department is required to sell the management ble; but the fact that its success is dependent on the on every dollar of its appropriation on the basis of quality of its direction and execution makes it a game the probable benefits to be derived from the projects of skill rather than of chance. The great industrial undertaken. might of this country has been built on gambles of But while research cannot be bought by the pound this kind, and it is becoming increasingly apparent or percentage, industrial corporations still have the that the biggest and riskiest gamble which industrial right to expect it to pay its own way and much more. corporations can take is to do no research at all.

Third Session Merritt A. Williamson, presiding Associate Director, Research and Development Department Pullman Standard Car Manufacturing Company WHAT IS NEEDED IN A RESEARCH EXECUTIVE by JOHN C. FLANAGAN Director of Research, American Institute for Research DURING WORLD WAR II a number of personnel psy- success and failure in that activity. An efficient techchologists were fortunate in being called on to assist nique for determining critical requirements has been in the solution of armed-forces problems on the clas- developed. This technique, called the "critical incident sification, training, and efficient utilization of man- technique," tends to substitute data for impressions power. The urgency and scope of the problems and and opinions. It provides a relatively precise and comthe large number of personnel involved provided an prehensive definition of effectiveness on a job in terms excellent opportunity to do decisive research on the of what people actually do on this job. The critical effectiveness of many of the personnel procedures incidents are reports by qualified observers of things which had been developed in the period prior to the done that were especially effective or ineffective in acwar. These tests of procedures confirmed some prac- complishing important parts of their job. The vague tices, such as the effectiveness of certain types of hunches, the stereotypes, and the poorly defined traits, aptitude tests, and raised doubts about others. such as character, imagination, and foresight, are reProbably the most significant outcome of these stud- placed by reports of observed incidents which are deies was in the increased realization on the part of tailed and specific. the research workers that the development of accurate To illustrate the technique and report some data job definitions must precede all other types of studies for a related field, a study of officers in the armed on personnel problems. If it is not known in precise forces will be summarized briefly. Several hundred behavior terms what an individual is supposed to officers were asked to report incidents they had obbe doing, including a description of the specific ways served in which something especially effective or inefin which he may succeed or fail, it is impossible to fective was done. Two or three illustrations will indievaluate any part of the personnel program concerned cate the type of data obtained. with this individual. The first example is an incident relative to making The underlining of this need led to the development decisions. This is an ineffective incident about a deputy of new and more precise techniques for defining job for operations: requirements. One of the new concepts growing out About two or three times a week he would come into my of these studies was that of "critical requirements." office and start the conversation by saying, "Say, Colonel, This was a substitute for the long lists of traits, du- you have to make a dicision." He would tell me the probties, and requirements for various jobs which in many lem on which he should have rendered a decision since he aties, were more comprehensive thantwhich in many had all the facts and I didn't. He was afraid to make a decases were more comprehensivea tchanqhey were enlight- cision that I would not approve. The final straw: A decision ening. It was proposed that techniques be developed was needed about sending a ship out in bad weather-he for obtaining short lists of the really critical require- knew the facts but didn't have the courage to decide. He ments. The critical requirements for an activity are came to me, and I told him to make the decision and those that are crucial in the sense that they have been notify me about it. He wrote a letter to all subordinates frequently observed to make the difference between telling them that he had been on the carpet, and the

Qualifications of Research Executive 42 reason was because his subordinates had failed to do as Per Cent Per Cent they were told. This was untrue, and several of his sub- for All for Col. Officers and Gen. ordinates complained to me about the letter, so I requested (N =2,907) (N =412) III. Proficiency in Planning and Directing The second incident regards the matter of taking Action................................. 16.8 39.8 prompt action. This is an ineffective incident about a 25. Taking responsibility................ 2.3 3.4 squadron adjutant: 26. Solving problems.................... 2.6 4.6 This officer was an adjutant in my squadron. My former 27. Making use of experience............ 1.9 8.3 adjutant was very energetic and handled most of the ad- 28. Long-range planning................ 1.5 5.1 ministrative functions with little or no supervision. This 29. Taking prompt action............... 1.0 2.9 new officer started out and was given the same authority 30. Suspending judgment............... 0.7 1.7 as my former one. The administration began to slow down, 31. Making correct decisions.............. 2.8 7.3 reports were going in late, so I began to check. I found 32. Making forceful efforts.............. 2.8 2.4 that this officer was very thorough, so thorough that even 33. Absorbing materials.................. 1.2 3.2 the most ordinary function was slowed up waiting for IV. Acceptance of Organizational his signature. One afternoon I overheard E.M. talking. Responsibility.......................... 13.0 9.2 The conversation was to the effect that they were through 34. Complying with orders and directives.. 1.8 0.0 work at noon and here it was 3 P.M. and the passes were 35. Accepting organizational procedure..... 2.7 0.7 not signed yet. I checked with the adjutant and found that 36. Subordinating personal interests...... 4.2 3.4 he had been too busy at noon to sign the passes but would 37. Cooperating with associates........... 1.6 1.7 get to it as soon as he finished a roster he was making 38. Showing loyalty..................... 1.9 1.2 up. I picked up the passes, signed them and got the men 39. Taking responsibility................ 0.9 2.2 started on their way. I checked with some of the men in V. Acceptance of Personal Responsibility.... 35.3 11.7 the unit and found that this delay had become a common occurrence. I then decided that this man was too slow for 4. Attending to duty................. 0. 41. Attending to details............... 4.4 0.2 my unit and got rid of him on the next shipment. 42. Reporting for appointments.......... 0.4 0.0 TABLE I 43. Meeting commitments................ 0.5 0.0 Relative frequency in per cent of incidents reported for each 44. Being fair and scrupulous............ 5.2 3.9 item on the Check List of Critical Requirements for Officer 45. Maintaining military appearance...... 1.0 0.5 Evaluation, comparing incidents reported concerning Colonels 46. Adapting to associates................ 5.0 2.4 and Generals with those reported for all officers. 47. Adapting to the job............... 5.7 1.7 Per Cent Per Cent 48. Conforming to civil standards........ 6.3 2.4 for All for Col. VI. Proficiency in Duty Military Officers and Gen. Occupational Specialty.................. 14.4 6.1 Item (N = 2,907) (N = 412) tItem (N v 2,907) (N = 412) 49. Possessing fundamental training...... 3.5 1.7 I. Proficiency in Handling Administrative 50. Improving effectiveness.............. 1.5 0.7 Details................................. 6.9 3.6 UnderDetailstanding instructions........... 6.9 3. 6 51. Keeping well informed in specialty.... 1.1 0.5. Understanding instructions..................... 0. 0.0 52. Applying training and information.... 4.5 1.0 2. Scheduling work............................ 0.3 0.2 53. Showing ingenuity in specialty........ 3.0 1.7 3. Getting information from records.... 0.1 0.0 54. Handling related assignments......... 0.8 0.2 4. Getting ideas from others............ 0.1 0.0 5. Checking accuracy of work........... 0.9 0.5 TOTAL FOR ALL AREAS.............. 100.0 100.0 6. Writing letters and reports............ 0.8 0.5 7. Getting cooperation.................. 2.0 1.5 The last incident concerns effective behavior re8. Presenting finished work............. 0.5 0.2 garding forcefulness: 9. Keeping records..................................... 1.1 0.2 A directive was received from higher headquarters re10. Keeping others informed.......................... 0.5 0.2 questing certain detailed information that appeared on 11. Rendering effectiveness reports....................... 0.1 0.2 the surface to be impossible to obtain in the limited time 11. Proficiency in Supervising Personnel.... 13.6 29.6 allowed. In the discussion on the matter, various ideas 12. Matching personnel and jobs......... 1.0 0.7 came up such as reporting certain phases of the required 13. Delegating authority................. 2.0 4.1 information as unknown, requesting higher headquarters 14. Giving orders and instructions........ 1.9 2.9 to extend the reporting date, etc. This officer spoke up and 15. Insuring comprehension.............. 0.2 0.2 said, "This is a directive. We should be getting the infor16. Giving reasons and explanations...... 0.6 1.2 mation instead of discussing alternatives; give me the 17. Supporting authorized action......... 0.6 2.2 job, and I'll get it." He was given the job as requested and 18. Encouraging ideas.................... 0.3 0.7 immediately formed a plan of attack and expedited it. 19. Developing team work............... 1.3 4.4 Information had to be obtained from over 100 locations 20. Setting a good example.............. 1.8 4.1 all over the United States. He formed a unit to bring the 21. Assisting subordinates in their work... 0.6 1.7 information in and another to compile it in the required 22. Evaluating subordinate's work........ 0.6 2.2 phases. The deadline was met on the report due to this 23. Looking out for subordinate's welfare.. 2.0 4.9 officer's initiative and force in handling a seemingly im24. Maintaining relations with subordinates 0.6 0.2 possible problem. This officer at the time was not assigned

43 Qualifications of Research Executive to specific duties, having just reported for duty. I forcibly TABLE II. requested this officer and he is working for me now. Relative frequency in per cent of incidents reported for each On the basis of nearly three thousand such descrip- item on the "Observational Record Form for Research Persontions of outstanding and unsatisfactory job perform- nel," comparing the incidents reported concerning research ance, a Check List of Critical Requirements for Of- workers classified in the two top professional grades with those'ficer Evaluation was prepared. For each of the fifty- reported for all research workers in two of the middle grades. ricer Evaluation was prepared. For each of the fiftyfour critical requirements, the proportion of the inci- % forr P % for 2 highest dents obtained from the total group of reporting of- middle levels levels of of res. res. ficers is shown in the first column in Table I. The third ITEM workers workers column shows the corresponding proportions of inci- (N = 641) (NP= 115) I. Formulating Problems and Hypotheses.. 5.0 6.9 dents concerning Colonels and Generals only. The A. Identifying and exploring prolems.... 1.4 2.6 table shows that about 30 per cent of the incidents B. Defining the problem..2.0 2.6 for the senior officers relate to proficiency in supervis- c. Setting up hypotheses................. 1.6 1.7 ing personnel, and an even larger proportion, about II. Planning, Designing the Investigation.. 15.6 11.3 40 per cent, has to do with proficiency in planning, A. Collecting background information 5.0 5.2 initiating, and directing action. Practically no incidents B. Setting up assumptions....9 describing either outstandingly effective or ineffective c. Identifying & controlling important behavior for these officers are found in the area relat- variables............................ 2.0 2.6 ing to proficiency in handling administrative details. D. Developing systematic & inclusive plans 3.1 1.7 A few incidents were reported regarding the senior E. Developing plans for the use of equipofficers' fundamental training and ingenuity in their ment, materials, or techniques........ 3.1.9 specialties. The remaining incidents dealt with more F. Anticipating difficulties............... 1.6.0 G. Determining the number of observations.2.0 personal aspects of their behavior, such as subordinating personal interests, cooperating with associates, III. Conducting the Investigation....... 18.2 20.8 taking responsibility for subordinates, being fair and A. Developing methods, materials, or equip. 7.5 11.3 scrupulous, adapting to associates, adapting to the job, B. Applying methods and techniques..... 3.0 2.6 sand conforming t civil standards.apting C. Modifying planned procedures...... 1.4.9 and.conformingto civilsn. D. Applying theory...................... 3.1 3.5 In the general area of supervising personnel, the E. Attending to and checking details...... 3.1 1.7 points which are most frequently observed for senior F. Analyzing the data...................3.9 officers as compared with junior officers are: developing IV. Interpreting Research Results......... 3 3.5 team work and looking out for subordinates' welfare. 3.5 A. Evaluating findings...................5 In the area of planning, initiating and directing ac- B. Pointing out implications of data.. 1.6. tion, solving problems, making effective use of experi- V Preparing Reports...................-. 92 5.2 ence, long-range planning, and making correct deciA. Describing and illustrating work...... 2.5.0 sions are the types of critical incidents which are gen- B. Substantiating and findings. 3.4 1.7 erally responsible for observed effective or ineffective C. Organizing the report. 3.0 2.6 behavior on the part of senior officers. D. Using appropriate style in Turning more specifically to the problem of the re- presenting report......................3.9 search executive, limited data are available from a vI. Administering Research Projects....... 27.0 33.9 study of research personnel carried out by the Ameri- A. Selecting and training personnel....... 5.9 6.1 can Institute for Research under the sponsorship of B. Dealing with subordinates............ 9.5 10.4 the Office of Naval Research. In this study, critical c. Planning and coordinating the work incidents were obtained from five hundred scientists of groups............................ 6.2 8.7 in twenty research laboratories. About twenty-five D. Making administrative decisions........ 2.7 4.3 hundred incidents were reported by supervisors de- E. Working with other groups............ 2.7 4.3 scribing something that a research worker did which VII. Accepting Organizational Responsibility 9.2 9.6 was especially effective or ineffective in getting his job A. Performing own work............... 3.0 3.5 done. In the detailed list of observed critical behaviors B. Assisting in the work of others........ 3.0 4.3 regarding research personnel, the two types of inci- C. Subordinating personal interests....... 1.7.9 dents which for research workers classified in the top D. Accepting regulations and supervision 1.6.9 two professional grades of Federal Civil Service em- VIII. Accepting Personal Responsibility..... 10.3 8.7 ployees are observed as critical more frequently than A. Adapting to associates............. 1.2 1.7 B. Adapting to job demands............. 3.9 2.6 for those in lower grades have to do with Area I, For- B. Adapting to job demands........... 3.9 2.6 - -C. Meeting personal commitments.........5.9 mulating Problems and Hypotheses, and Area VI, Ad- D. Being fair and ethical.............. 3.4.9 ministering Research Projects. The data concerning E. Showing interest in work.............. 1.2 2.6 these incidents is reported in Table II. WTithin the first area, the critical behaviors observed TOTAL FOR ALL AREAS.............. 100.0 100.0

Qualifications of Research Executive 44 regarding the research executives included identify- of this event to the subordinate concerned, he immediing and exploring problems, defining the problem, and ately contacted the man and invited him to his office for setting up hypotheses, but the behavior most distinc- interview. Upon arrival, the man was warmly congratutive of the research executive as compared with his lated, given assistance and suggestions in connection with subordinates was identifying and exploring problems. the final routines necessary to certification, and then comIn the area concerned with administering research plimented upon the work he was doing. Discussion of the latter followed and resulted in beneficial suggestions for projects, all the subcategories appear to be Important pursuance of the problem. In so doing, the division head factors in evaluating the effectiveness of the research gratified his subordinate and made him feel that his place executive. These included: selecting and training per- in the organization was important, and inspired him to sonnel, dealing with subordinates, planning and co- a greater degree of effort and initiative. Thus, the over-all ordinating the work of groups, making administrative progress and effectiveness of both the individual and the decisions, and working with other groups. These types institution were measurably enhanced at the expenditure of behavior are also critical for research personnel on of only a few minutes of the executive's time. the next lower levels. The administrative subcategory Because of the differences in the categories used in in which responsibility appears to be most frequently tabulating the data regarding military executives and shared with the next lower level of research super- research executives, a precise comparison is not posvision is that of selecting and training personnel. sible. Certain comparisons which can be made do A few examples illustrating the behavior of re- throw some light on these positions. In both instances, search executives are given below. The first incident it appears that effectiveness of performance with redescribes ineffective behavior on the part of a research spect to the type of administrative detail involved in administrator: staff work, such as checking, record-keeping, getting inAn administrator was informed that one of his sub- formation, and keeping others informed are not reordinates, a section supervisor, had recommended for pro- ported as critical behaviors for either of these types motion a laboratory worker who was notoriously lazy and of executives. Proficiency in administration in the incompetent. Without obtaining the complete story or sense of office detail does not appear to be a critical checking the facts, the administrator summoned his sub- requirement for these types of executives. ordinate and began administering a severe reprimand. During the course of this reprimand, it developed the In both groups, the general area designated "Supercomplaint was based on misinformation and that the sub- vising Personnel" or "Dealing with Subordinates" is ordinate supervisor was completely innocent. Thereupon, a substantial critical factor. Judging from the proporthe administrator merely stated he would look into the tion of critical incidents reported, it is not quite as matter further. Several weeks have passed, and the inci- important for the research executive as for the military dent is apparently considered closed, and the subordinate executive. Proficiency in planning, initiating, and diwas left dangling in a justifiably offended frame of mind. recting action include the largest proportion of the critThe second incident illustrates ineffective behavior ical incidents reported regarding both types of execuon the part of a supervisor in one of the lower grades tives. The major difference in the requirements of of research supervision. Data of this type should be these two groups in this general area is indicated by valuable in deciding whom to promote to higher su- a larger proportion of incidents for the research execupervisory positions: tive dealing with getting new ideas or using imaginaThis supervisor refused to accept the recommendation tion in formulating a problem, plan, or program. made by a shop foreman to simplify and improve the de- Fewer incidents are reported regarding fairness and sign of a new supporting bracket. The supervisor had as- ethics for the research executive. The final comparison signed the project to a subordinate who was inexperienced is in regard to the area designated "Accepting Organiin mechanical design and did little research prior to issuing zational Responsibility." The proportion of critical the plan. The supervisor did not spend any time assisting incidents concerning matters of fitting into the work his subordinate or laying down any general require- plans of others is slightly larger for the research execuments but gave him the job to do completely. The fore- tive. It should be emphasized that the relative fretive. It should be emphasized that the relative freman's idea was approved by the supervisor's superior but discarded by the supervisor because he felt the "production quency of critical incidents of either an effective or department had no business telling him what to do." This ineffective nature is not an indication of superiority inflexible attitude kept production forces away from him, or inferiority of one group over the other in the comwasted time and effort, and produced a poorer product. parisons discussed above. It is merely an indication The last example illustrates effective behavior on of the extent to which the problems and activities in the part of an administrator having to do with super- a particular area lead to critical behaviors as observed vising personnel: by the reporting research workers. A division head learned indirectly of the success of one The foregoing analyses should be regarded in the of his subordinates in obtaining qualifications under a nature of a pilot study with respect to requirements Civil Service Examination after a long period of delay, for executive personnel. Although several thousand correspondence, and appeal. Recognizing the importance incidents were collected in each instance, the number

45 Qualifications of Research Executive relating to individuals in high-level executive posi- failure or whether the results must be completely unusable tions was only something more than a hundred in each to be classified as unsatisfactory. case. For an adequate study, it would be desirable to 4. The observer must be qualified to make judgments recollect a few thousand incidents. garding the actvity observed. Typically, the supervisor on The basis of this new approach to job and personnel the job is in a much better position to make judgment as study is the desire to substitute relatively objective to whether behavior is outstanding or unsatisfactory than data for opinions and hunches. In dealing with the is the job analyst or psychologist. On the other hand, the behavior of research executives or any other groups supervisor on the job is ordinarily lacking in the training nbehavior of research executives or any other groups essential to make an inference as to the particular mental of human beings, it has been found very difficult to trait which caused the behavior to be successful or unsucmeasure with anything like the precision of the physi- cessful. cal sciences. Usually we must fall back on counting. 5. The last necessary condition is that reporting be acAs a rough approximation, this works out quite well curate. The principal problems here are those of memory if we take care in collecting and classifying the items and communication. It is also important that the observer's which we later count. attention be directed to the essential aspects of the beIn establishing the critical requirements in terms havior being observed. of actual job behavior for the research executive, five Even with a very careful formulation of our prospecific conditions must be satisfied if valid results are cedures and the use of large samples, the problems into be obtained. These are as follows: volved in studying human behavior are great, and our 1. It is essential that actual observations be made of the results must usually be reported in terms of probabilion-the-job activity and the product of such activity, ties which are very far from certainties. Recent studies 2. The aims and objectives of the activity must be known have demonstrated that improved techniques such as to the observer. Unless this condition is fulfilled, it will the "critical incident technique" can provide data of be impossible for the observer or judge to identify success verifiable predictive value. The day of expert opinion or failure. For example, a research director might be rated as very successful if the objective of his activity were taken regarding problems of selection and classification, as getting along well with the workmen under him. At the training needs, job requirements, the evaluation of same time, he might be rated as very unsatisfactory if the performance, and job design is coming to a close. Reobjective were the quality and quantity of research ac- search of the type which enables the personnel psycomplished. chologist to stand on the shoulders of the preceding 3. The basis for the specific judgments to be made by the contributors in this field is taking over. Much work observer must be clearly defined. The data can be objective needs to be done, but we can predict with confidence only if all observers are following the same rules. All observers must have the same criteria for making judgments. that in the next few years it wil become possible to For example, the definition must clearly state whether or read a paper stating clearly, and with known precinot a minor imperfection will be regarded as evidence of sion, the needs of the research executive. DISCUSSION OF DR. FLANAGAN'S PAPER DR. FURNAS: I would like to compliment Dr. Flana- revisions and some additions in the Air Force at the gan on the pioneer work which they are doing. At present time. last I see some element of objectivity entering into the This study also forms the basis for a new form being personnel problems. put into use in the Marine Corps. The Marine Corps, I would like to ask: What is the next step? What after a general study of a lot of procedures, decided are you going to do with this information after you that this approach was best. They made some changes. have it? Can your findings be directly translated into For example, they cut the number of items down to some effective system for merit rating? twenty-seven from more than fifty in the original form. DR. FLANAGAN: Yes, we think so. We regard this We are now working on a project to survey experias a first step. As a definition of what is required for ence in the Air Force and are trying to improve the success on the job, in terms of behavior, it seems to be present form, so that it will be even more effective. a first step, and we believe that work on selection, We are also working on the general problem of work on training, and work on evaluation of success merit rating of hourly-wage workers with one of the on the job comes after that. large automobile manufacturers. We have, for example, prepared the officer-effective- We have done some preliminary research on developness report that some of you have. This was prepared ing a merit-rating procedure in connection with studies for the U.S. Air Force and is being used with slight of research personnel for the Office of Naval Research.

Qualifications of Research Executive 46 This study was conceived as a five-year program, and DR. FLANAGAN: We would not use them in our we are now in the third year. We have completed the organization. first step, determining the critical job requirements, MR. Dow: Dr. Flanagan just passed over very and the second step, developing a selection test. We briefly the question of aptitudes. The performance of have also done quite a bit of work on the problem any individual in a given situation is probably a funcof developing a merit-rating procedure for research tion of both his inheritance and his developmental personnel. We are now working on two things: the de- personality. velopment of a proficiency standard to measure skills The field of developmental personality is going to and the readiness of the person to accomplish inde- be a very complex field to work out in this approach. pendent research. This study does not aim to evaluate It is an exceedingly interesting one, but the field a person's aptitude to learn to do research but rather of inherited aptitudes may be a little simpler and his present ability to do it. Besides, we are trying to more immediately useful. It would not be the final andevelop more objective procedures for evaluating prod- swer, because a man may have the ability and still ucts, specifically research reports. There are a num- not perform up to his ability because of personality ber of unsolved problems. We have established that restrictions. You mentioned that during the war you general impressions are practically worthless in most have done some work on aptitudes, which in my mind situations for merit rating and that some method has are largely inherited, and I wonder if it would be to be developed for getting records of observations of possible to go a little further into that. actual behaviors. DR. FLANAGAN: I might just say that that work How long people can remember observations and has been reported. We wrote nineteen volumes, which report them with accuracy, we don't know. In working the Government Printing Office put out at the end with foremen in industry, we found the foremen lost of the war. The volumes run from 200 to 1000 pages three-fourths of the incidents at the end of two weeks, each. as compared with what they would have had if they But to summarize very briefly, we can say that we had reported them each morning. They put a check did achieve success: firmly established, verifiable sucmark in the correct space on the man's record each cess in predicting success in pilot, navigator, and bommorning to indicate things they had seen the previous bardier training, and also more limited success in preday. With some twenty-five workers they usually had dicting combat results. three or four incidents to check each morning. Just It is very hard to find out whether a person is sucwhat the right period of time is for making observa- cessful in combat, and we did not get started on the tions for research personnel, we don't know. problem until late. However, we did considerable DR. FURNAS: Do you consider that the merit-rat- follow-up work to check on the tests. In pilot training, ing forms of the usual kind, which are more or less for example, we took in a special experimental sample helter-skelter, are useless? Also, we apparently will of one thousand people that came to Aviation Cadet have to wait a number of years before the world's Boards and asked to be trained as pilots. This study best rating system is available. Is that correct? was hard to sell since it cost several million dollars. DR. FLANAGAN: I would not like to put a date on We took these applicants no matter what they did. it, and I would hesitate to say anything was the world's on their aptitude test, provided they were physically best. and otherwise qualified and had not been in prison. DR. FURNAS: Putting it another way, considering There was self-selection in that they came in to apply, the time for the culmination of your work, it will be but the boys that came in were given the test and, no at least a couple of years before you will have a recom- matter what they did on the tests, they were sent on mended rating system. to pilot training. DR. FLANAGAN: We hope to be able to submit a It was not a bad group; nor was it a good group; report to the Naval Research personnel in a couple just a random selection from boards all over the of years and say: "This is the system we think you country: one thousand people who said they wanted ought to put in." to be pilots. After taking the first screening examinaWhen we tell them to put it in we will probably tion, they were given a battery of twenty tests; they were assigned numbers, as were all applicants, from also tell them that it has shortcomings and that fur- one (lowest aptitude) to nine (highest aptitude). ther research is desirable. In the end we hope to have Because of the failure to use the first screening, something which is practical and also follows sound there were as many as one hundred and fifty with the principles, but we haven't got there yet in the research- lowest aptitude (score of one). personnel field. To report very briefly: out of the one hundred and DR. FURNAS: IS my first impression correct that you fifty Number l's there was not a single one left by consider the present or available rating systems to be the time they got ready to graduate from advance essentially useless? Perhaps that question is unfair. training. Of the 2's and 3's there were sixteen or seven

47 Qualifications of Research Executive teen who graduated. Of the 8's and 9's there were very the new form was preferred as giving a fairer, more few, only sixteen or seventeen, who failed. valid estimate, in spite of the fact that it entailed a To improve our procedures we made a special fol- little more work. low-up on those that failed or succeeded contrary to This does not prove that it is better. There are our expectations. In general, the correlation (the pre- many difficulties in obtaining valid effectiveness redictive accuracy) was about.70. This was fairly close ports, and we have just made a survey of what difficulto the maximum we could get with the criteria of suc- ties are being encountered with the Air Force officer cess in pilot training that we had. report used at the present time. Further work is being done now. For example, modi- I do not want to go into a lengthy discussion, but fications are being made in procedures and equipment it is being misused in a lot of instances. For example, to select jet pilots. it has been used to determine whether the person Very little information was available at the begin- gets a promotion to the next grade. Rumors have gotning of the war on what was necessary for success or ten around that you have to have a 4.0 rating to be failure as a pilot. Vocabulary tests were no criterion promoted and that certain commands are rating higher of probable success. Ability to do simple addition than others. This affects the ratings. For example, one showed a little consistent, negative correlation with ratersaid: "I went through and marked this fellow ability to assimilate pilot training. But we did find just the way he ought to be marked, then I looked at that certain coordination tests, reaction-time tests, it and I recognized the fact that I am traditionally mechanical, spatial, and visual tests predict quite ac- a low rater, so I went through and changed all my curately. marks, so as to raise him one point." Now, if you have that sort of thing being done, adCOL. SEILER: I am somewhat interested in the new rating forms for officers, because, as I recall, the criti- ministrative actions based on just this one report, you cism of the old form was that we had too many people are not going to get satisfactory results. We are trying ending up with the same rating; in other words, we to work out a procedure which will have the officer did not have enough spread of the rating of the people make some of his judgments about administrative to get a good distribution of their effectiveness, actions independently of whatever the description of actual performance on the part of this officer is. In reviewing the new rating forms on the subordin- a If anybody has any other ideas for improvement, I ate officers, I find that this form, for a small segment would like to have them. of officers, also tends to end in ratings that are very In one command, the Air Materiel Command, they close to each other, but I was wondering whether you required their officers for a period of time to make had an opportunity to examine a large number of required their officers for a period of time to make observations frequently and to keep records. The prothese, to see whether this effectiveness report has at-e put cedure had one thing lacking, which we have put tained in some way a better effectiveness rating of the in ourother procedures. The officers were supposed military personnel. to observe and report all actions, whether they were DR. FLANAGAN: "Better" either implies a criterion, mediocre, poor, very good, and so forth. something to compare it with, or some sort of logical Recently, while working on some procedures, we analysis or judgement factor, and I do not believe said, "Let's not record everything let's report only that we have any such criterion. the critical behaviors that are especially effective or I would say that the officers in the field try-out, about especially ineffective." two thousand officers, liked it better than other forms This saves a great deal of observational time. Just they had used. I think that is still true in general. exactly how this can be made into a system that the In the recent Marine Corps try-out, in which four busy superior will be able to use, we don't know as thousand officers used a number of different forms, yet, but we are trying out some new approaches.

Qualifications of Research Executive 48 - II - by ALBERT A. LOMBARD, JR. Scientific Assistant, Directorate of Research and Development Headquarters, United States Air Force Introduction tion to insure the timely improvements in concepts, I am pleased to participate in this program on the techniques, and materiel of promise to the Air Force discussion of "What is Needed in a Research Execu- mission, taking fullest advantage of research being tive." In view of my connection with the Air Force conducted outside the Air Force. An important remy talk will be slanted toward Air Force research, sponsibility of the research executive is to see that recognizing that many of the same factors also apply this policy is carried out in practice. to research in other Government departments and to Air Force Research and Development Program some extent to non-Government research as well. s Administration of Air Force research is a technical Why Air Force Research? job. The research executive must be technically trained First of all, it is essential that the Air Force research in order to determine what projects will be undertaken executive understand the role of the Government, and and to judge the competence of individuals or organiof the Air Force particularly, in the research field. zations conducting the work. In order to explain these There are some people who would have the Govern- technical requirements in some detail, I would like ment take over everything. However, our nation has to discuss the various phases of the Air Force research been built up on the basis of private initiative and and developmental program in terms of its broad obresourcefulness with incentives and rewards as results jectives. of the effectiveness of private enterprise. We want this DETERMINE MOST ECONOMICAL PROGRAM to continue in the research field, as in others, to the Background research is being conducted on a broad greatest possible degree. Many types of research have scale today to determine the most economical probeen and are being supported completely by private gram for the development of weapons and weapons capital. systems of maximum effectiveness with a minimum Industries are supporting research to an increasing cost. Important studies in this field are being perdegree today, as is evidenced by the interest in attend- formed by the RAND Corporation, a nonprofit corance at this conference. As several speakers have indi- poration operating under a contract, to investigate cated, industrial research is primarily pointed toward and improve methods and techniques in the conduct the development of new products or production of intercontinental warfare. These studies consider the methods of value to the individual company, although performance, reliability, vulnerability, and logistic many companies do sponsor some basic research as factors pertaining to the weapons and to the weapons well. In addition, considerable basic research is being systems as a whole. In such an analysis as this, the carried on in universities and other research establish- evaluation of a new development in relation to a ments, supported largely by direct grants and contri- time scale is necessarily very important. For example, butions from philanthropic persons and organizations we must predict with reasonable accuracy the expected who expect no direct personal return but are interested time at which new heat-resistant materials will be in the conduct of this research because of its obvious available for incorporation in engines, which in turn benefits to mankind. This research and its support will require further predictions as to when these enmust continue by all means; it provides the broadest gines will be available for aircraft. The engine developbase for scientific inquiry with the greatest possible ment with its substantial influence on speed of airfreedom to the individual. craft will thereby necessarily have a substantial inWith the substantial funds for research available fluence on the armament development program and on from the public treasuries today, the Government re- the percentage of effort which should be devoted to search executive must have a keen sense of balance to new armament. avoid the pitfall of putting Federal Government funds The research executive must therefore not only unnecessarily in competition with other funds in re- evaluate the fields of promise but must, to the fullest search, which might unduly influence the trend of extent possible, predict the time table on which basic research and might drive private enterprise out of items can ultimately result in improved weapons sysprofitable and fruitful fields. tems. It is, therefore, the policy of the Air Force to per- Because we have weapons systems that are continuform or sponsor necessary studies and experimenta- ally changing or subject to change, the Air Force is

49 Qualifications of Research Executive sponsoring basic research in many areas of promise fore being devoted toward improved radar techniques. to the Air Force. Necessarily, these areas are very broad, This objective of locating targets and recognizing them including all branches of science. The Office of Naval necessarily brings the research executive again into Research is sponsoring research in physical sciences. contact with many problems of present-day physics, Other Air Force activities are dealing with other electronics, mechanics, and mathematics. sciences. While many research executives are devoting HIT TARGET ACCURATELY their full time to one or more of these sciences, there Once the target is located and recognized, weapon is also a need for individuals of broad appreciation Once the target is located and recognized, weapon who can recognize, evaluate, and coordinate this re- effectiveness is measured directly in ability to hit the who can recognize, evaluate, and coordinate this research across the board, looking always for greater target accurately. The research executive must see effectiveness and greater economy. to it that all promising avenues are explored which improve the accuracy of hitting the target. For each RAPID AND EFFICIENT TRANSPORT per cent increase in efficiency here, there can be a A large part of the Air Force mission involves the corresponding reduction in materials, personnel, and rapid and efficient transport of bombs, ammunition, dollars to accomplish a given mission. personnel, and cargo. Two fundamental military ABILITY TO CAUSE DESTRUCTION characteristics of great importance are speed and range. Also, we need to maximize destruction. Through Also, we need to maximize destruction. Through Aviation has grown from the development of these research and development we must improve bombs, two characteristics. Speed is essential for effective pene- warheads, and fuzes, which involve the whole range tration and survival of aerial vehicles and, in many of developments of explosives, propellants, and the cases, for the accomplishment of a mission within a.. cases, for the accomplishment of a mission within a like. It is interesting and somewhat gratifying to see limited time scale. Range ins likewise important. how explosives research is tying in with the propulsion Increased performance in both these areas depends fields. Both these phenomena involve chemical reacprimarily upon research in aerodynamics and propulprimarily upon research in aerodynamics and propul- tions. The only difference is that explosions are often sion. more rapid than propulsion phenomena. A whole new Fundamental research in aerodynamics and propulFundamental research in aeroynamics and propul- field of chemistry is opening up to understand better sion is being financed directly by the Air Force. In to understand the kinetics of chemical reactions, to understand why addition a substantial and significant contribution some reactions are fast and others slow, why heat, cold, is made by the National Advisory Committee for Aero- and catalysts accelerate or retard reactions. Our renautics, an independent Government agency which is devoting its primary attention to these two fields. search executive must, therefore, appreciate the problems of chemistry. I am glad to say that there is close coordination between the administration of Air Force research and FUNCTION INDEPENDENTLY OF WEATHER AND DARKNESS that of the NACA. The Air Force is represented on Research and development is pointed toward the the NACA committees and subcommittees, and, in requirement that the Air Force be able to function turn, the NACA is represented on the Air Force Scien- completely independent of weather and darkness, to tific Advisory Board. Thus, you will see that a sub- take off and land in a continuous procession, using a stantial phase of the job of an Air Force research relatively narrow flight strip in conditions of poor executive is one of coordination, of keeping the Air weather and visibility, as well as operate under all weaForce program from duplicating that of the NACA in ther conditions. In addition, we want to take adthe development of rapid and efficient transport. vantage of weather, as, for example, using winds to LOCATE TARGETS AND RECOGNIZE THEM increase the range of aircraft, clouds for cover, and bad weather to reduce enemy fighter opposition. The Aviation has long since passed the time when mere Air Force Cambridge Research Laboratories, in adability to get off the ground has significance. Many dition to doing electronic research, are therefore purphases of militar operations.requirelocatinga dition to doing electronic research, are therefore purphases of military operations require locating targets suing a geophysical research program aimed at a better and recognizing them.. understanding of the weather and predicting future In many of these instances long-distance navigation meteorlogical conditions as well as determining what becomes an essential factor, guided by celestial obser- effort is required to operate independently of the vations, dead reckoning using inertial systems, Loran systems, radar, or a combination of these. In order to weather. This activity increases the fields of interest systems, radar, or a combination of these. In order to simplify the work of the air crews and to make guided missiles possible, projects are under way to make these DEFEAT ENEMY INTERFERENCE systems automatic to the fullest extent.'We must defeat enemy interference to insure that Also, the target must be recognized. Visual methods our airborne vehicles can perform their assigned misof recognition are hampered by weather conditions sions. In the active phase of this program, long-rangeand by the altitudes and speeds at which airplanes penetration fighters are being developed to strike at are now being operated. A substantial effort is there- the source of enemy resistance, with new promising

Qualifications of Research Executive 50 developments to increase the effective range of these therefore be a great receptivity to new ideas regardless fighters. The passive phase of defeating enemy inter- of origin, and a humble desire to aid and encourage ference can be accomplished through superior per- the research organizations and workers themselves. In formance of aircraft, confusion of defenses, evasive this way, he can make certain that public funds are tactics, countermeasures, and bomber-defensive arma- wisely spent to bring results of value to the Air Force. ment. This requires considerable attention by the Air A good administrator is therefore much more valuForce research executive to aerodynamics, propulsion, able as an executive than a clever technician or invenelectronics, armament, materials, etc. tive genius. Furthermore, the research executive must COMMUNICATION FROM GROUND-TO-AIR AND AIR-TO-AIR be a good salesman, not in the high-pressure sense but Utilization of aircraft for long-range bombing and rather in the true sense of what constitutes salesmanUtilization of aircraft for long-range bombing and ship. He must know his product- research- and the reconnaissance, for defense, for short-range tactical ship. He must know his product - research - and the capabilities of his organization to undertake research offense, and for intercontinental transport require dis- t and to produce. He must see the applications of his tinct types of communication equipment from ground- products to the solution of useful problems of h to-air and air-to-air. Here attention is being given not sponsor, and he must be able to present through reonly to the power and frequency of equipment but on. y. q y q ~. ~ports, talks, and other sales techniques the ideas which also to the provision for multiple channels, automatic control, and miniaturization to conserve space and weight. The principal requirement of this develop- Facilities ment falls on the electronic research executive, but It is an Air Force policy to use existing facilities, there are also human aspects when it comes to making wherever situated, to the fullest extent rather than go communications intelligible. out and build additional facilities. In many cases, DEFEND HOME TERRITORY however, new fields of research do require new facilities, and the costs of these modern research facilities The possession of nuclear weapons by a potential are so great and so specialized that private capital is enemy increases the necessity for defense of home ter- neither available nor disposed to make the necessary ritory, which likewise involves research in aerody- investments. In these cases, the Government necessarily namics, electronics, propulsion, and armament. The must construct research facilities. Examples of such existence of high-speed aircraft reduces the time avail- facilities are the following: able to the defensive forces and emphasizes the need 1) Wind Tunnels. A few years ago very acceptfor superior performance in interceptors, superior per- able wind tunnels could be built for a few hunformance of the detection equipment, and reduced dred thousand dollars. They were usually built with time to translate the warning of an enemy approach private funds and located either at industrial plants into the direction of the interceptor aircraft. This is or at colleges and universities. However, today with a challenging problem in systems development. the great emphasis on transonic and supersonic flight, EFFECTIVE UTILIZATION OF PERSONNEL very much larger and more expensive facilities are reUntil now we have discussed primarily the Air Force quired. Under the unitary plan, Congress has given research and development program in relation to phy- the NACA $75,000,000 for new supersonic tunnels sical equipment. Equally important is the research and has authorized the Air Force to construct the program in human resources to increase the utilization Arnold Development Center, at a cost of $100,000,000, of personnel by the best possible selection, technical which will embrace three major facilities: training, and job engineering. The research executive a) A high-altitude engine test facility for testing must be familiar with and understand the techniques new power plants. and results of this research program in human re- b) A 16-foot propulsion wind tunnel for full-scale and results of this research program in human resources. testing at supersonic speeds under simulated altitude conditions. Conduct of Research c) A 40-inch hypersonic wind tunnel operating up Having discussed the types of research in which the to speeds of approximately Mach 10. Necessarily the Air Force is interested and with which the research location of the site for this facility was determined to executive must be familiar, I would like to discuss the a large extent on considerations of electric power. manner in which the Air Force conducts its research. Consequently, Tullahoma, Tennessee, was chosen While in many organizations research is conducted among the possible sites. almost completely within the organization itself, Air The cost of facilities such as the above-mentioned Force research is, to a very large extent, done under places a significant responsibilty on the research excontracts with universities, research institutes, indus- ecutive, for he must determine that they are needed trial concerns, and through other governmental agen- and properly utilized. cies, such as the Bureau of Standards. The predom- 2. Flight Test Range for the Airplane and Guided inant characteristic of the research executive must Missile. An important part of the research and develop

5 1 Qualifications of Research Executive ment program involves actual flight testing of the the facilities will be the key to a new development. finished experimental vehicle to determine its per- Moreover, the facility must be completed before reformance, behavior, and capabilities. For this purpose, search can be undertaken. The research executive the Air Force has established a large flight test base must, therefore, be continuously alert to these new known as Edwards Air Force Base, at Muroc, Cali- developments in order that the facilities can be initifornia, on the Mojave Desert, where many natural ated in time to become available when needed. runways and landing areas abound. The Navy, as well Conclusion as the NACA, uses this facility for aircraft testing. For flight testing of missiles, the military depart- In conclusion, therefore, the research executive, ments operate missile test ranges at Pt. Mugu, Califor- whether an officer or civilian, must be a person with nia, and White Sands, Almagordo, New Mexico. The the following qualifications: Air Force is developing an additional long-range prov- 1) He must understand why the Air Force is coning ground under the joint cooperation of the United ducting research and under what conditions; States and Bahamanian governments, in which mis- 2) he must be familiar with the over-all mission of siles will be fired from Banana River, Florida, south- the Air Force and able to translate that mission into easterly, along the chain of the Bahamas. This location all fields of science, and, in turn, translate scientific was chosen because of the geography of the area and capabilities back into military ones; the suitability of intermediate siting stations. Thus, 3) he must be able to see research in a clear enough the facility picture broadens our research executive perspective to predict and evaluate its rate of developstill further as many specialized facilities are involved ment on an adequate time scale in order to keep all in the program. elements in step, shift emphasis from one area to In all these facility matters, the fluidity of the tactical another, and plan facilities as needed; developments is a very important factor. Besides, the 4) he must be a leader, receptive to new ideas refacilities must be completely modern. It is rather ob- gardless of their origin, who aids and encourages revious that a subsonic wind tunnel would be extremely search, but does not dictate it; and limited in predicting the performance of a supersonic 5) he has to be a salesman, who knows his product airplane. It is not always quite so obvious that other and the capabilities of his organization, sees applicaresearch and development facilities must likewise be tion to solution of problems of his parent organization, kept in step with new technical developments. Often and presents his ideas convincingly. DISCUSSION OF DR. LOMBARD'S PAPER COL. SEILER: The comments of the last two days to take one concrete example, has the facilities to do indicate to me that industry is operating research on research in hypersonic aerodynamics, among other a fairly short-time basis, that is, if the project does not things. We visualize that the results of this research seem to indicate a possible payoff in two or three will be of great value in 1960. We are looking at the years, there is small probability that it might be under- facilities picture in terms of the kind of research we taken. I was wondering whether Dr. Lombard would will be doing in 1960, which means that we should care to comment on the time range of research for the start to process the facility expansions now, if we are Air Force. going to have the facilities fully operating by that DR. LOMBARD: Well, our time scale is much longer time, or shortly before that time. I would say, therethan that of some industries. It seems that any new fore, that at least elements of the Air Force research development in the physical sciences, in fact, in any program are aimed at very long-term objectives. of the sciences, can have a direct application to Air Even the problem of developing an airplane itself, Force activity. That is a' very broad statement. It has is a five- or eight-year project, and we are looking for been our observation that research has enormous research that will lead to new developments far away fields of application to the Air Force, and our time in the future. scale is not only the solution of immediate problems DR. KLOPSTEG: I wonder if Dr. Lombard has a on a two- or three- or four-year basis (and those are comment on the "research mindedness" of Congress, definitely the fire-drill program problems), but it considering the fact that the committee on appropriais at the same time looking forward to research that tions cut out the $475,000 proposed for the National should be conducted and will be conducted on a Science Foundation. much longer time scale. DR. LOMBARD: I think Dr. Killian is giving a preThe Arnold Engineering Development Center, just sentation of the National Science Foundation this af

Qualifications of Research Executive 52 ternoon. I will let Dr. Killian answer that for the cal and climatic conditions are there. We also have a Department of Defehse. I have no comment on that. climatic hangar in Florida, where we can get down I do have one comment, even though it may be minus sixty-five degrees, even in Florida. trespassing on this afternoon's conversation on the A VOICE: One approach in technological research National Science Foundation. It is our concept that the is the cost of the production of airplanes. Have you National Science Foundation will not represent a any information as to the cost of research and developmonopoly of Government research. ment as related to the cost of airplane instrumentation The N S F will attack certain phases of the prob- and things of that nature? lem, but we still visualize a very substantial usefulness DR. LOMBARD: Let me see if I understand. We for Air Force research, for Navy research, Army re- know what our research and development appropriasearch, as well as for private research, and research in tions are. They are approximately a couple hundred the NACA, and other Government agencies, even million dollars a year-that is, the entire research and with and after the full operation of the National development appropriation, including the building of Science Foundation. prototype aircrafts and all the equipment that goes DR. SEEGER: May I add a footnote: Government into them. research, particularly military research, is often distinc- The production appropriations are one or two biltive in that Government must support long-range basic lion dollars a year, so that the research and developresearch in certain fields in which the law of supply ment budget is of the order of ten to twenty per cent and demand may be inoperative for certain stagnant of the overall "hardware" program of the Air Force. periods, e.g., ordnance research and development, in- In view of the conversation we have already had cluding aeroballistics, high explosives, etc. on definitions, I am reluctant to state how much basic Incidentally, an additional question is not just what research we do in the Air Force. We do consider it to we are to expect of research executives, but rather be something of the order of twenty million dollars, where we are to get persons for this type of work. depending on how you define it. MR. MORSE: I am interested to know: (1) to what MR. MAIER: With your permission, Dr. Lombard, extent is the Air Corps going to attempt to centralize I think I can answer that question. On an engine deany of their research activities and engineering facili- velopment, it normally takes a period of eight years. ties; (2) is the Office of Air Research really going to One specific engine that I know about cost fifty-two exist or not? million dollars to develop, and the price of the engine DR. LOMBARD: Colonel Seiler is the chief of the in mass production runs somewhere around eighty Office of Air Research, and it is his intention, and it thousand dollars apiece. That will give you a time is mine, and it is that of many of the higher Air scale, as well as the cost. That does not include, howForce officers that the Office of Air Research will exist ever, the cost of special facilities for producing the and become an important agency for basic research. engine that had to be supplied. In answer to your other question as to centraliza- DR. HOLLAND: One other related question, to get tion of facilities, I believe that the tendency is going a kind of basic comparison with costs of doing research to be in the opposite direction. The tendency in the in industry, is the investment in research and developfuture will be to locate research facilities where the ment per professional research worker. I don't attempt predominant resources are that are essential for the to differentiate between research and development, operation of that facility. but taking the Air Force overall R and D activity, At the A E D C the predominant requirement is what is the investment in facilities per professional electric power, and consequently it was built in Ten- research worker? nessee. In the case of electronics research, the predom- DR. LOMBARD: I am sorry, I just can't answer that inant requirement is not electric power but manpower question. I wish I could, but I don't know. I know skilled in the fields of electronics, guidance mechan- what the question is, but I don't know the answer to isms, and the like. There is a very substantial resource it. in that field in the Cambridge, Massachusetts, area. DR. FURNAS: If I remember correctly, the NA CA That is why the Cambridge Research Laboratory is investment in facilities at the present time is about located there. $150,000,000, and they have seven thousand emOther research activities are being located in other ployees. areas where there is a predominant characteristic which DR. LOMBARD: The A E D C will be a very expenmakes it favorable to the particular kind of research. sive project. They will have facilities valued at $157,We have research going on in Alaska and on Mt. 500,000 and one to two thousand employees. That will Washington in New Hampshire because the geophysi- be fifty to one hundred thousand dollars per man.

53 Qualifications of Research Executive - IIIby RAYMOND STEVENS Vice President, Arthur D. Little, Inc. The attack on any research problem is helped by If we now supplement a literature search with obserrestatement in new terms and in new perspective. Fre- vations on industrial research and research administraquently a new and more basic problem emerges. Less tion as it exists in some of the three or four thousand frequently, what seemed an irritating difficulty reap- organizations, we are struck at once with the tremenpears as an attractive opportunity. In this instance an dous variation in policy and practice. It is almost like intriguing and constructive new aspect of general cor- looking at a series of embryos and tracing the history porate management seems involved. of the evolution of the organism. The man in charge The problem as given originally was a description has various titles, with about four out of five called of the qualifications of a Director of Industrial Re- "Research Director." Other designations are obviously search. To resolve the matter properly we need a fresh dated, being still embedded in the past history of review of the organized provision made by management search and management. They describe the eras when for a company's future. The professional requirements research was an odd man, not a department or even are stated by Dr. Norman Shepard in a chapter in a permanent staff member. Then, a little later, he was the Industrial Research Institute book, Research in a chemist troubleshooting for management. At a later Industry. Technical knowledge and skill coupled with period we see the chief chemist and the cornerstone experience in research are accepted as essential pre- of a research department. When the physicists, biolorequisites. The research executive must have absolute gists, mathematicians, and economists entered the conauthority over the research program, and its formula- ception of research, the embryo took on the charactertion is his major task. He must prepare the research istics of the species. Then the director of research budget, with the technical and business judgment that appeared. By the 1930's it was recognized generally is presupposed. He must sell his budget to manage- that research belonged in some way to top management. He must maintain a healthy and constructive ment, and we find the title Vice President-Research. mental atmosphere within his staff and must concern Without spending much time on the company whose himself with salaries and status factors. He must build research executive is still a "chief chemist" concerned and maintain adequate liaison with production and primarily with technical quality control, we pass sales. through the various stages to the higher development Dr. Shepard also notes the frequency with which where an officer of the company sitting with the execumen are found in the post without the necessary scien- tive committee and the board of directors superintends tific training and research experience. Unfortunately, a modern profit- and growth-producing activity that it is equally true that the post is frequently filled by a has far outgrown the industrial research function we man unwilling or unable to qualify himself in the knew even a few years ago. policy, financial, and top administrative aspects of One of the most recent phases of functional dehis important assignment. His stature in the com- velopment in industrial research is systematic operapany's policy deliberations does not then equal that tions research, an outgrowth of the war. It brings of other executives. He and his associates are frus- applied mathematicians into the research orbit in a trated by decisions made for them "above the point large way. It emphasizes the breadth essential in the where the facts are gathered," and research fails to modern research executive, for it is concerned with top realize its potential in the company. policies-and with top policies in which science and Giles E. Hopkins has written that research manage- engineering in their narrow sense are negligible facment provides the company with background for tors. Such matters as, for example, distribution, traffic, company policy, for interpreting technological trends plant location, and inventory policy are among the with an understading of economic values, and selling more simple and obvious elements that are involved. the results to those who should utilize them. Apprecia- Yet the basic method is that of the research man, and tion of dollar values and leadership are among the re- the administrative experience and the skill required of quired abilities. In his opinion research administration him are identical with the requirements for highly is a growing new profession. Rector emphasized that developed research. It is merely another phase of the the research director needs the drive which translates developing embryo We are not even sure it is research accomplishments into business results, and a the last and highest. few other elements of a job specification for a research We find a correlation between this high developdirector are found scattered in the technical literature. ment and the strength, growth, and profit record of

Qualifications of Research Executive 54 the companies concerned. In some companies with There are other characteristics. We find, for examspectacular growth and diversification records we find pie, a large percentage of men schooled in a particular the term "Research Director" clearly inadequate as class of subjects. The physical sciences, engineering, title for the responsible officer. In the gradual develop- and mathematics are especially prominent. But the ment of research, connotations have grown like barna- most important and essential characteristic of the new cles on the words used. The term "research" itself has activity at its best is a new basic philosophy. This philbecome inadequate, and the term "director" may carry osophy is often the one significant difference between an implication that his responsibilities lie below the the successful and the unsuccessful sponsorship of retop policy-creating echelon. The dictionary gives little search. support for "manager" outranking "director," but in An industrial corporation should be a living thing, some companies the precedence exists. Where it does, recreating itself in frequent generations, adapting itunrealized opportunities probably also exist. self to changing conditions, growing where growth is Many attempts have been made to define industrial desirable, alive, and healthy. It cannot be life-like if research briefly. Industrial research needs redefinition, its various functions are all static, as were so many and redefinition in more than mere words. Only when gay-nineties production and sales functions, for exthat redefinition is accomplished can the qualifications ample. It is indeed the exception to find the otherwise of a research director be defined properly. It is not an undesirable sales, production, or financial executive easy task, but some aspects of it can be considered who is constantly seeking major change. The one functo advantage. In the face of the observed tremendous tion occupied particularly with responsibility for difference between the various kinds of "industrial change, growth, and adaptation to surroundings is the research," a brief and all-inclusive definition is im- function with which we are now concerned; call it possible. The terms "research," "industrial research," research if you will, but drop any limitations to the and "research director" now have many meanings to term acquired by early industrial usage. many men. If we start with this concept of the responsibility It may help to drop the barnacle-clustered words of an executive for whom we are writing a "job speciand start without them. How can we describe in simple fication" sheet, we have at least a reasonable chance English the highly developed and still developing at a definition that will fit the few known highly deoperation that makes some companies living, growing, veloped types of our species. The "job specification" dynamic organisms, companies so much more humanly sheet becomes fairly easy. Matching it with an availinteresting and financially attractive than those which able candidate is another matter. lack this new thing? It may be of interest to look at a description of a As vague outlines begin to.emerge, the first reaction c"man wanted" for research director at the policy level. may be a bit of a shock. Human and financial aspects Here is one used recently by Rogers and Slade, a New loom large. Dominant is the general approach to prob- York firm specializing in the location of industrial lems with which physical scientists have been indoc- executives: trinated, an approach that requires a period of patient "He will be a member of the company's top manageapprenticeship for its full comprehension and most ment group. In addition to planning, directing and supercertainly for its skillful application. The general vising his own Division, he will be expected to maintain method is useful over a wide range of problem types- effective liaison with the sales, procurement, plant engineering and production departments... many of them far outside the laboratory. many of them far outside the laboratory. "Works in cooperation with top management in the Another familiar touch is emphasis on measurement. planning of an over-all technical research program, deBoth metric and English units are required —and curl- signed to meet both long-range and immediate operating ously the one series of metric units most universally problems. used is that based on the dollar. It is a unit familiar "Plans, develops and directs the activities of the Reto the executives of the older functions of industry search Division, designed to assist the company in mainand one into which scientific measurements must be taining its excellent position of leadership in the industry. translated for talks with them. "In collaboration with operating executives, assists in There also looms brutal realism, an insistence on finding solutions to production and related problems. "Prepares ahd issues practical interpretations to manfindings facts, facing them, and acting on them. Every agement regarding current technical problems and trends effort is aimed at combatting wishful thinking, ration- affecting the industry." alization of unsound situations, impulsive action. Another states that: When facts are available, some of the human touches When facts are available, some of the human touches "Since the company is deeply interested in the develof old-time management seem in danger, but in their opment of a long-range, practical and broad-gauged replace come a broader vision, fuller realization of pos- research program, a man would be preferred who possibilities, fewer tragic failures. The whole industrial sesses these general qualifications: organization is catalyzed into a new and better way of "'Skill in defining and presenting recommendations life. with respect to both long-range and immediate objectives;

55 Qualifications of Research Executive in planning work programs in accordance with approved ties and-yes, of its personnel. For only with this objectives and policies, and in directing and supervising knowledge can the future be planned and shaped. a technical staff in securing results in conformity with With this type of definition we see the shortcomWith this type of definition we see the shortcom"'A keen appreciation of the problems of department ings of any narrowly trained technician. We see what heads and their respective staffs, particularly with respect handicaps must be surmounted by the scientist or ento production, sales and finance. He must be practical and gineer who seeks this type of responsibility. We begin possess sound business judgment, balanced with vision and to see why men who have recently become great in the courage of his convictions. this field are seldom recognized for their personal ac"'He should possess a highly developed sense of team complishments as scientists or engineers. But, by the play, and the capacity to put into practice the standards same token, the truly successful men have been thorby which top management measures the effectiveness of oughly indoctrinated in the philosophy of the scientific the activities for which he is responsible. Too, he must be able to interpret to top management the potential results method and the methodoloy of physical-science rewhich may be attained through an adequate research pro- search. They have known how to use the methods gram and activities related to such a program.'" and how to talk the language of their staff of scientists, One used recently by Handy Associates, Inc., states engineers, and economists. in part: Because the language of science is little known to "A qualified Director of Research will be experi- the business man, and because the language of the enced in laboratory techniques and practices... He business dollar is little known to the scientist, the will have a valid sense of commercial application and top executive is, above all things, an ambassador and can objectively evaluate products in respect of merit, an interpreter between two worlds. Because the lancost and consumer acceptance.... guage of science and engineering is an obvious re"As a counselor in research matters to the manage- quirement, he is most apt to be deficient in the lanment of his company, a research director will have guage of the dollar. He must know the sources and a capacity of personal salesmanship to the point where limitations of funds and the relative merits of the many he can get acceptance of his ideas by those responsible demands for the corporation's money. Balance sheets for carrying them out." and statements must be as familiar to him as critical These specifications are at a high level. They are tables. He must know something of markets and pracfar removed indeed from George Eastman's first out- tical commercial economics generally. Dr. Shepard has line of the research man, back in 1890: "It will not with reason termed The research director's job... be long before your firm will need a practical chem- the most important one of all the technical positions ist... the best way is to make application to the Pro- in a company, and, without danger of over-emphasis fessor of Chemistry in some good technical school and one of the most important in the entire executive have him recommend two or three first-class boys. You family, when the long-term success of a company is can... take your choice. If he is any good he will be considered. the most profitable man you can hire." Summarizing: We need a new concept of industrial Yet, returning to the delineation of the good pre- research devoid of the limiting connotations acquired sent-day research executive, even these specifications in its early development. We visualize a crystallized when compared with a broad definition of the func- corporate function concerned with company growth tion involved leave something to be desired. What we and change. The executive responsible should be inreally are talking about is responsibility for the future doctrinated in the basic methods and philosophies of the company insofar as that future differs from the developed since Bacon in the laboratories of the physipresent activity. We know that many a highly suc- cal scientists. He must maintain and inspire a staff of cessful modern company will be completely different specialists whose language he must speak fluently and twenty years from now-in size, in product, in market, whose mode of thought he must instinctively comprecertainly in personnel. The production, sales, and per- hend. He must interpret staff findings honestly and sonnel managers are at no one time primarily con- clearly to his lay associates, and, conversely, he must cerned with this future. The executive we are now dis- be qualified to interpret the area of management policy cussing is almost solely concerned with the future. and finance to his staff in their guidance toward propIt is here that we begin to define our problem. In- er targets. He must have breadth, vision, something sofar as the many restrictions permit, our man is re- of the urge of the crusader, and something of the sponsible for the character of changes in his company enthusiasm of the pioneer and the promoter-all for much of its future. He must be a prophet, a gam- tempered with awareness of the practical possibilities bler, a man of vision, a man dissatisfied. He must and an honest and realistic allowance for the limitachange, adapt, create. He and his staff must know bet- tions of men and money. ter than any others the limitations and possibilities This is no place for a small or lazy man, but, forof the company's products and markets, of its facili- tunately, a superman is not required either. The of

Qualificationls of Research Executive 56 ficer in charge can now supplement his own blind are highly developed, but many of the others need spots with men experienced in any of the several prin- only expansion in scope in order to cover wider fields. cipal areas. The general procedures and policies to be If we are to judge by the results of the very few piofollowed can be examined in a few working models neering leaders in this activity, the promise is great now in successful operation. Only a very few indeed to those companies which follow. DISCUSSION OF MR. STEVENS' PAPER MR. ABRAMS: I would like to inquire how successful of the detailed knowledge of balance sheets and statethese management-engineering corporations are in ments and financial policy. selecting research executives, as proved by subsequent In addition to his scientific technical training, he performance. must get the equivalent of what some of the business MR. STEVENS: I would say that they have been administration schools are giving. And he must acvastly more successful than the choices that have been quaint himself and learn how to deal with and talk made on the recommendation of one or two men in the language of the businessman. the field. I believe that most of the men here would meet I have seen several failures of men who were ap- those requirements. pointed on the recommendation of some research M4R. ASBURY: Many of MlIr. Stevens' remarks are director who was a friend of the president of the pertinent when it is a question of bringing in a man company. from the outside to fill an important job in the reDR. FURNAS: I wonder if you can summarize your search organization, but there is nothing that helps remarks in one phrase? Is the essence of your observa- the morale of an organization more than to take a tions that the man you are talking about is the vice man from within the organization. president in charge of change? I presume Mr. Stevens was thinking of companies MR. STEVENS: Dr. MNaurice Holland used a term who are starting a new research organization and are last night, "Vice President in Charge of the Future." trying to find somebody not available in the present But the old-time organization chart had a president, organization. However, with a going research organization, the a sales manager, production manager, and a treasurer. The president probably thought that all the functions problem is to have a system for developing men to take the place of the executives that are already there he supervised were covered. Obviously, they weren't, ~~~~~~~~~'he.pvs e vrl.e un't and to fill new executive positions, and under such because he had left out the one life-giving function exective that has been sorely neglected ever since industrial circumstances executive training becomes an imporcorporations began to assume their present form. And r asort of a man might wtell be "Vice MR. STEVENS: I am seriously concerned if I left the title for that st well "Vice that impression, because bringing in a man at the President in Charge of Change." top level on some of these organizations is a very MIR. ABRAMS' ~Vhat recommendation would you serious problem, and the record has not been too make to a young man, let us say around thirty, as to how he can become a research director some day? What I would much prefer to see the men in the organi-ort ofacio soldheI would much prefer to see the men in the organization begin to work toward the new positions and MR. STEVENS: MV[y first recommendation would be begin to see some of the responsibilities it envolves. to decide whether he really wants to be one or not. DR. HOLLAND: I would like to make one point that It is a hard life, because one has to have two fields was touched on yesterday and that is the question, and cover them thoroughly. whether we shall grow from the hard technological One has to be an expert in the scientific field and core of our own company research organization or at the same time command respect at the policy level grow by the merger route. and in the administrative field. Otherwise, he is not A friend of mine in one of the large petroleum a competent man of the type that I am talking about. corporations for years has been making a study of He certainly should take every chance to meet with some of the chemical companies in the United States. and obtain the feeling and the sense of the lay bus- I think his studies clearly indicate that the growth is iness man, who is concerned primarily with policy. largely by accretion, the merger route. I'note that If he doesn't understand the language of these men, some of our friends are becoming the interpreters to he is not competent. Our technical schools do not management as to what organized research means in teach the men much along that line, certainly none terms of company growth and development. They are

57 Qualifications of Research Executive becoming technical advisers to the president on what what is needed in a research director, I began to think I call "the technological strategy" of company growth. of applicable metallurgical and chemical terminology. I would like to present this idea: Let us come up Some of these terms are rather interesting and approwith the concept of the Vice President in Charge of priate, others are interesting but not appropriate. Research,... of the Future,... of Change, or what- "Elasticity" is one, also "high endurance limit" and ever you want to call this overall assignment. Our "resistance to corrosion." Vice President of the Future is a product of a techno- The research director is "abraded" and "eroded" by logical background and of business management, who sales, management, production, engineering, and all keeps his fingers on the pulse of business and competi- the rest. tion and taxes and all the rest of it in order to be in "High strength" is also good, and "low fatigue." In fact a "technological strategist" for his company. That metallurgy, we hear an "order-disorder transformation" is what the job amounts to. that might be applicable, since a research director Some of the chemical companies, such as General should be able to transform things in a disordered Aniline and Film, have a director of development, state into an ordered one. who is a Ph.D. He is not director of development in You can go on to say "bright and polished, like a the sense we think about; he is director of the develop- metal surface." "Miscibility," the ability to mix well, ment of the corporation. His job is to view the corpora- is another attribute, for the research director should tion as a whole and how best to use technologically be a good mixer. "Orientation" is another term-seeing trained men. I think such a concept is beginning to himself in the proper perspective in the organization. emerge. You might have a little fun kicking around "boiling DR. KLOPSTEG: If the word development is synony- point," too, also "freezing point," also "throwing mous with or connotes "change," it would seem ap- power" and "toughness." propriate to have a vice president in charge of develop- Of course there is the phenomenon that is very well ment, tying in with Dr. Furnas' idea. Incidentally, known in metallurgy under two names. One is "aging," the use of the word "development" in that connota- the other "age hardening," which means getting hardtion might help wean us away from the misuse of the er as you grow older! I don't know whether you can word "research." have any fun kicking that one around or not! MR. WILLIAMSON: I entered the research business As far as "brass" is concerned, I prefer to leave that by way of metallurgy and chemistry. In thinking about out! ROUND-TABLE DISCUSSION R. J. SEEGER, Secretary It was generally agreed that there is not yet any tive. The ability and desire to handle people were also valid concept of an ideal research executive. Such an stressed. individual, at present, may be called upon to perform The primary need now is a reliable list of items in various functions, the exact pattern depending upon a usable form for evaluating or selecting research exthe specific job and its place in a particular organiza- ecutives. The concepts must be defined precisely in as tion. The essential responsibilities seem to be: (1) quantitative terms as possible. Records of specific inobtaining ideas, i.e., planning; (2) selling ideas up stances of good and bad performance were recomand down, i.e., promoting; (3) administering the re- mended as an initial procedure in this direction. sultant program, i.e., supervising, etc. This Round-Table group, consisting of about 30 Mostindividuals, h oweer, werotepiiotat persons, recommends that the Conference request Dr. Most individuals, however, wer o the of the opinion that Flanagan to conduct an investigation to ascertain a agreement could be reached on the basic virtues de- short list of critical requirements for a research execusired in a research executive. A research background tive on the basis of information to be solicited from is regarded as the primary requisite for such an execu- the members of this conference, et al.

Fourth Session Albert E. White, presiding Director, Engineering Research Institute University of Michigan OVERHEAD As a Factor in Sponsored Research W. K. PIERPONT Controller, University of Michigan THE SUBJECT of overhead as a factor in sponsored fice of Naval Research will very likely spend someresearch reminds me sometimes of the White Rabbit where between 20 and 25 million dollars in colleges in Alice in Wonderland. Wherever Alice was or what- and universities. It has been estimated that the Fedever Alice was doing, the White Rabbit usually ap- eral Government through all its agencies will spend peared on the scene under one disguise or another. A possibly up to 100 million dollars this year for research discussion of overhead, it seems to me, almost always work in colleges and universities. This large amount appears on any program devoted to the problems in- of money to be spent on the part of the Federal volved in the administration of research, and I am Government indicates that research has a very imporvery happy to have the privilege today to discuss tant place in the operations of the Federal Governthis subject with you. ment, whether for the protection of the people or for I think it might be helpful to take a brief look at their well-being in other ways. From the point of view sponsored research itself, the amount of such research of the universities, this sum of money spent on rein dollars and cents, and the place of such research search gives sponsored research an important place in the Government, in the universities, and in private in universities. For example, the total research proindustry. gram at the University of Michigan for the past When World War II ended, all of us were well aware year amounted to about 5 million dollars, so that of the contributions made toward the winning of the the Federal Government's research funds would carry war by research work in many fields of knowledge. Not twenty programs the size of ours in twenty different only were we impressed by our great achievements in universities. And our program here is not small by producing the machines for war, but we began to any means, as university programs go. appreciate more fully than ever before the great con- In addition to the funds provided by the Federal tribution of physicists, chemists, engineers, and others Government, there are, of course, large sums provided working from the theories of basic science to the draw- for research by grants from foundations such as the ing boards of manufacturing plants. There has been National Foundation for Infantile Paralysis, the a definite recognition on the part of Government of- Rockefeller Foundation, the Carnegie Corporation, ficials, industrial leaders, and university personnel the Kellogg Foundation, and others. that research in basic, fundamental knowledge should The expenditure of these sums of money, when rebe continued if we are to have a sound base upon lated to the previous amounts spent on sponsored rewhich to build better techniques and better processes search or to the laboratory facilities and personnel for better products and better services of all kinds. available for research work, has resulted in the estabThat sponsored research is a well-recognized and That sponsored p o research is a well-recognized and lishment of new facilities on a large scale in many well-supported program of research is evidenced by the. fact that the Atomic Energy Commission expects to industrial plants and universities and in a fuller spend about 14 million dollars this year outside of its utilization of previously unused capacities. The use own laboratories. Its own laboratories, of course, will of these new facilities and previously unused facilities spend many millions more. The Public Health Service, costs money, to cover both direct costs and the overfinanced by the Federal Government, also expects head costs which are created in substantial proportions. to spend about 14 million dollars this year. The Of- With this brief background, let us look in detail at

59 Overhead the costs of research and the specific problem of over- research project for a sponsoring agency until it has head costs. been reviewed by certain members of our administraI think we should recognize first of all that research tive staff to determine if (1) the project proposed is is an activity for a purpose. There is a product to be compatible with our overall objectives, (2) we have developed, and this product may be, and is in fact, an a contractual agreement to be reimbursed for all costs idea, a theory tested, a technique evolved, or any incurred, (3) salary rates to be paid are in line with one of a limitless number of possibilities in the realm established rates for our institution, and (4) we have of knowledge. From the point of view of the nature the space available for the project. Once this is done, of the costs involved, this kind of a product is not we have to select and appoint the personnel for the different from any of the products of industry or com- project; we have to pay them, and, of course, we have merce, of the universities, or of the Federal Govern- to be sure that their net pay is correct after deducting ment itself. income taxes, medical and hospitalization payments, The costs of a product or service, tangible or in- retirement payments, bond deductions, and what nottangible, from the point of view of budgeting and fi- each one of these deductions must of course balance nancing such costs and from the point of view of con- exactly to two decimal points, and each one of these trolling such costs, may be separated into three major deductions requires a separate record for each individelements. Almost every elementary cost-accounting ual-; we have to purchase materials, supplies, and textbook devotes the first several chapters to three equipment for the project and do this right now, or elements: else the project will be held up; we have to keep an 1) Personnel employed specifically, exclusively, accurate accounting of all funds involved and prepare and measurably in a convenient sense for the par- invoices itemizing the costs incurred; we have to heat ticular product, or job, or service, or process under and light the areas used in the project; we have to prostudy-ordinarily called direct labor; vide water and sewage facilities; we have to sweep the (2) material, supplies, equipment, communi- floors, clean the ice and snow off the front steps, build cations (such as long distance telephone tolls and and maintain parking lots; and a thousand other telegrams), travel of personnel, and other non- things which, if not done, would be used as good exprofessional items of cost purchased or used speci- cuses for poor research results. On top of all this, in fically, and also in measurable quantities from a most instances, we have to advance the funds to pay practical sense-ordinarily called direct material salaries and other expenses and then seek reimburseand supplies and other direct costs; and ment. Usually, we obtain reimbursement without any 3) personnel, material, and supplies, equip- trouble, but if we stray from the well-worn path, we ment, utilities, and other costs which are not have to spend our time, money, and energy to obtain easily measurable in terms of a specific product repayment for the funds which we have previously or operation or which for purposes of con- advanced. venience are not charged directly to the product The first point I wish to emphasize today is that or operation under study-ordinarily called in- overhead costs exist in all reality; they are not the direct or overhead costs. figment of an overactive imagination nor the result In determining the total cost of a particular product of accounting hocus-pocus. We have innumerable items or service, overhead costs are assigned to the product of cost which we incur because of the sponsored reon the basis of the amount of a known factor of cost. search carried on by our staff and which we cannot, In most cases the amount of direct personnel costs is or do not choose to, charge directly to a research projused for the assignment of overhead, since, by and ect. I wanted to stress particularly that point, as it large and in most cases, there is a direct relationship seems to me that oftentimes we overlook this fact between the amount of direct personnel costs and the and are concerned only with paying for direct labor amount of overhead costs properly assignable thereto. or direct material, or travel, or equipment and are not The relationship of direct personnel costs and over- quite sure of the existence or the extent of indirect head costs for an entire enterprise or for a particular costs of one kind or another. part of an enterprise establishes a rate applicable to You may say, to be sure, that we would build and each particular product, and this rate is known as the maintain parking lots anyway. That is right, we would overhead rate. build enough for our instructional and administrative Now, one of the first points to keep in mind in this staff, but without research personnel on the campus question of overhead is that overhead as a cost prac- financed by sponsored research funds, we would not tically always exists in obtaining any product or serv- need to acquire the land and build and maintain ice. It may exist to a more or less extent, but it will so many lots. If some of you do think that parking lots always be present. This is true for various reasons. are not a serious problem here, I suggest you come We, at the University of Michigan, and we are not back in a couple of weeks when we are in full operation different in this respect from others, do not begin a and try to find a place to park your car, after 8:45

Overhead 60 in the morning, somewhere near the campus. of it but which, in my opinion and upon close analyA second point to keep in mind in studying the sis, has little to support it. The proposition, as I costs of products or services and the overhead rate is understand it, is essentially this: Overhead is admitted that there is considerable overlapping between the to exist, but the payment for overhead costs or, more three elements of cost, that is, direct labor, direct ma- precisely, the reimbursement for overhead costs, deterial, and overhead, and that the amount of costs pends upon who initiates a request for the particular represented by each element in a given situation and research project. To be practical about this propowithin a given total may vary, depending upon the sition, it means that we have to identify and attribute organizational structure of the enterprise, the account- to a particular person or group the origin of an idea or ing procedures used, the objectives of those who con- a theory which will be the subject of a research projtrol budgeting operations, and other such considera- ect. If-this would be unfortunate for the university tions. For example, the total cost of a given project involved-a professor thinks of something which should may consist of the following distribution of costs: be investigated further (and that, by the way, is one Example Example Example of the primary functions of professors), presumably A B C the university should be appreciative of any financial Personnel 60 65 70 assistance provided and should be content with reimMaterials and Supplies 10 10 15 bursement for direct salaries and materials needed for Overhead 30 (50%) 25 (38%) 15 (21 c) the research project. If, however, the Government -- -- -- or industry wants something investigated, the total 100 100 100 costs, including overhead, are a proper charge to the The overhead rate in these examples varies from 50 sponsoring agency. per cent on personnel in example A, to 38 per cent You are all well aware, of course, that ideas flow in example B, to 21 per cent in example C. Specifically, from discussions and from conversations as well as from items of cost which may be considered as direct costs solitary thought in a library cubicle. To determine the or as overhead costs, depending upon the circum- source of an idea for a research project is not only stances in the given case, include vacation pay, retire- a subject for argument but also a waste of time. Whoment and pension payments, telephone costs, secre- ever sponsors a research project should have no partarial help, janitors, and freight bills, and the treat- ticular reason for differentiating direct costs from inment of these items as direct or indirect in the cost direct costs; whoever initiates a project must recognize of a project will shift the overhead rate as much as that direct costs are always accompanied by indirect was indicated in the three examples just given. costs. And both types of cost have to be covered by Low overhead by itself is no indication at all of an some source of funds. efficient operation, and decisions made on the basis A variation of this proposition has been introduced of overhead rates or percentages without proper study by the Atomic Energy Commission. This Commission of all the costs involved may just as likely be in- has introduced two new terms to separate research correct decisions as correct ones! projects on the basis of interest. "Programmatic reThe second point I wish to make today is that the search" is research primarily of interest to the Atomic amount of overhead or the overhead rate should not Energy Commission and is directly related to the obbe a controlling factor in the placing of research with jectives of this commission. "Nonprogrammatic rea given enterprise. The rate of overhead is a factor search" is research primarily of interest to educational in the reimbursement of costs incurred in carrying institutions or to industrial enterprises. For programon research, and that is all. matic research, the Atomic Energy Commission is willIt is not within the scope of this discussion to elab- ing to negotiate the reimbursement for overhead costs orate on the factors involved in determining where incurred. For nonprogrammatic research, the Atomic research work shall be carried on. It is sufficient to Energy Commission is willing to allow 8 per cent of state that the particular training of the personnel total costs for overhead reimbursement on the assumpavailable, the books and pamphlets in the library, and tion that if the university is interested in the research the interest of the personnel in the objective of the work to be done, the university should provide funds research are examples of real controlling factors in for the rest of the overhead from other sources. (This selecting the site for research and far outweigh in figure of 8 per cent, by the way, has no particular importance the factor of overhead. In research, prob- validity and, apparently, was picked up from the overably more than anywhere else, it is very easy to enter head rate used in Public Health Service grants.) into a low-cost project and obtain low-grade results. With such a proposition to work with, if I were in It is not money that obtains research results, it is a the Atomic Energy Commission, I should search the research worker with a desire to find something new. country for as many individuals and institutions inThere has recently been advanced an idea or a prop- terested in Atomic Energy Commission projects as I osition which is somewhat intriguing on the face could find. Then I would let all work on a nonpro

61 Overhead grammatic basis and do no work on a programmatic colleges should not undertake research projects or acbasis, or in Atomic Energy Commission laboratories. cept research grants without recognizing that the I have not seen a financial report of the activities eventual burden inherent in a failure to cover all costs of the laboratories operated by the Atomic Energy of research, even if on only a few projects in the beginCommission, but I venture to guess that they are not ning, will create a serious financial problem. operated on an 8 per cent factor for overhead. There is one other point of view I would like to There is a fundamental issue involved in this propo- examine briefly today. In carrying on research, which sition. If the Atomic Energy Commission is not to us is our business and which has many of the charinterested primarily in a project, why should it spon- acteristics of any business, it is one thing to estimate sor any of the costs? Are taxpayers' dollars less rigor- or budget the cost of a project and another to agree ously controlled if they are expended for direct costs on an amount or, in other words, a price to carry on than for indirect costs? On the other hand, if a univer- the research project under consideration. In these sity is not primarily interested in a project, why should agreements not much difficulty is encountered with it undertake the project at all, regardless of the finan- direct costs, but overhead costs appear to be different. cial support proposed. It seems to me that a mutual There are, in general, two basic procedures which interest in a project by a sponsor and by an organiza- may be followed in determining the amount of overtion willing to undertake the project is the only sound head costs which must be covered to finance fully the basis upon which to work. If either party is not inter- total cost of a research project. One procedure is to ested, the project should go elsewhere or be sponsored negotiate an amount for overhead or an overhead by someone else. And this decision should be made rate for each project. And this negotiation between the without consideration of the problem of who shall parties may be repeated for each renewal or extension pay for overhead costs. of the project. In this negotiation, all the factors which You will recognize, of course, that research work for are pertinent may be brought to bear on the discuswhich reimbursement is not obtained for total costs sion in determining the price, and finally a price is is financially sponsored by some one, in a university mutually agreed upon. This procedure will soon like ours, either by student fees, state appropriation, result in the overhead rate of each project being the or other income of an unrestricted nature. It is this main point of argument, and the price will vary assumption of overhead costs to be met by other than directly with the amount or rate of overhead finally research funds which must not continue to go on in agreed upon. an unrecognized fashion. A second procedure is to use a standard or normal This idea of sharing the cost of research projects, overhead rate applied to a common factor in all reparticularly the overhead costs, has had wide accept- search projects to cover the overhead costs incurred. ance in the medical field during the past few years. Such a rate, admittedly, would be an average rate, Public Health Service grants, American Cancer Society high on some projects, low on others, but it would grants, and other grants have been made without full cover in total the overhead costs applicable to research recognition of the overhead-cost problem and have work. Since considerable discussion has taken place included in their grants an item of 8 per cent or 5 recently over the propriety of a standard rate for per cent of total costs to cover the overhead costs of overhead costs for all the projects of a particular the work to be done under the grant. At the same time university, I would like to consider certain aspects of that medical schools are carrying on research without the use of such a rate. proper consideration being given to the overhead prob- Universities, as you know, are nonprofit institutions. lem, these same schools are experiencing severe fi- They also hope to be "nonloss" institutions. In such nancial troubles. This may not be any more than pure institutions, there is no desire for more than a fair coincidence, and I would not want to say that a full overhead rate to cover actual overhead costs. Dividends recognition of overhead costs for medical research are not going to be increased, and the bonus is not would solve the financial difficulties of these schools. going to be increased, since such distributions of inBut I would draw your attention to the fact that we are come do not exist in our operations. There is a desire, discussing these days Federal subsidies for medical however, to hold administrative costs to a minimum education at the same time that the Federal Govern- and to obtain the most effective use of the trained ment through its Public Health Service is being sub- scholars and research staff on the university payroll. sidized by medical schools and universities in the re- When a policy is followed of negotiating each project search work carried on with grants which do not in- or contract independently of all others, it is necessary clude sufficient funds for overhead costs. I suggest to arrive at an overhead rate for the reimbursement of that a clarification of this situation will stop any fur- overhead costs which is applicable to a particular projther aggravation from this source of the financial prob- ect and that project only. Let us see where such a lems of these schools. procedure leads us. Also, I would like to suggest that other schools and From a university's point of view, this procedure re

Overhead 62 quires internal financial analyses of an endless variety. ever seems to think about increasing the amount paid In our own case, we have several hundred separate for overhead over and above the actual cost incurred. research projects using varying amounts of space; of If universities were to enter into an agreement which equipment, of heat, light, and power, of purchasing resulted in less than fair reimbursement for overhead department personnel, and administrative, supervi- costs, with which agency of the Federal Government sory, and consulting personnel time and effort. Ac- should it enter into such an agreement? Should it be cordingly, we would need to have a full-time staff work- the agency with the smallest amount of research on ing with our technical and research personnel if we the theory that the financial significance of this rewere to try to determine the overhead costs applicable search is not of much concern? Should it be with the to each specific project. This requirement would in- agency with, the largest amount of work on the theory crease by many times our administrative costs, which that such a sponsor is a good customer and must be must be covered by overhead amounts received, and retained? Should it be with all agencies to preclude would be a serious drain, on the time and energy of unfair treatment of one as compared with another? our research personnel who should presumably be de- A privilege granted to one sponsor in one project, voting all their efforts to the objective of the research if allowed to spread to other projects or other sponsors, project and not to determining the total cost of the will soon be a privilege lost to all projects and to all project. sponsors. It will become an intolerable financial burI might add that these analyses take some time den on an institution and will be reflected in a general to complete, and their completion before the begin- lowering of the quality of operations of the instituning of a project would be a necessity if those of us tion. who are concerned about financial matters were to These considerations indicate the desirability of ushave any peace of mind. We can all imagine what ing a standard rate which covers all costs of an overwould happen if the beginning of a research project head character. Such a standard rate precludes overwere held up to any significant extent because the head-cost analyses for individual projects and provides parties involved had not yet decided upon the over- a fair share of overhead costs in each project. The use head applicable to that particular project. of a standard rate takes the question of reimbursement From the point of view of a sponsor of research for overhead costs out of the sphere of argument and work, the preparation of overhead-cost analyses and reduces it to a place of minor significance in a negotiathe concluding of an agreement as to the amount of tion of a research contract, which is its proper place. overhead properly reimbursable will postpone the be- During the war period, which ended five years ago, ginning of a project and will call for a certain amount some rough calculations were made as to the amount of administrative effort on his part. If a sponsor has of overhead costs applicable to research projects caronly one or at most a few projects, the cost involved ried on at universities, and standard overhead rates may not be significant, but there still is a delay in were used in many wartime contracts. In 1947 the obtaining the research results wanted. use of a standard rate was given further impetus by But, let us look at the Federal Government as a the issuance of a set of principles entitled Explanation sponsor of research to be carried on by industrial en- of Principles for Determination of Cost Contracts with terprises or by educational institutions. This sponsor Educational Institutions. has many agencies, all with certain objectives to ac- These principles were approved for use by the milicomplish on what always seems to them as a limited tary services and were acceptable to the universities. budget. Moreover, these agencies are sufficiently inde- They were not when adopted and they are not now pendent so that the operations of one may differ mark- a perfect set of principles fully agreeable to either the edly from the operations of another. Does it not seem military services or to the universities, but they offer self-evident, then, that there should be a consistent a way to determine a standard rate of overhead for treatment of overhead costs by all the agencies of one research projects which can be used for all research sponsor? On the other hand, should not the universi- sponsored by the Federal Government. With the use ties consciously avoid preferential treatment of any of these principles to determine a standard rate, it is one agency of the Federal Government? now a matter of ordinary contract procedure to inI realize that the operating personnel of a single clude a proper allowance for overhead costs in research agency of the Government may wish to reduce the contracts with the military services. payments made by the agency to a minimum for each It is hoped that this practice of using a standard rate project in order to accomplish as much as possible of overhead, which has been found successful in the with its limited budget. And in trying to accomplish administration of many contracts for research, will this minimum expenditure per project, an attempt be adopted by agencies of the Federal Government is likely to be made to reduce the payments for over- which do not now adhere to that practice. It may also head costs to as low a figure as possible. Parentheti- be adopted, in a proper manner, by nongovernmental cally, I should like to add at this point that no one organizations for sponsored research projects.

63 Overhead In conclusion, I should like to state that- 3) when the amount of this allowance can be com1) when full recognition is given to the existence of puted in a simple but accurate manner by the use of a 1) when full recognition is given to the existence of overhead costs on sponsored research projects, standard overhead rate adjusted whenever necessary overhead costs on for sponsored research projectsesfor special circumstances2) when an allowance for overhead costs is consid- then the problems of overhead in sponsored research ered a normal item in the budgeted cost of a sponsored will be considerably lessened, if not substantially elimresearch project, and inated. DISCUSSION OF DR. PIERPONT'S PAPER DR. FURNAS: I don't know whether I can add any- stance, that in the computing of the overhead costs thing in particular to Mr. Pierpont's excellent paper, of the universities, any consideration is ever given to except perhaps to state my experience from another the necessary existence of a department of English. point of view. Nevertheless, the matter of having people with adeI feel that the gripes, some valid and some other- quate cultural background and ability to express themwise, were very much too modest, and I believe that it selves is very much an indirect part of a whole research is really a quite serious situation, and his explanation program. As a result, the universities always come out of the seriousness was also much too modest. on the short end of the horn as far as covering the I think I can speak objectively, because my own full costs of the running of their institution is conorganization, while it is owned by a university, is a cerned. Merely paying for the added increment of takseparate, though wholly-owned, subsidiary of a uni- ing on a few more research people does not pay for versity. It does not operate on the university type of the full costs. Who, then, covers these complete costs? contract; it operates on the industrial type of contract, The person who covers them in some instances is the and it has to be completely self-sustaining financially. poor student paying the fees. That means he does not In other words, it has to pay its own living, in that get as much return on his educational expenditure it has to recover all its costs. as he has a right to expect. Sometimes the intention Is is my opinion that the governmental agencies, of the man who has supplied an endowment is not and I am including all, seriously short-changed the followed. In other cases, those costs are paid by those universities in the matter of overhead, and I think it who are presumably paying for the strictly educational is not in the national interest to do so. activities of the university, sometimes the state legisIn the matter of reimbursement of overhead to the lature itself. universities, the Navy uses the so-called "blue book," The results are either a trend towards financial bankthe Air Force and the Army use the "university type" ruptcy or a trend towards educational bankruptcy. of contract, and these do not, I am convinced, allow for I think I can point out some examples, the first complete and full costs. Of course, there is always the being the case of the medical colleges referred to by matter of the bargaining position, and I do not know Dr. Pierpont. In Harper's Magazine-I believe it is the whether that kind of bargaining is particularly dif- June, 1950, issue-there is an article which asks the ferent from any other type of bargaining, except that question: "Can we afford our medical schools," or you are dealing with a nonprofit institution, which something like that. It goes into a discussion on the doesn't have the bargaining position of saying, "In plight of the financing of the medical schools of the order for us to exist, we must have a profit." country. It makes quite a strong point of the topic Apparently, the basis of the bargaining, as nearly which was brought up, that the more research they as I can tell, is that the increment of overhead which do, the poorer they are, simply because of the way is taken into the discussion is that which is associated the overhead is computed, since the reimbursement with the added increment of cost to the university for they get does not by any means cover the complete undertaking this work. I have sat in on some of these cost. All of you who are interested in the general strucarguments, and I have heard a number of them second ture of the overhead as financed on research by the hand, and I have never seen any philosophy entering Federal Government should be interested in reading in the bargaining, which indicates that the full costs that article. of doing research will be recognized if you take into In suggesting the possibility of educational bankaccount these things which are supposed to consti- ruptcy, one speaks with a great deal of trepidation, tute the heart of the university, namely a community for such remarks tend to make enemies. One of the of scholars. country's leading institutions, according to its annual If I may be more specific, I don't believe, for in- statement last year, had approximately seventy-five per

Overhead 64 cent of its budget involved in research, largely on con- of research projects to pay the full cost, whether or tracts. That means that there are certain people who not the money comes from a grant or a research conare presumably hired for academic pursuits that hap- tract. pen to be riding two horses, and one can't ride two MR. TOUR: I would like to compliment Dr. Pierhorses very satisfactorily. pont. It is certainly pleasant to hear that universities I have heard certain rumors to the effect in the last are beginning to appreciate that overhead is a real few years that there have been certain repercussions factor in the cost of industrial research. in the academic aspects of this institution, simply be- I would like to comment on several items, however. cause the best people-and they are the ones involved Dr. Pierpont mentioned that according to various in research-do not have the time to spend with the ways of figuring overhead, the result might be forty per students who come there to get an education. Perhaps cent or thirty per cent or even as low as fifteen per this is not as dangerous as it appears, but it is a trend cent. He included such things as vacation and ill-time which is more or less prevalent in all universities. I in that overhead. I am surprised at the at the arithmetic. am just pointing out that there is a possibility that It is impossible to include holidays, vacations, and this might possibly lead to the bankruptcy of aca- sick leave in an overhead and have so low an overdemic activity as well as toward financial bankruptcy. head as fifteen per cent. Professional and technical The more research you have, the poorer you are be- employees may average more than fifteen per cent of cause the contracting agency does not allow payment a year's time as lost time. of all the costs involved, so that the money has to be A survey was made of our staff as to the available taken from somewhere else. In the case of universi- time per year of our productive professional and techties, the Federal Government is definitely not acting nical employees. The initial deductions were for two in the national interest in its policy of paying less than weeks' vacation, two weeks' ill-time, and eight holidays full overhead. each year. Next came, how many days are spent at I believe that the universities as a group have been technical-society conventions and committee meetings too modest in their attempts to get adequate overhead. each year, how many hours or days are spent each As a national trend the situation is serious. It is not year on reports and papers for committees and techniin the national interest for the Federal agencies to be cal societies. We expect each professional and technical so penny-wise and pound-foolish, and the universities member of our staff to be active in that type of work should try collectively to help them see the light; in his field, and we must allow him the time to do otherwise the next generation will be seriously in- so. A research worker that will not keep up to date in jured educationally. his field is not the man we want to keep on our payDR. SPENCER: The remark made by Dr. Furnas roll. Making these additional deductions from gross needs emphasis. Overhead is a part of the cost of con- yearly time and plotting the available time as the orducting research and it must be paid. If the sponsor dinate and annual salaries as the abscissa gave almost does not pay the full overhead, it is paid from some a straight line. A three-thousand-dollar-a-year man has other source. It may come from student fees, endow- available something of the order of eighteen hundred ment income, state taxes, or anonymous donors. work hours per year. A fifteen-thousand-dollar-a-year The Government employs industry to conduct a man has only fourteen hundred hours a year available large amount of research and development. It pays the for productive work. The higher the salary, the lower full cost, which includes all direct charges plus the the number of work hours available. full overhead. In addition, it usually pays a small prof- In the ordinary cost-accounting system based on a it. Moreover, industry's overhead is higher than any forty-hour week and fifty-two weeks a year, there are university's overhead. It is difficult to see why the two thousand eighty hours a year, and annual salaries Government should expect the universities to conduct are divided by this figure to arrive at a direct labor research for less than its costs. cost per hour. For example, a man receiving a $6,240.00 The difficulty is, undoubtedly, associated with a per-year salary is figured at a direct labor cost of $3.00 university tradition. For years, universities have sought per hour. However, he only has 1700 hours of time and accepted'grants-in-aid.' With such a grant it is available for productive work. If the missing 380 hours understood that the university must add to the grant are charged to overhead, this item of lost time becomes some of its own funds. Under such a procedure it a 22.3 per cent overhead on direct time. In the case would be possible for universities to reach a position of a $12,000.00-a-year man with only 1500 hours of where they receive only grants-in-aid and research productive time available, this overhead item of lost contracts that do not pay full overhead. They then time becomes 38.6 per cent on direct time. This time might find it impossible to proceed with any of the is not lost because there is no pay work for the man research projects, because the additional funds re- to do but is lost because of the nature of his work. The quired for each project could not be found..$6,240.00 man must be charged as $3.67 for direct Universities ought to encourage or require sponsors labor, and the $12,000.00 man as $8.00 for direct labor,

65 Overhead and this type of lost time must be included in over- leave, and what else, amount to 15.1 per cent of the head on direct labor, or else the overhead rates must payroll, just those items of payroll cost. start somewhere above 33 per cent. The point that I tried to make and that we want to How can this item be included in the fifteen per be sure we recognize is that we should not have any cent overhead that Dr. Pierpont spoke about? Already preconceived idea of what a percentage rate ought to we are over fifteen per cent in the hole, and we haven't be. We have several kinds of operations here. In some started to pay rent, light, heat, insurance, and a host cases some items are charged "direct" and in some of other items. cases the same item may be charged "indirect." It is I would like to point out another item of overhead very easy, of course, to set up a project by itself and that seems to be missed by a number of our universities charge almost everything conceivable as a direct cost and institutions. That terrible little item is called de- of the project, and probably no questions will be preciation and amortization. asked or answered. There is a tendency on the part of the Federal de- We have some Government auditors here in the partments to say that if you got your building for audience who will be able to substantiate my ideas. nothing, and if you got your equipment for next to They know the books better than I do, I believe, in nothing, you have no right to charge the Government some cases. The point that, I think, we want to be with depreciation and amortization. Some day the sure we recognize is that a rate of one hundred and institutions should get to the point of living on their fifty per cent or two hundred and fifty per cent or own incomes. Some day buildings and equipment will fifteen or fifty per cent, or whatever it is, can be obhave to be replaced. Depreciation and amortization tained. Every institution may have a different way of buildings and equipment should be based upon fair to get at its rate of overhead. appraised values and not on first costs. Depreciation The point I want to make beyond that is that the and amortization costs based on such appraised values overhead rate is not a particularly significant thing in should be included in overhead. determining who should get a contract for a research Other items of overhead that were mentioned by Dr. project. We have been told-and it has been used in Pierpont are Federal old-age benefits, social security the negotiations in setting a price-: "Well, we will go taxes, workmen's compensation, and so forth. In the some place else." State of New York these items amount to five per cent If you want to come to the University of Michigan of gross payroll. Productive payroll is not gross pay- or Ohio State or the University of California or Mellon roll. Productive payroll is only that portion of the pay- Institute or some place else with a research project, roll that covers the productive work hours of the pro- it is because there is somebody there whom you want ductive workers. Office workers, cost accountants, time to do the work. That is really the number one requireclerks, telephone operators, and all the people who ment. When you get off that particular point, then argue with the Government auditors about how costs you begin to worry about getting a low price, and you are kept, janitors, charwomen, and messengers are non- may get what you pay for. productive workers on the payroll. The five per cent So that is the thing that, I believe, we ought to be of gross payroll for direct payroll taxes, when con- sure we think about on the overhead rate. It is someverted into a cost on direct or productive labor pay- thing that is a matter of reimbursement and comes roll, becomes close to fifteen per cent overhead. after the project is pretty well set up. Where is a fifteen per cent or even a forty per cent I agree with Mr. Tour, fifteen per cent would not total overhead a possibility? cover the down time. We charge both vacations and I enjoyed what Dr. Pierpont has said about over- sick-leave as direct costs to the project. It seems to me head as a factor in sponsored research, but I am sorry the best procedure is to charge everything you can he did not start with one hundred and fifty per cent charge to the direct cost of the contract, and you won't overhead and say that perhaps it could be gotten have so much to argue about. down to one hundred per cent, rather than starting DR. WATERMAN: I feel that perhaps somebody with forty and getting it down to fifteen per cent. ought to speak for the Government. Perhaps I had DR. PIERPONT: You would get down below fifteen better do it, because I am not particularly competent per cent, if you charge everything "direct." Fifteen on the subject of overhead, nor have I very much to has no validity, ten has no validity. It might take five do with contract negotiations, but I would like to hundred per cent if you had the choice of what is make a few observations from the standpoint of one direct and what is indirect. Take the point on vacation who has taken a very active interest in the effect on pay, for example. I will agree wholeheartedly with the the actual work that is going on in these discussions comments on sick-time and vacation time and so forth. about overhead. As a matter of fact, the so-called fringe benefits that As we see it, the effect that should be considered the University of Michigan has for its labor force, in- is the effect on the institution and the effect on tlhe cluding vacation, group insurance, retirement, sick- general program of research throughout the country.

Overhead 66 Another reason I am glad to make these remarks research standing of the institution. If this were done is because it gives me a chance to talk on the other by institutions generally, it would affect higher educaside of the question of overhead. tion very fundamentally and, I believe, adversely. I have been arguing with so many people that over- I have no formula to apply here, but only a few head is justifiable, that I welcome a chance to talk observations to make. It seems to me, in the interest against some of the remarks that were made earlier, of making sponsored research a success, the best move about what the Government is trying to do in the case is to approach the matter in a cooperative way, bearing of overhead. There are two extremes, it seems to me. in mind the interest of the institution and the interOne is the case where the supporting agency is in- est of the agency. terested in supporting a field of research that is an old DR. KLOPSTEG: I am glad that Dr. Waterman made subject with the institution, one in which it is very his observations, because I have several that fall in the much interested and in which it has unusually com- same category. petent men. That is why the supporting agency goes I fully agree with everything that Dr. Pierpont says there. regarding what we might call development or indusIn this case the university is just as much interested trial research done by educational institutions, namely, in fostering the research and encouraging it as is the the type of work in which the agency makes the apsupporting agency. It becomes much more of a co- proach, just as it would in case of a commercial conoperative venture, and I think that is helpful. sulting laboratory. This is in the hardware category It means that in this case a rate of support may be that Dr. Waterman mentioned. Without question, set, as often done by O N R, that permits additional the sponsoring agency should pay the whole bill, inpeople to be secured for research, while the permanent cluding all the applicable overhead. staff of the institution is still paid by the institution. Now, on the other point that I want to make I I think this is an excellent arrangement. speak not as a representative of a particular instituIn an academic institution, the permanent officers tion, but as one who has been observing institutions in a sense are the board of directors, with regard to in their efforts to do research, especially institutions educational policy, and it does not seem right in the like Northwestern, that are operated out of private enlong run, that these should be supported in any major dowments and student fees and do not have a legisway by an outside agency. Those who are afraid of lature to hand out the great sums of money each Government pressure should welcome this arrange- biennium. We rather feel that the state institutions ment. In fundamental research, then, it makes a great are in a much better situation that way. deal of sense to divide the cost between the institution In private institutions, the amount of money that and the supporting agency. is available to support the scholarly research that the Nor should one forget that research is a real asset faculty may want to do is definitely limited. Now, supto the institution-it gives prestige, it will give favor- pose a Government agency has funds which may be able publicity, and it has a money value. All these devoted to the building up of this backlog of knowlreasons prompt me to say that in fundamental research edge that we have been hearing about the last few already established at the institution, the overhead ex- days, and that, I believe, was first emphasized in the pense may justifiably be shared. report by Dr. Bush, Science, the Endless Frontier. Now take the other extreme, the case of a truly Moreover, suppose an institution is interested in addacademic institution having a staff member who has ing to that backlog of knowledge, and its own funds a flare for a certain type of development which is out- are rather scarce, and an agency which has ample side the normal activities of the institution. An out- funds comes along, and a person in the institution side agency knows this and wants to get the man to eager to do a piece of research sees the opportunity develop a piece of hardware, and it approaches the in- of getting funds from that source. Now, the question stitution. This is not a normal venture for the institu- that I want to raise is: tion. In this case, it seems to me, the supporting If the university can strengthen its basic research agency, since it needs the work done and wants it activity by accepting such funds, whether or not there done there, ought to pay the entire cost. This is quite is overhead, should it do so? Or, should it reject the reasonable, since the problem is one not normally funds, because the supporting agency is not going to undertaken by the educational institution. pay the forty or fifty or whatever the per cent of overAnother observation: if all work sponsored in a head is? nonprofit institution were to be supported entirely In my estimation, there are really two categories of by outside agencies, with all salaries and the entire research of which we must take cognizance. The one, overhead paid by supporting agencies, then, I believe, the actual sponsored research, where the Government the institution would tend to take on a business rather seeks to get tangible, demonstrable value out of what than an educational character. In time this might it buys; the other, where the agency has funds for the interfere very seriously with the educational and support of basic research, and where the university

67 Overhead can strengthen its basic research program by accepting versity is not asking for complete support of an extrasuch funds. neous project but for the supplementary contribution If the answer in the second case, that is in the case to an integrated program. (I would, to be sure, differof supported basic research, is: "No, we can't accept entiate this type of undertaking from a development those funds unless the agency pays all the applicable project, which, I personally believe, should be done overhead," then I raise the question: Should the uni- by a university only in cases of emergency.) This prinversity take the same position toward the granting ciple will be actuated differently in different types of of such funds by private individuals? universities: the poor ones and the rich ones, those DR. SEEGER: Having a full-time relationship with publicly supported and those privately. Each case will the Government and a part-time one with a university, have to be decided on its own merits; uniformity seems I can perhaps see something on both sides of the ques- to be undesirable-if not impracticable. tion. DR. MORGEN: I would like to agree with Dr. KlopDr. Pierpont's comprehensive presentation of a uni- steg and Dr. Waterman. However, there is one point versity's point of view should undoubtedly have been which, I think, we ought to make clear. In our unifollowed by equally factual arguments on the part of versity budgets we assume that we are going to do a a representative from Government. I am certain that certain amount of research. We assume that a certain he would have begun in the same vein as Dr. Pierpont, staff member is interested in research and that he is namely, "We are tired of having all these different going to do it. Therefore, we have included in the rates of overhead with so many types of universities overhead of the university, from our regular appropriaand industries, with private institutions, and with state tions, wherever they are, provision for that research. institutions. We desire a uniform rate." In either case, Then, if we could get a grant or research aid for the same question persists, "How does one determine graduate students or other assistance for that man, a single rate for even one university, or for all univer- we could in that particular case take the aid without sities, for even a single Government agency or for all receiving the full overhead. That amount of overhead Government agencies?" has been calculated in the original institutional Take, for instance, the cited eight per cent rate. Ap- budget. parently, that is customary, but certain people object Each institution, then, will have to determine from to it because of its very uniformity. its own budget how much of that type of aid it can A VOICE: It is too low. take before it begins to exceed the calculated overhead DR. SEEGER: All right, it is regarded by some as uni- of the university. I don't think we can give a specific formly too low. I am familiar with a university which answer to the question for all universities. Each unihas such a uniform rate of its own for all contractors versity is going to have to determine that for itself, on and accepted by all Government agencies, apparently the particular budget that it has. because it is uniformly low. The underlying reason is DR. WHITE: I would like to make two observations. that the university regards itself as a partner in these I suppose most of us have dealings with the faculty enterprises and not as an institution selling something and find it rather difficult at times to convince them of to the Government. Government support is regarded the necessity of overhead. Also, we talk about Governby this university more as a federal grant-in-aid than ment, and my experience is that Government is of the as a purchase of information. In such a case, overhead opinion of some one individual that has been vested becomes of minor importance, being merely an ad- with authority. It is an individual, rather than a comministrative procedure for rendering account of a bined viewpoint. public trust. One's viewpoint, therefore, determines Then I find that different institutions are treated in largely what is to be included in overhead for Govern- different ways. We cannot help but raise our eyebrows ment research contracts. It is amazing what one finds when two or three months ago considerable pressure sometimes included in it on a quid pro quo basis. I was brought to bear on the University to take a conhave heard of demands for partial support of such tract from a Government unit at eight per cent, and items as, not just parking lots, but dormitories and the following Sunday a gentleman came through from stadiums. the Pacific Coast who had a contract for a similar type On the other hand, starting from the premises that of work, only about ten times greater in amount, a university is essentially an educational institution which, we were led to understand, carried an overhead and that higher education is intimately linked with of about forty per cent. basic research, one arrives at the following university attitude: "We have a research program which our attitude: "We have a research program which our Instances of this kind make us question the fairness faculty and students cannot prosecute as they desire. with which overhead is administered. We invite the Government to share in this program, May I now ask Dr. Pierpont to close the session with which we believe is of specific interest not only to our- further discussion? selves but also to the Government." In short, the uni- DR. PIERPONT: I will take but just a few minutes.

Overhead 68 There are two or three things I would like to comment that it is financed. I think that sometimes we lay an on. awful lot to the Government, which is nothing but I tried to indicate that I think this overhead prob- a way for us to escape some of our responsibility. It lem is considered too serious a problem and that we is our responsibility to decide who is to teach school, get too much wrapped up in it. On the other hand, how many hours they should teach, how many hours I don't want to minimize the problem either. they should spend on research time. I think we should all recognize that the university We don't expect at our place that the Government research business- is somewhat different from what it overhead rate on every contract will be the same rate. was fifteen years ago. Certainly it is larger. It certainly I think that that is a mistake. We do not expect our is more diversified, it covers many more fields, it has rate to be like Maryland's or that of any other place many more people in it. in the country. I doubt very much if you will find I would like to say that we do not expect one over- the overhead cost of automobiles in General Motors, head rate for all contracts. That would be a pretty Chrysler, Ford, or Kaiser-Frazer alike. It is an entirely easy way out, but it will never work. Neither do we erroneous conception of overhead to think it is going expect one rate for all universities. We believe that to come out the same no matter where you are. would be a grave error. Comment was made that maybe some universities It has been suggested that universities are looking should charge overhead and some should not; that the for research and should reflect this in overhead charges. ones who are well able to pay the overhead should I don't think that has anything to do with who pays pay the overhead, and the ones that cannot afford it the overhead, whether Washington people are here should not pay it. The rich will be poor and the poor or whether we are in Washington. I believe that will be rich under this philosophy, I am afraid, and we get off base in our thinking on overhead if we are it won't be very long coming. concerned about what kind of a project it is and who There is one other point I would like to make. Dr. wants it. Waterman pointed out and very well, I think, that I agree that research should often be a cooperative one of the philosophies that is inherent in the Governdeal. We should not have any project on this campus ment approach to this problem is that they want to unless we want it. Nor should we have it unless the be sure that we study the situation very thoroughly Government wants us to have it. before we take a research project, that we are aware The philosophy of taking a contract to support a of its cost, and that we are willing to bear part of its few graduate students and not worrying too much cost. If the Government were to offer full costs on all about the overhead, I agree to. But let us have just research projects, we would be rushing for everything one contract without overhead reimbursement or a we could get, and pretty soon we would be only a low overhead rate, and every Government agency will research organization and forget our twenty thousand quote us the example, chapter and verse, day in and students. day out. I think such a situation is of concern, but I don't Against the advice of Professor White and Professor believe it is a Government problem. That is our probGood, we signed a contract a couple of years ago at lem. I think that we should be the ones to be sure five per cent overhead. We have been plagued many that we teach school properly. Our first and fundatimes since with this rate, but these two gentlemen mental objective, as was said before, is to teach have been very kind not to say to me, "Well, you school. Along with that objective is the objective of should have known better." research. It is up to us to decide how much research We have had that particular project brought up so we can do at our institution and handle it properly. many times that I really wish it would never be re- One of the reasons we have project directors, and newed. We don't want it any more, it is a very small one of the reasons we have administrators of research, project, but every Government agency I can think of is to see to it that from every particular project that has used that project as an example to try and obtain we have here somebody gets some good results. a low overhead rate. Part of the problem of research in universities is The overhead problem is not entirely a problem of that it has increased pretty fast. We haven't control the Government beating us down. We are still inde- of all the problems, either internally or externally. We pendent organizations, we still have people working have gone a long way in the last few years in trying on the job of running the University. to solve the problems involved. The so-called blue The problem of the high-paid faculty members, the book was certainly a good step in the right direction. top-flight faculty members, not teaching school but The arguments we have in these discussions and doing research, is not the Government's problem. Let privately with people in Washington or here in our us not lay that to the Government. It is our business own place, I think, are all to the good in bringing out to find out who is to do the teaching at our place the essential nature of the problem. I think that owing and to see to it that it gets done. It is our job to see to these discussions we shall keep going down the right

69 Overhead road. Nevertheless, we probably will come back a year now, and still have some problems on research overfrom now, five years from now, twenty-five years from head matters. ROUND-TABLE DISCUSSION W. K. PIERPONT, Chairman The Round Table devoted to a discussion of over- beginning in defining the elements of cost but that adhead problems discussed the following specific points: ditional work could be done to the benefit of all par1. What precisely is sponsored research? ties concerned. There was an indication also that it 2. What items of cost should be included in overhead might be possible to establish certain norms for testand is it possible to establish some norms for overhead ing overhead rates of universities, though it was also rates? pointed out in this connection that such norms might 3. The problems involved in the determination of the vary so greatly that they would be useless in practice. cost of a research project should be distinguished from There was a discussion on the influence on the the problems involved in determining the amount which overhead rate of various internal university procedures should be received or paid for research.asing, in the treat4. Universities must recognize the nonmonetary rewards and policies in accounting, in purchasing, in the treatobtained from research work and must relate these rewards ment of retirement and pension costs, and in other to the amounts received for research work. areas of university activity. If this is the case-and it 5. Research work accepted for less than full reimburse- was generally agreed that it was - it is desirable to ment of cost must be recognized as such by universities and separate the problem of determining the amount paid the full implication of such a procedure on the financial for research. position of universities must be considered. The cost of research is determined by using accountIn the discussion on what is sponsored research, ing procedures which are not exact and which introthere were quite a few comments to the effect that re- duce a number of considered judgments, whereas, the search work varies all the way from that done by gradu- amount paid for research involves the problem of deate students under fellowship grants to research work termining a precise payment for a given research projdone by full-time research personnel under contracts ect. By separating these problems, it is possible to rewith outside sponsors for work expressly requested late financial considerations to the specific character by the sponsors. Such wide differences in the kind of of the research work to be done, to the source of funds research work done on university campuses should be sponsoring the research, and to other such matters taken into consideration in any discussion of the over- which might affect the amount paid for research head reimbursement to the university by the sponsor- though they do not necessarily affect the cost of reing agency. It was pointed out that corporation or search. governmental grants which are used in the form of A number of the members of the Round Table fellowship grants for graduate students to carry on pointed out that universities derive considerable nonresearch work possibly should not bear a full cost of monetary rewards, such as the publicity given to the overhead expense, whereas, at the other extreme over- university and to its personnel for research work carhead costs should be fully reimbursed. There was a ried on with funds provided by sponsors and the adfeeling that the term "sponsored research" should be mitted aid to the teaching work of the university from expressly defined, at least in connection with a deter- the point of view of the students and faculty. Recognizmination of the amount of overhead costs on spon- ing these nonmonetary re-wards, it may be possible to sored research. relate them to the amount of the reimbursement to the Following some pertinent discussion there was an university for some of the costs incurred in the reexpressed desire on the part of several members of the search work. In any case, however, the university which Round Table for a detailed study of what items of accepts research work and receives less than full recost should be included in overhead and what items imbursement for cost must recognize that the costs of cost should be considered direct labor or direct ma- are still inherent in the operation and must be covered terials and supplies costs. It was pointed out that the by some other source or sources of funds. recent work of the Armed Services and the universities There was a request by several members of the in preparing the Principles for the Determination of Round Table for the inclusion of a session on overCost under Government Research and Development Research and Development head and various aspects of the overhead problem on Contracts with Educational Institutions was a good the agenda of next year's conference.

Fifth Session John I. Mattill, presiding Secretary, Engineering College Research Council NEW GOVERNMENT SERVICES TO RESEARCH THE NATIONAL SCIENCE FOUNDATION AND RESEARCH by THOMAS J. KILLIAN Science Director, Research Division Office of Naval Research SINCE WORLD WAR II the United States Government assumed new responsibilities for the promotion of basic has rendered many tangible services to research. It has research and the development of scientific talent. recognized that basic research is the foundation of The immediate history of this official recognition of the science and technology vital to our national wel- basic scientific research as a national resource began fare and security. in 1944. On November 18 of that year President RooseIn particular, three outstanding steps have been velt addressed an historic letter to Dr. Vannevar Bush taken: (1) About four years ago, the Office of Naval asking for a plan in which the successful research exResearch was established, which since its inception perience developed by the Office of Scientific Research enjoyed the full-hearted cooperation of the Army and and Development could be used after the war to imthe Air Force, both of which have been well aware prove national health and the national standard of of the importance of basic research; (2) the State De- living. In particular, President Roosevelt was conpartment Scientific Office has been established re- cerned about what the Government could do to incently, with plans for Overseas Scientific Staffs; and, crease our future research strength and to discover and (3) the National Science Foundation has finally been develop scientific talent. authorized. Science, the Endless Frontier was the stirring answer Each of these governmental actions, when studied to this request. This report Dr. Bush submitted to separately, seems logical and clear cut. But there are President Truman on July 5, 1945, one month before strong interrelationships. Let us examine some of the the surrender of Japan. Dr. Bush recommended the present and future complex interrelating effects. creation of a National Research Foundation to proWe will begin with the National Science Founda- mote a national policy for scientific research and edution. On May 10 of this year, President Truman signed cation, to support basic research in nonprofit instiPublic Law 507, which authorized a National Science tutions, to develop scientific talent in American youth, Foundation. The stated purposes of this "National and to support long-range research on military matScience Foundation Act of 1950" are: ters. "to promote the progress of science; to advance Then the legislative wheels began to grind. It took the national health, prosperity and welfare; to a five-year gestation period for the recommendations to secure the national defense; and for other pur- bear fruit. poses." The principal recommendations of the Bush ReThus, after more than five years of effort, a new phase port were soon embodied in a bill introduced by Senaof Government activity has begun-science for its own tor Warren Magnuson of Washington. A short time sake. For the first time our Government has acted in later Senator Harvey Kilgore of West Virginia intropositive recognition of the vital importance of science duced an alternative bill. Most of the controversies to our national health, prosperity, and security. It has over the legislation then and later were focused on

71 Government Services to Research three important differences between these two bills. Thus, it was to start from scratch as if the work of First, with respect to organization and administra- Bush and others had never been done. It was no tion, the Bush Report and the Magnuson Bill recom- wonder that all action on a National Science Foundamended that the Foundation be governed by a board tion was tabled in the summer of 1946, blocking furof nine members, none of whom should be full-time ther action before a new session of Congress. government employees, but who should be recruited Nevertheless, an important positive step in the imfrom private life, chiefly, though not exclusively, em- plementation of a national policy recognizing the siginent scientists. The Kilgore Bill, on the other hand, nificance of science to national security was the action provided that the Foundation should be headed by of President Truman in signing Public Law 588 on an administrator appointed by the President and aided August 1, 1946. This law established the Office of by an advisory board without direct power. The ad- Naval Research. Let us look at the introduction to visory board was to be larger than that proposed by Public Law 588: the Magnuson Bill. It was to be composed equally of "To establish an Office of Naval Research in the Depersons from Government agencies and of others out- partment of the Navy; to plan, foster, and encourage side Government. scientific research in recognition of its paramount imporSecond, there was a marked difference in the pro- tance as related to the maintenance of future naval power, posed patent policies. The Magnuson Bill followed and the preservation of national security; to provide within the OSRD patent policy (and incidentally that of the the Department of the Navy a single office, which, by contract and otherwise, shall be able to obtain, coordinate, Navy), which left it open to private interests to patent and make available to all bureaus and activities of the results of work supported in whole or in part by Fed- Department of the Navy, world-wide scientific information eral funds, unless there were specific contract restric- and the necessary services for conducting specialized and tions to the contrary. The Kilgore Bill went to the imaginative research; to establish a Naval Research Adother extreme. It included a blanket prohibition of pat- visory Committee consisting of persons preeminent in the enting results growing out of Government-supported fields of science and research, to consult with and advise research. the Chief of such office in matters pertaining to research." And third, the Magnuson Bill did not include any All the aforementioned controversial issues were specific provision for the social sciences. Although the avoided. O N R is under the Secretary of the Navy. original Kilgore Bill did not provide for the social The social sciences are not directly mentioned. No sciences either, a revised form introduced in October, change in patent policy is indicated. Furthermore, ex1945, did include the social sciences. cept for the Naval Research Advisory Committee, the In a message to Congress on September 5, 1945, elements of the organization were already in existence. President Truman broadly endorsed the policies of the The Naval Research Laboratory and the Special DeKilgore Bill. He strongly recommended that Congress vices Center were active field organizations, while the enact legislation establishing a Science Foundation and Planning Division of the Office of Research and Inthat an administrator and not a board have primary ventions had already launched its university contract responsibility. He further recommended that the social program. sciences be included. The situation late in 1946 was this: A compromise Then began a series of events which some people bill had passed the Senate but died in a pigeonhole in feel could happen only in Washington. The Kilgore the House. Nonetheless, an agency unique in GovernBill was referred to the Committee on Military Affairs ment had been established and was at work, the Office while the Magnuson Bill was referred to the Com- of Naval Research. merce Committee. However, joint hearings were ar- A new note was being added. Late in 1946 the ranged under the auspices of the Subcommittee on President appointed Dr. John R. Steelman as chairWar Mobilization of the Committee on Military Af- man of the President's Scientific Research Board and fairs, of which Senator Kilgore was chairman. directed him to make a thorough survey of all research In the meantime the division between certain and development activities, both in and out of Governscientific groups began to widen. The number of bills ment, and to make recommendations for insuring introduced in the Senate and House began to increase, that the scientific personnel and training and research and they varied greatly. For instance, Senator Willis facilities of the country are most effectively used in of Indiana introduced a bill establishing a National the national interest. This was a Herculean task. The Science Foundation consisting of fifty members ap- board consists of the Secretary of Agriculture, the Secpointed by the President from nominees of the retary of Commerce, the Secretary of the Interior, the National Academy of Sciences. The Foundation was to Secretary of the Navy, the Secretary of War, as well set up its own constitution and define its scope of as the heads of the Federal Loan Agency, the Federal activity subject to the approval of Congress. The final Security Agency, the Federal Works Agency, the Fedclause asked for the appropriation of $100,000 for the eral Communications Commission, the National Advimaking of an "initial report and recommendations." sory Committee on Aeronautics, the Atomic Energy

Government Services to Research 72 Commission, the Tennessee Valley Authority, and the Department in national scientific policy and activities Veterans Administration. With the aid of a Board of and their interrelationships with foreign policy, and Alternates, a report consisting of five volumes was is- (b) appropriate organization and staffing required to sued on August 27, 1947. carry out this responsibility. The conclusions parallel those of the Bush Report. Dr. L. V. Berkner was appointed special consultant Specifically, it recommended that Congress be urged on October 4, 1949, and immediately went to work to establish at its next session a National Science with his characteristic energy and enthusiasm. There Foundation within the Executive Office of the Presi- were the traditional high-level steering and advisory dent and that it be authorized to spend $50,000,000 in committees. The Steering Committee consisted of ten support of basic research during the first year of its Assistant Secretaries of State with the Under-Secretary existence and increasing amounts thereafter, rising to as Chairman. The Advisory Committee was headed by an annual expenditure of at least $250,000,000 in 1957! Roger Adams and included Vannevar Bush, I. I. Rabi, It recommended that no restrictions should be placed Alexander Wetmore, Robert E. Wilson, Alfred N. upon the fields of inquiry eligible for support. Richards, and Detlev W. Bronk. And there were the In the spring of 1947 Congress did pass a bill estab- usual low-level, but hard working groups. lishing a National Science Foundation. It provided Again the result is a report, in this case a rather for 24 part-time members appointed by the President long one (170 pages), entitled, "Science and Foreign with the consent of the Senate. The members, in turn, Relations," which was submitted April 28, 1950. In were to appoint their own Director who would be re- addition to general recommendations that the State sponsible to them and not to the President. Expressing Department become aware of the scientific implicahis deep regret for being forced to take such a course, tions of foreign policy and of the international charthe President vetoed the bill on August 6, 1947, be- acter of science and technology, there are two recomcause it provided "a marked departure from sound mendations of great significance to science in this principles" of administration. Many students of gov- country. ernment are in agreement with the validity of the The first recommended the establishment in the DePresident's decision. partment of State of a Science Office headed by a ScienThus, the issue remained dormant during 1947, tific Advisory appointed as Special Assistant to the 1948, and 1949. In the Bulletin of Atomic Scientists of Under-Secretary of State and supported by a small staff February, 1950, Dr. L. A. DuBridge wrote: of three or four scientists. Dr. Herman A. Spoehr has "After four years of waiting and working there is still been appointed to this office. no Science Foundation. The second recommendation provided for the es"This failure of the National Science Foundation legis- tablishment of Science Staffs at selected United States lation to pass might easily lead one to suppose either that diplomatic missions abroad. These would be of two this proposal was not of very great importance to national categories, differing in size and geographical coverage. welfare, or else that it had stirred up an extremely active Scientific staffs at a few key posts would act as centers Scientific staffs at a few key posts would act as centers opposition. Yet neither of these suppositions is true. The proposed Foundation is of vital importance to the nation, for large geographic areas, not limited to the country and it has not encountered major opposition. It apparently to which the staff is assigned. Initially, staffs in this suffers solely from its own inertia. No individual or group category would be established as follows: London, in Congress has taken the responsibility of pushing this (Western Europe), Johannesburg (South Africa), Rio particular piece of legislation over the many obstacles de Janeiro (South America), and Canberra (Australwhich stand in the way of its passage and approval. asia). "Time is now running out. Unless the present bill before The second category would involve very small staffs the House of Representatives is passed at the next session in individual countries. Initially the posts might be: of the 81st Congress, the chances of ever having a Science Paris, Rome, Berne, Stockholm, Ottawa, Lima, Oslo, Foundation may be small indeed. Copenhagen, The Hague, and Brussels. "The case for a Science Foundation has never been more Copenhagen, The Hague, and Brussels. adequately stated than in the orignal Bush Report, Science, The report was approved unanimously by the Adthe Endless Frontier. The arguments set forth in that re- visory Committee and the Steering Committee. The port are as sound today as when first presented. In fact, Secretary of State directed that the recommendations four years of experience have strengthened some of the be implemented. This is now being done. Thus, the most essential arguments for an independent Foundation." second and third significant steps have been taken In our rapid survey of science and Government, let in transforming positive scientific policy to reality. us go back to 1948, when we find the Department of Let us now look at the "National Science FoundaState being systematically studied by the Hoover Com- tion Act of 1950." It provides for a National Science mission. A special "task force" recommended that a Board and a Director. The Board of 24 members is scientist of national repute be asked by the Department appointed by the President with the advice and conto serve as a temporary consultant to analyze and sub- sent of the Senate. The Director is also appointed by mit recommendations on (a) the role of the State the President with the advice and consent of the Sen

73 Government Services to Research ate. The Board may make recommendations with re- hydrography and oceanography; underwater acoustics; spect to the appointment of the Director, and the Di- electromagnetic wave propagation; studies of the merector is not to be appointed until the Board has had chanical, electrical, and magnetic properties of matter an opportunity to make such recommendations. as they may bear on materials used in ships and airAt present four divisions are provided for: (a) craft; nuclear power for ship propulsion; underwater Medical Research, (b) Mathematical, Physical, and explosives; hydrodynamics; cavitation studies; all Engineering Sciences, (c) Biological Sciences, and (4) types of background noise, such as thermal, magnetic, Scientific Personnel and Education. However, other electromagnetic, and acoustic; mathematics essential divisions may be established as the Board deems neces- for the improvement of fire control, weather forecastsary. Thus, the door is open for the social sciences. The ing, operational analysis, and computers; safety and controversial problem of patents is disposed of by a health-hazard studies with special reference to nuclear generalization calling for the protection of the public radiation; human resources, personnel selection, aptiinterest and the equities of the individual or organi- tude studies, and training; human engineering; studies zation concerned. Foreign activities of the National of man in relation to his environment, with particular Science Foundation are subject to the approval of the reference to arctic, tropic, desert, and other extreme Secretary of State. The Science Office of the Depart- conditions; studies of the normal man, etc., etc. A ment of State will, of course, provide the necessary careful analysis of projects in these and other fields channels. The act authorized $500,000 to be appropri- of Navy interest will indicate that many should not ated the first year and sums not to exceed $15,000,000 be considered for transfer to the Foundation. each later fiscal year. A third screening of the program may be for projThis third step is very important. Not only have ects which bring the activities of outstanding scientists nearly all the objections to previous forms of the bill into the research and development program of the been removed, but numerous improvements have been Navy. O N R's basic research program has made availadded. This is particularly true with regard to security able to the Navy and the Department of Defense, on a provisions and research related to national defense. broad scope, the advice and counsel of many of the The Foundation is not required to carry on research best scientists of the country. The university program related to national defense, although it may initiate has built up a vast network of experts in most of the such research. Research not connected with defense fields of scientific interest to the Navy. The Navy is free from security regulations. should continue to have available an effective and During the past few years, the Office of Naval Re- complete group of consultants in these areas of science. search has given a great deal of consideration to the The effect of a National Science Foundation on problems of its relationships with a National Science O N R personnel has also been considered. We may be Foundation. The establishment of the National Science requested to loan, or to transfer permanently, some of Foundation will undoubtedly have effects on the plan- our people to the Foundation. This may include adning, programming, and budgeting of O N R. There ministrative and contract, as well as scientific, personis general agreement among the committees of RD B nel. We are prepared to give all possible assistance in that the National Military Establishment and the this respect. In addition, we must consider possible deNavy, in particular, should be allowed and encouraged mands for an ON R liaison group to represent the to continue the support of basic research at approxi- Navy in the Foundation. mately the present levels. This is in accordance with Let us remember that what I have loosely called the sound policy that every agency with large develop- the ON R program includes many joint projects to ment responsibilities should have a basic research pro- which the Army, Air Force, A E C, and other Governgram in its fields of interest to insure scientific balance. ment agencies not only contribute but which, in some Yet it is obvious that certain projects on our research cases, they administrate. It is a program of nearly 1200 program can and possibly should be transferred to the projects in 200 institutions. It amounts to about Foundation. Precisely which projects cannot be de- $20,000,000 a year and involves about 3000 scientists termined at this time. Each must be considered indi- and 2500 graduate students. Such joint activities truly vidually. Any transfer must be through mutual agree- indicate the excellent practicabilities for joint and coment with the Foundation, the Office of Naval Re- operative programs between O N R and the N S F. search, the university, and the investigator. The State Department program at present involves It is worth mentioning some of our thoughts on only two small offices, but it is growing. To date the this subject. First, using the stipulation that applied N S F has no Board and no money, and is struggling research be undertaken with a definite naval applica- to come into being this year (1950).* However, we are tion in mind, we can screen all projects which can be * Since this talk was given, an appropriation of $225,000 was classified as applied research. These would not be con- alloted and the twenty-four members of the N S F Board have sidered for transfer. Second, there are certain areas been appointed by President Truman. They include some of of the frontiers of science in which the Navy has a the nation's highest-level educators, engineers, scientists, and very vital interest. I may mention a few as examples: industrialists. The members are:

Government Services to Research 74 dealing in futures. O N R looks forward to the healthy o. w. Hyman, Dean of Medical School and Vice President, growth of both. We hope they will profit from our University of Tennessee mistakes as well as our successes. In such a case the Robert F. Loeb, Bard Professor of Medical Services, College future health, welfare, and security of this country are of Physicians and Surgeons, Columbia University assured. Donald H. McLaughlin, President, Homestake Mining Co., San Francisco Sophie D. Aberle, Special Research Director, University of New Frederick A. Middlebush, President, University of Missouri Mexico Edward L. Moreland, of Jackson and Moreland, Consulting EnChester I. Barnard, President, Rockefeller Foundation Robert Percy Barnes, Head, Department of Chemistry, Howard Robert Percy Barnes, Head, Departent of Cheistry, Howard Jos. C. Morris, Head of Physics Department and Vice President, University Tulane University, New Orleans Detlev Wulf Bronk, President, Johns Hopkins University Detlev Wulf Bronk, President, Johns Hopkins University Harold Marston Morse, Professor of Mathematics, Princeton Gerti Theresa Carl, Professor of Biological Chemistry, Wash- University ington University Medical School, St. Louis Andrey A. Potter, Dean of Engineering, Purdue University James Bryant Conant, Harvard University John WV. Davis, President, West Virginia State College James A. Reyniers, Director, Bacteriology Laboratories, Notre Charles Dollard, President, Carnegie Corporation, New York Dame University Lee A. DuBridge, President, California Institute of Technology Elvin C. Stakman, Division of Plant Pathology and Botany, UniEdwin B. Fred, President, University of Wisconsin versity of Minnesota Paul M. Gross, Dean of Graduate School of Arts and Sciences, Charles Edward Wilson, President, General Electric Company Duke University Patrick Henry Yancey, Professor of Biology, Spring Hill College, George D. Humphrey, President, University of Wyoming Spring Hill, Alabama DISCUSSION OF DR. KILLIAN'S PAPER DR. FURNAS: I would like to know what is the chance priations Committee passed an omnibus bill, a bill or the probability at this time of getting out with $471,000 for starting the work of the National enough money to get this started or implemented this Science Foundation. It was not in the Senate appropriayear. Or is that asking too much of the crystal ball? tion bill; however, hearings were held last Friday. DR. KILLIAN: I tried to find out yesterday, but I Larry Halstead was there, I was there and two other could not. You probably are concerned with the arti- people, and the Senate Committee, what there was of cle in the New York Times, that the House Committee it, was very cordial. Leverett Saltonstall was really in felt that the President did not want any new programs there swinging for us, and we came out feeling rather to interfere with the national defense efforts. to itfewhhntn n. encouraged. Instead of saying, "Do you think it will I think that was farthest from the President's mind encouraged. Instead of saying, "Do ou think it will I think that has now been made clear to the Appro- pass?, we made up a pool on who could guess the priations Committee. closest amount of money we would get. DR. WALKER: I believe that I can give you an an- We have a lot of hope for getting the bill before swer as of last Friday. Last Friday the House Appro- Congress. THE FEDERAL HOUSING RESEARCH PROGRAM by RICHARD U. RATCLIFF Director, Housing Research, Housing and Home Finance Agency This conference itself - the fact that we have gathered search efforts in other areas seem plodding by comparihere-illustrates strikingly the extent to which research son. The relative progress of research in housing durhas achieved recognition as an essential factor in prog- ing these years and that in nuclear physics or plastics, ress. During the past decade, research in some fields for example, has made me feel sometimes a bit as Alice has moved swiftly and dramatically, enhancing the rec- must have felt when the Red Queen said to her: ognition of research generally. But this has made re- "It takes all the running you can do to keep in the

75 Government Services to Research same place. If you want to get somewhere else, you Mounting defense needs put a special emphasis on must run at least twice as fast as that." the objectives of lower costs and economic stability. These days, however, I am feeling much less like In this situation, two considerations become central: Alice than I used to. Because now we have a robust First, the conservation of manpower and materials in housing research program, which is going places. It housing construction and operation; second, applicais certainly something new in Government research, as tion of controls over housing credit and construction, I see it billed on this morning's program, and it is based on a sturdy foundation of fact and analysis. The indeed something new under the sun. value of research in our national strivings toward all HISTORY AND OBJECTIVES these objectives is clear. The Division of Housing Research as presently set THE ROLE OF GOVERNMENT IN HOUSING RESEARCH up came into existence late in 1949. Since then, we The need for housing research is apparent enough. have conceived and put into action a program designed But why is a program to meet this need being sponto meet critical research needs of the housing and sored by the Federal Government? The answer can housing finance industries. Research projects well un- be put simply: The current program was created bederway are focused on a wide range of subject matter, cause critical research needs are not being met by the ranging from light-weight aggregate concrete to family housing industry itself, or by other nongovernmental residential mobility. The current international situa- means. The job of Government in this field is detion has given defense-related activities of the Division fined by these unmet needs. As housing research by an added pulse of urgency. industry, business, and other private groups expands, Congress authorized the Division of Housing Re- Federal activities will be modified to avoid overlapsearch in the Housing Act of 1949. Funds were made ping and duplication. available late in the year. Title IV of this Act created This approach to the problem conditioned the moldthe Division as part of the Housing and Home Finance ing of the Title IV legislation. It has become a guidAgency. The Housing Act recognizes clearly that re- ing principle to us in planning our research program. search is essential to the production of better, less ex- It is linked with the conviction that this program pensive housing. A study of the legislative history be- must be as broad as the society it serves. Preoccupation hind the Act reveals also an increasing recognition of with any one segment of the housing and home finance the chronic need for speeding up, expanding, and inte- industries must be avoided. The program must congrating housing research throughout the housing in- sider them all. dustry, and across the Nation. The Division must constantly seek to comprehend The language of the law and its legislative history the housing process, structure, and environment, in reveal further that no narrow definition of housing is their entirety. The thinking behind all major deciadmissable in the administration of a housing research sions in the program must penetrate to the end prodprogram. In the minds of the legislators, the subject of uct of the housing process-to the individual homehousing was not bounded by the four walls of a build- for this reason: Giving maximum benefit to the housing. The scope of our program therefore reflects a ing consumer yields the greatest benefit to all of the broad concept of housing. The substance of housing housing industry; at the same time it achieves the priresearch is imbedded in all categories of knowledge, mary goals of housing research set by Congress. across the full breadth of nature and society. Thus, in addition to sponsoring research projects. Congress defined the goals of Government housing the Division serves as an observation tower, as it were, activities in terms of several national housing objec- from which the housing industry can view the vast tives. These goals give the research program a definite and intricate panorama of its own process. This servdirection and thrust. In general terms, the national ice is especially important in housing, where there housing objective may be summarized as follows: has been no central vantage point or clearing house 1) Improve housing and its community environment. for research. The Division has a unique and challengThis improvement should satisfy both physical stand- ing opportunity. ards of strength and durability, and human values of In addition to direct research activities, the funchealth, safety, comfort, and personal satisfaction. tions of the Housing and Home Finance Agency re2) Reduce the financial burden of shelter to consum- search program are these: To integrate all housing reers. This involves cutting costs of production, opera- search activity through full interchange of information tion, and maintenance, and providing sound, efficient to the end that the impact of research results will be channels for housing finance. cumulative; to stimulate housing research activity by 3) Stabilize the housing and housing finance indus- others, by spot-lighting research needs and providing tries. The goal is a balanced supply of housing and technical guidance, to translate scientific findings into housing funds. This would contribute greatly to the practical applications, to disseminate information on harmonious growth of the national economy as a research developments to all potential users; finally, whole. to encourage the general adoption of proven innova

Government Services to Research 76 tions. All these are part of the process of achieving ted free and unrestricted publication privileges for the greatest possible yield for business and industry the contractor at the expiration of that period. out of every quantum of energy expended on hous- Another matter requiring some attention was the ing research. arrangement for inspection of the research in progress RESEARCH AND THE "11 H F A FAMILY" by representatives of our Agency. No real difficulty Before talking about specific research activities and arose here since the Division of Housing Research is as anxious as the university contractor that he be perhow they work, I want to give you a quick look at the the s Division's structure and its relation to the Housing itted complete freedom in pursuing the research and Home Finance Agency "family." project. However, it is important that our staff keep HoA was created bny Confress in 1947 to com- in close touch with the work, partly to permit integraH HFA was created by Congress in 1947 to comtion with related projects at other institutions, and, bine and integrate most of the permanent housing.. si partly, to be sure that there are no substantial modiactivities of the Federal Government under a single fications in original objectives and scope. administrator. Currently, its major components are The contracts are on a reimbursable basis and proThe contracts are on a reimbursable basis and prothe Home Loan Bank Board, the Federal Housing vide for an allowance for overhead costs. ReimburseAdministration, the Public Housing Administration, and the Office of the Administrator. The Administra- ment is for actual cost, but the Agency endeavors to tor's office includes the Division of Slum Clearance perate without requiring a detailed and comprehento Redevelopment ands the Division of Research, sive audit. In determining overhead costs the most andoth of which were orn in the Housing Act of 1949. recent available audit, regardless of its purpose, is both of which were born in the Housing Act of 1949. usually accepted as sufficient for establishing a proInside the Division of Research, the administrative visional rate applied to the total of salaries and wages. structure comprises four major staff units and five main branches. The staff units are labeled: (1) Statistical In most cases the Naval Cost Inspection Service has been able to furnish the necessary information. Research and Development, (2) Agency Reports and Research and Development, (2) Agency Reports and At the time the contract is negotiated, a provisional Statistics, (3) Research Intelligence, and (4) Publica- tion. The functional branches are (1) Housing Tech- rate is established. Then arrangementsare made for nology, (2) Housing Economics, (3) Housing Fi- the ultimate determination of a fixed rate to apply for the duration of the contract. The fixed rate is innance, (4) Local Housing Regulations, and (5) Uran Studies. tended to compensate for actual cost and follows the ban Studies. general principles outlined in the War DepartmentCONTRACT METHOD Navy Department Explanation of Principles of DeterUnder the Housing Act, the Division is instructed mination of Costs Under Government Research and to use existing facilities for research wherever feasible Development Contracts With Educational Institutions, rather than setting up elaborate machinery of its own. dated August, 1947. The contract method is the obvious device for carry- The range in overhead returns in contracts already ing out this instruction. Our experience with the con- entered into is approximately from 10 to 55 per cent, tract method has been more pleasant than it might with most cases in the vicinity of 40 per cent. These have been otherwise by reason of the familiarity of differences are accounted for largely by the degree most universities with Government contract work. to which items of direct cost can be identified and estiBut in one respect this familiarity has made it more mated so that direct reimbursement can be made. difficult, since most institutions have become familiar Thus, the greater the coverage of direct costs, the less with contracts with the military establishment. They will be the residual overhead charge. Another factor did not initially appreciate, in some cases, the fact influencing the overhead rate is the proportion of the that our program was operating under a different basic contract work to be conducted off-campus, where the law and thus under a different set of limitations. direct costs are more readily identified and where cerOne of the features of our contracts with universi- tain of the overhead costs such as heat and light would ties which resulted in considerable discussion is that not apply. relating to publication. The basic problem was recon- One other minor problem has arisen from our conciling the statutory requirement that research results tracts which are in force. This relates to the purchase must be placed in the public domain with the prac- of equipment by the contractor. We have been asked, tical necessity of reserving for the Government the for example, to authorize the purchase of typewriters right of prior publication for a reasonable length of and filing cabinets. As a matter of policy it is assumed time, and the natural and proper desire of university that normal facilities are to be provided by the conscientists for independence and free right of discussion tractor both in laboratory and office equipment. It is and publication. These differences were resolved to only in the procurement of special equipment or that the satisfaction of all our contractors in a provision which is substantially in excess of the normal needs which reserved the right of prior publication for the of the contractor that purchase may be allowed out of Government during a six-month period, and permit- Government funds. Ownership of this equipment vests

77 Government Services to Research in the Government when reimbursement of its pur- under the direction of Professor Herbert F. Taggart, chase price has been made. in the School of Business Administration. CURRENT RESEARCH ACTIVITIES The second Michigan project is: Fifty-eight research projects have been undertaken Labor Relations in the Building Industry so far by contracts between H H F A and other organi- One of the largest components of cost in house conzations, private and governmental. Contract funds struction is, of course, site labor. This study is exploracommitted to these projects total $1.4 million. These tory in objective and aims at describing current condicontracts have been signed with twenty-one universi- tions in the labor market as they relate to the building ties, eight governmental agencies, one private non- trades and at identifying major trouble spots to which profit organization, and the Academy of Sciences. future research can be directed. The work is being The Bureau of Census, Bureau of Labor Statistics, supervised by Professor William Haber, long recand Forest Products Laboratory are among the gov- ognized as the leading authority in this field. ernmental agencies conducting housing research for The third project here at Michigan is: us. Universities handling HHFA projects include A Survey of Buyer Considerations in Recent House Columbia, Harvard, Illinois, Tuskegee, California, Purchases and Michigan. Other contracts include the Bureau of This project is to be carried on by the Michigan Entomology and Plant Quarantine of the Department Survey Research Center through the use of the interof Agriculture, and the Southwest Research Institute. view techniques which they have developed to such a The first final research report on any of these projects high degree of effectiveness. The purpose is to identify was received in our offices two weeks ago. Scheduled and evaluate the behavior of recent home purchasers completion dates for other projects range up through in choosing among various housing features. The ultiMay, 1954, but the majority are to be finished during mate objective is to provide sounder design criteria 1951. in home planning, particularly in the small house. Some current projects are aimed at economies In addition to sponsoring individual research projthrough improved materials and better use of materi- ects, the Division is seeking by other means to facilials. Studies on the use and properties of light-weight tate sound business and public policy decisions affectaggregate concretes in housebuilding are examples. A ing housing production, financing, distribution, and project of this kind is being done for us now by the providing Government agencies and business organiNational Bureau of Standards. zations with a steady flow of market information for Another group of projects is aimed at savings their use and guidance. Especially in connection with through more efficient structural systems and structural defense activities, the research staff is devoting an incomponents. The Division's work on the development creasing amount of time to gathering this informaof simplified plumbing systems and more economical tion and presenting it most usefully. floor construction are examples of this. All told, thirteen current contracts are being handled More economical procurement of materials and by the Housing Economics Branch of the Division of equipment is another field of concentration. Current Housing Research, thirty-six are assigned to the Housprojects in this area include an intensive study of ing Technology Branch, five to the Housing Finance present channels used to distribute all kinds of build- Branch, and four to the Urban Studies Branch. Exing materials. amples of urban studies projects are a Study of ResiAnd the Division is striving to help the housing in- dential Mobility, i.e., the movement of families inside dustry develop more effective assembly and erection urban communities, and Growth Patterns in Metromanagement practices. One project is an analysis of politan Areas in the United States. the best methods now in use, including the contribu- HOUSING RESEARCH CONTRIBUTION TO DEFENSE tion of factory fabrication. Three current research projects are being conducted Adjusting activities of the Division to the internaThree current research projects are being conducted Ffor H FA right here at the University of Michigan. tional situation is primarily a matter of gaining a maxifor H H mum yield for defense purposes from research directed Aside from the "local angle," it is appropriate for me to specify them here because they give a good indica- toward the long-term housing objectives set by Contion of the scope of subject matter our program covers. gress. The artificial situation created by economic controls and specialized defense needs increases the diffiThe firsv t of Cothese projects iculty of economic analysis, but makes such analysis Development of Cost Accounting Systems for Home even more important. Build ers Insuring the greatest contribution of housing reThis research will provide home builders with ef- search to national defense does not involve a basic fective means for cost control and cost reduction. Part redirection of the program but rather a narrowing of of the research job will be preparing texts or manuals its scope to bring about a sharper focus on problems on cost accounting for house builders. The project is aggravated by necessities of national security. The

Government Services to Research 78 policy of concentrating more intensively upon defense activity in housing research, attracting minds which problems will, in most cases, produce results not only penetrate housing problems from new viewpoints of of immediate value but of broader application to the study and experience. Ultimately, the cultivation of long-term chronic housing ills. a housing research movement in America may be the HOUSING RESEARCH MOVEMENT IN AMERICA most important aspect of our program. [Extemporaneous Remarks:] We hope that the current H H F A housing research Now, how has the Korean situation affected this program will serve as the catalyst for development of Now, how has the Korean situation affected this a real housing research movement in this country. program? It has very clearly affected everything that The yield from each dollar now being spent on re- goes on in Washington, including research. search in housing can be multiplied if such a move- Our basic aim is to find ways of saving materials ment emerges and matures. There is substantial evi- and manpower, of conserving resources-all of which dence that it is now being born. It is the product of a objectives become even more pointed, in a time of growing confidence in the H H F A research program. labor and material shortages, when substitutes have It will be nourished by the recognition of its results. to be used. And so, in housing technology research, in Not only the results themselves must be recognized, particular, we have no great problem in diverting our but the fact that they were produced by research. This activity and focusing it on current as well as anticiis necessary in nurturing any research movement. It pated defense problems. will assist us in multiplying the impetus of the If we get into the materials allocations at some later H H F A program so that housing research activity stage, which I hope we can avoid, then we will neceswill be generated far beyond the limits of our own sarily become involved in research pertinent to alwork. It will lead to creation of a permanent fund of locations problems. DISCUSSION OF MR. RATCLIFF'S PAPER MR. WORTHINGTON: I wonder, Mr. Ratcliff, if you ment agency wishes to publish a more complete recould tell us how your agency's program will tie in port on the same subject, that is perfectly all right. with the industry-sponsored research agency called the There is a way of doing this that the Structural Building Research Advisory Board. Clay Products Institute has used, with which some MR. RATCLIFF: While you are quite right that in- of you may be familiar, i.e., issuing a series of research dustry groups are responsible for its creation, B R A B publications after the material has previously appeared is part of the National Academy of Sciences. in the technical journals. The article carries the auTheir objectives are much the same as ours, in the thor's name and institutional affiliation directly besense of aiming to integrate and stimulate housing re- neath the title, with a legend across the bottom of the search activity. Their field goes beyond housing, cover, "Number blank of a series of reports on the though, into all kinds of construction, whereas ours is progress of research activities in the Structural Clay limited largely to housing problems. Products Industry." This is satisfactory and suitable We have one contract project with them now-we since it shows who did the work and where it was are financing the research and they are carrying it out. done, while certainly providing proper acknowledgIt involves bringing together information on all hous- ment of the support of the work by the sponsoring ing research activities now being conducted by the organization which it needs to obtain support for subGovernment, industry, universities, and others. We sequent appropriations. hope this will develop into a kind of a research refer- MR. RATCLIFF: I don't suppose we want to instience service for the housing industry, carried on by tute a debate on this subject. We have wrestled with our division. We have other joint activities with it a great deal during our operation and discussed B R A B, and we are working very closely with them. it with our contractors at great length. DR. OWENS: I believe it is true that unless the Being fresh from a university campus, I think that scientists in the universities can actually publish the I was able to see both sides of the question, and in results of their investigations under their own names most cases, not in every one, we have come to an any agency, industrial or Government, will have a understanding that the university people can have difficult time getting research done under contract by freedom of discussion of the technical aspects of the the caliber of personnel that is desirable. problem with other scientists and other people in the The first publication should be in an established field. scientific or technical journal. Then if the Govern- We have found, in discussing the problem with

79 Government Services to Research them, that the kind of thing they want to publish is which the university professor wants to publish the not the kind of thing we want to publish. However, results of the same project. arrangements can be made, so that very likely publi- MR. LOUGHRIDGE: It seems to me this is somewhat cation will be simultaneous. No matter in what order analogous to an experience I ran into about a year ago. publication occurs, the problem, when you think of I would first like to say that I agree with Dr. Killian it in general terms, is a more serious one than when on the desirability of getting research of a military you think of it in specific terms. nature published as quickly as possible, taking due We tried to work it out, and we came to that con- cognizance of security restrictions. I ran into a rather clusion. All we ask for is the right if we choose to peculiar and, to me, surprising result about a year elect it-and in many cases we certainly would elect ago. A question came up as to the degree of duplicait-of getting out a publication somewhat in advance tion between some of the Quartermaster's research in of another publication which would deal with the vital the field of foods and that of the Department of Agriparts of the project. culture in the same field. There are many other kinds of publications that will I was asked to look into the matter and see whether develop as the project proceeds, which we will en- this duplication did or did not exist. I, of course, concourage; but I don't think you can settle the point tacted both the Quartermaster and the Department now, and it is one on which we are open-minded and of Agriculture and ran into this rather peculiar atwill negotiate within the limits of the law under which titude on the part of the Department of Agriculture. we operate. They admitted, as did the Quartermaster, that there DR. SPENCER: Mr. Ratcliff, why does this particu- was some duplication, as, of course, you find in most lar Government agency want the prior right of publi- fields. In research you do not worry too much about cation? It would appear that the results of this type of that. However, some research results had been withresearch should be published as quickly as possible. held from the Quartermaster, due to the fact that MR. RATCLIFF: Well, there are a number of in- they had had some unfortunate experiences and critifluencing factors. One is that our law rather specifi- cism coming to the Department of Agriculture from cally places an obligation to do just what you say is industry. The complaints were based upon the claim desirable, namely, to disseminate the results of re- that the Quartermaster published too much and that search as widely as possible. This obligation obviously in the past, when they had freely allowed information involves publication in most cases. to go back and forth between the Department of AgriOur law specifically provides that technical develop- culture and the Quartermaster Corps, the Quarterments, and in fact all the research product, must be master had released in their quarterly publications put in the public domain. Thus, no contractor may some information which industry objected to on acpatent a process or device growing out of contract count of its being in the nature of trade secrets. work. The philosophy is to secure wide distribution This is exactly the reverse of what we usually hear, without restriction. Take the copyright problem. When namely, that the military services keep too much close a university publishes, or even most scientific journals, to their chest and release essentially nothing. I am the material is generally copyrighted. Our lawyers wondering whether this element may not enter here raised the question of whether such copyright was a since, without doubt, you are tied up, and necessarily restriction of the use of the research product and thus quite closely, with some industries. May there not be untenable under our basic law. an objection by industry in releasing some of this There is also a practical problem, our being able to information? demonstrate to Congress that the money they put into MR. RATCLIFF: I don't think there is in our case research produced useful results. Since publications any particular problem with industry. All our probare the form in which research results are generally lems are worked out with industry, with advisory com"packaged," they constitute primary evidence of work mittees, and we usually obtain representation from completed. And as I mentioned a moment ago, the industry. law under which we operate makes us responsible for Furthermore, we have no control over what the dissemination of our research results. Naturally, we contractor publishes, except this one limitation-havdo not choose to dodge that responsibility on the as- ing prior right of publication for the duration of the sumption that publication by the contractor will suf- contract plus six months. Ultimately, the contractor fice. can publish the whole works, or any part of it, withCertainly, it seems to us perfectly possible to let out any limitation whatever, so that in the end we everybody get all the credit they want and to get the cannot control what comes out of the project in any widest possible dissemination of results. Our rather way. mild restrictions do not impede this-involving our Our only interest is in taking a part that is pertinent right to publish in the channels in which we want to -that scientific knowledge be applied to specific probpublish. These often would not be the channels in lems in industry and that information comes out in

Government Services to Research 80 such a fashion that it is going to be put to work. Our publication funds are available, it will be published. whole objective is lost, unless we can effect economies DR. MORGEN: There is one implication, I don't in the industrial processes of house building. know whether it is true or not, in your statement that MR. GREEN: Perhaps I can shed a little light on you reserve the right for publication by the agency. this. Five years ago the President signed an order, and There is a possibility that you would publish without this order created the inter-departmental Publication giving due credit to the persons who conducted the Board. There are six officers of cabinet rank, and I hap- research for the institution. That is one implication pen to be the executive secretary of that board. I can which, I believe, ought to be clarified. say that their policy has been that it is not too impor- MR. RATCLIFF: No, the matter of giving credit tant who has the prior publication, whether it is the has to be very carefully protected. In fact, we prouniversity or the Government, as long as it is reason- tect the scientist in two ways. We agree to give him ably prompt. full credit for the work done on the basis of the pubBut we are concerned with the degree of dissemina- lication; but, at his option, so that if we should by tion. I think we will agree that sometimes publication chance twist the results or express them in such a in university bulletins does not reach as wide an audi- fashion that he is not willing to be associated with the ence of taxpayers, who actually paid for that research, publication, he may decide not to be recognized. as might be done if a Government agency, which has A VOICE: What right have you to twist his finda direct channel, develops at least a supplementary ings? operation. MR. RATCLIFF: Well, it might be done unintenI am saying for the Publication Board, which actu- tionally. ally has overall authority on this subject, that we be- DR. MORGEN: It seems to me that this raises a very lieve it should be worked out on a case-by-case basis, important point which is relevant to this discussion. that priority of publication is not significant, but There is a distinction between publishing results in rather the degree of dissemination. the social sciences and publishing results in the physDR. LOMBARD: Having heard from the Army and ical sciences (these remarks are directed to Dr. RatNavy, I think the Air Force should also be heard on cliff). It is quite possible, isn't it, in social science rethis point. Research is not worth anything if you do search that (1) the conclusions may be based on qualinot publish it, if you are hiding your light under a tative information, and (2) the conclusions may have bushel. policy implications. Further, almost reverse policy Our idea is to get as much out and disseminate it implications can be drawn from the same set of reas widely and as quickly as possible. Established jour- sults by two different people. Consequently, the Govnals usually have wider distribution than Govern- ernment agency publishing the research results may ment agency journals. publish conclusions of a different sort from those As far as the point of getting credit is concerned, which an individual might publish. Since the first conit is reasonable to require that any article describing clusions of a research project that reach the press are research done under the auspices of the Housing Ad- the ones that get accepted, there is some problem of ministration should give due credit to the sponsoring publishing in this area because of the real difference agency. The publication of an article in a recognized in interpretation given to the same data by two difjournal, giving credit to a sponsoring agency, has ferent people. weight in requesting additional appropriations as has DR. RATCLIFF: I agree substantially with what you the publication of an article in the agency's own said; but it has not been an important consideration journal. in our thinking about the contract deal. We very careIn many research contracts, a report is all that is fully protected the contractor, so that he can publish physically submitted to satisfy contractual require- any interpretation he wants to. ments. It is practicable and desirable to include a pro- However, I don't think that is too important a point vision that the publication of an article disseminat- in the kind of things we are interested in, because we ing the results of the investigation in a recognized are more interested in techniques and processes (even technical journal will satisfy the contractual require- in the social sciences) and in ways of effecting economents for a final report. mies, and much less interested in basic policy, in ideoDR. KILLIAN: I want to bring up some points. logical approaches to housing problems. When research is published in a professional journal, I would agree with you, but I say it has not been it is brought before the peers of the man who pub- too important in this connection. lished it. Now, perhaps it would not be accepted for DR. SEEGER: I would like to make two observations publication, in the case of the A I E T or the IRE. that may be of general interest. Certain Government The latter is a board that goes over this material, and organizations have to be exceedingly careful about all they may decide that it is too trivial to be published. projects under their cognizance owing to possible reThat is not true of a governmental publication. If percussions from industry in the event of premature

81 Government Services to Research conclusions. In this case, the administration is bound with several on this-will not accept an article that to be quite stringent. On the other hand, certain lab- contains substantially the same information as that oratories prefer to make their preliminary observa- which has been published previously in any source tions available at once in informal laboratory reports, readily available. I do not believe that such precedence which are of immediate interest to workers in the field is necessary in Government publications. (one or two months' time requisite), and then to sub- That, I think, is the really important point. We want mit the material for formal publication in national to be sure that research results get into appropriate journals, which reach a large public (three to twelve technical journals, and the only way I know of that months' time requisite). will assure that is to have them appear there first; DR. OWENS: I want to dispel any illusion that or, if some journals will agree to a simultaneous rethe universities do not want wide distribution of re- lease, so that the university research men are not kept sults. They most certainly do. from publishing in appropriate journals, there is no Dr. Ratcliff and the universities are in entire agree- desire to keep the Government from publishing simulment that the results should have the widest possible taneously or at a later date because we all want wide distribution. dissemination of the results. I agree with Mr. Lombard that proper acknowledg- DR. RATCLIFF: Our program obviously cannot sucment of an agency's support is good evidence of ceed unless we have the cooperation of university the worth of the agency's activities. I believe to have scientists. We are depending on them, so that for seen a publication of the H H F A which contained our survival, if for nothing else, we have to work out in the text an acknowledgment to the persons and a deal which is acceptable. institution at which the work was done, but this was These things are always matters of negotiation, trythe only indication that the H H F A did not itself con- ing to arrive at an arrangement that is mutually satisduct the research. A far more prominent acknowledg- factory. But it involves some concessions on everyment of the authorship of the work would leave suffi- body's part. cient evidence of the sponsorship and support of I suppose that I am influenced by the fact that H H F A as a basis for future appropriations. in all the negotiations we have had in the past nine I believe that all university representatives will agree months, we did not lose one single university contract with me that there are two things that are important: by reason of our publication clause. That is not to say One is that you can still obtain good technical men we didn't get a little resistance here and there, but and scientists for work on sponsored research in uni- somehow we persuaded the university scientist the facts are versities, even though they are not getting rich out was not as bad as it looked At any rate, the facts are of it, because part of what they desire is the prestige that we did not lose any of them, and our contracts are and the commendation, if you wish, of their col- with leading universities in the country, from Harvard, leagues in their field. This comes through publication in the East, to the University of California, in the West. of research results under their own names. So I cannot believe that it is as bad as it sounds. The second important thing on which I disagree However, we are open-minded and we will make such completely with Dr. Ratcliff concerns his statement modifications as are necessary to carry out our obthat the university scientists can publish whatever jective and at the same time protect academic freethey wish afterward. After material has once been dom, in which we are just as much interested as are published, most technical journals-and I've checked the universities. - III RESEARCH FOR SMALL INDUSTRIES by JOHN C. GREEN Director, Office of Technical Services U. S. Department of Commerce "Assistance to Small Business" is a popular phrase values at the polls involved in espousing the cause. in Washington because of a genuine interest in the However, it is not at the seat of Government alone that preservation of this vital segment of our economy and the phrase rings in the air. There are many commerbecause of an equally genuine appreciation of the cial and trade groups representing a diversity of in

Government Services to Research 82 dustries or a segment of a single industry which are thing, the large firm is usually working on types of equally vocal and vigorous in their support of the research where enormous investments are necessary smaller competitor. simply to tackle the kinds of problems undertaken. In Among the types of "aids" which have been offered fields like acetylene chemistry, it would do little good as panaceas for the ills of smaller firms is "research." for a small company to invest heavily in research bePermit me to strike a discordant note at this conference cause a large concentration of capital and managerial on the merits and importance of research-to indicate talent is needed simply to capitalize on the findings. a belief that the public has been "over-sold" on the Again, small business profits time and again from big benefits to be derived from this term. Certainly it is business research. Thus, the manufacturers of women's understandable that the tremendous scientific achieve- stockings did not have to develop Nylon but it was ments of World War II, the Atomic Energy program, made available for their use; producers of store lightthe new industry, television, and the many other ing fixtures did not have to invent fluorescent lighting, glistening products ably reported by the advertising but that invention created vast new sales opportunities agencies through all information media have built up for them, and many relatively small radio manufaca public impression that the scientist is a magician turers today are successfully using inventions in telecapable of solving all ills and insuring peace and pros- vision resulting from pioneering research performed perity. Perhaps the scientist himself is a bit guilty too; by large corporations. he remembers all too well the relative obscurity of Large-scale industrial research can only be afforded scientific workers in universities and industry prior by large companies and we should recognize that fact. to World War II. And it is true that where responsible This does not mean in any sense that research is the people have spoken out concerning the difficulty of prerogative of the "big." The small, research-conscious predicting results and the probabilities against amaz- firm with its closely-knit management, labor, and sales ing overnight strides, their comments have not been forces has definite advantages in technological compewidely appreciated. tition since it can "turn around" more quickly than Research should be given its proper importance in its larger competitor and get the improved product the business operation. However, it can never substi- to market fast and aggressively. tute for other and equally important factors of indus- To get back to our firm with less than $500,000 a trial success. Capable management, a skilled labor year gross sales and with $10,000 a year to spend on force, a competent sales organization backed by a research. How can they produce new and improved product of popular appeal priced to meet competition products (the end items of industrial research) at a are at least of equal importance. rate commensurate with their larger competitor with If we admit that small firms can fail because of a research operation budgeted at more than our small normal errors of judgment, incompetency, and plain friend's annual gross sales? There are a number of bad luck, we can take a more objective look at the techniques, none perfect, but applied intelligently role of smaller businesses in the economy and methods to the problems of the individual firm, they can narproposed for maintaining their vigor. For example, row the competitive gap. it is easy to argue that small firms face an insuperable Let us analyze some methods of procuring research disadvantage in manufacturing competition because benefits, realizing that all is not necessarily applicathey simply do not do a sufficient volume of business ble to every company. Let us also imagine that our to afford an adequate research staff and facilities. "small business" is reasonably competent in all asYou may have heard the problem of the small busi- pects of the business operation except research. They ness firm and research stated in this way: have just heard of research-perhaps over the radio, Studies have shown that two and one-half per cent in the corner bar, or through their children's comic of sales is a typical figure for an industrial firm's re- books. Naturally they want to buy some. What to do? search budget. That means that a fifty-million-dollar By this intellectual osmosis they have surmounted firm has over a million to spend on research every the first obstacle-they appreciate the potentials of year. But the small concern whose sales would prob- research and want to know how to apply it. This is ably run less than half a million a year-and over 95 truly a significant step since all too many firms take per cent of American businesses fall in this category- an attitude toward research: (a) we don't need it, (b) would have only about $10,000 available. Ten thou- we couldn't possibly afford it, and/or (c) what good is sand dollars does not give much of a permanent re- it anyhow? search staff, nor the initial cost and expense of main- But our firm is alert and wants to take advantage taining a laboratory. of the powerful competitive tool inherent in research. To the extent that the preceding computations are First, they should have a technically trained man on true, however, they overlook a number of important their staff. He will be a part of management able to distinctions between large and small business. For one talk the language of the researcher and to interpret

83 Government Services to Research management's problems. This liaison runs two ways the technical assistance available through a local instisince research accomplishments will have to be effec- tution, he is certainly at liberty to go to the appropritively integrated into shop production. If our man ate persons there and hire their facilities and assistance. has time to keep abreast of the literature, attend scien- It is one type of tangible help to the businessman who tific meetings and keep aware of the latest develop- lacks his own research unit. ments and their possible significance to his company, The nonprofit research foundations are a rapidly continuously reporting them to management, so much expanding type of research support for business. As a the better. matter of fact, the charters of many of these organizaAt present the greatest number of technically trained tions "spell out" service to small business as a basic graduates go into large firms. The reasons are obvious. objective. Unfortunately, lack of understanding of It is the awareness by such concerns of science and their qualifications and abilities has deterred small technology as a competitive factor that provides job firms from making as effective use as their larger brothopportunities. If small manufacturing firms would ers. Here the research foundations themselves carry systematically employ technically trained men, the on an active program of public education designed to problem of placing graduates in industry would be acquaint more and more businessmen with their exforgotten. Certainly jobs in small business with its istence and specific aids. Many of these foundations chances for rapid advancement, exercise of initiative bring together aggregations of talent, equipment, and and early responsibility would be attractive to gradu- experience easily the equal of the research organizaate personnel. tions maintained by the largest firms. The volume of Next, the engineering and research departments of business annually handled by a typical foundation is material suppliers are a source of technical assistance. comparable to that of a medium-sized company in our Many firms maintain sales engineers, backed up by economic structure. research personnel and facilities, to solve problems re- The Government itself, through its diversified agenlating to the use of their material. The supplier is cies, develops a mass of new and vital research inforalways interested in remedying difficulties which re- mation which our technical officer should know of and strict sales or in exploring new uses which would ex- bring to the attention of his firm when appropriate. pand the market for his material. We have found that many small firms which think The advice and guidance of a consulting engineer they need research actually need better access to known or firm can be of great value. Consultants are special- technical information. Their problem has been solved ists who can assist small firms on a part-time basis for and reported somewhere in the literature but they a reasonable fee because they have a number of clients, have lacked the ability to locate the answer. It is in the sum total of whose business provides satisfactory this type of small-business technical assistance, short annual remuneration. They can be most helpful in of research, that the Federal Government, our great ironing out production difficulties, suggesting im- libraries, and the engineering societies perform a proved methods or materials, and in advising on most needed service. fruitful areas for research projects. Again, the National Bureau of Standards maintains The commercial laboratories provide another area a "Research Associate Plan" whereby industry groups of research for hire. These tax-paying enterprises are are able to maintain personnel at the Bureau to parusually small businesses themselves, qualified to give ticipate in related work and take advantage of the technical advice and guidance at reasonable charges. unique facilities and experience available there. They usually carry out testing of the manufacturer's Trade and industry associations are more and more products and some have facilities, equipment, and per- sponsoring "cooperative" research wherein a number sonnel to undertake research of a high order. Our tech- of companies jointly underwrite a project of benefit nical liaison should acquaint himself with the facilities to all. Here there are no arbitrary barriers of size of reasonably adjacent to him and those which are firm; all may contribute and enjoy the results. This specially qualified in the industrial area he represents. is one of the most important techniques whereby a Next, the land-grant colleges, state universities, and small, alert firm may help shape a research program other educational institutions often have highly trained of direct benefit to them at a relatively modest exfaculty personnel and equipment available for industry penditure. help. Our liaison should develop contacts with these It should be noted, however, that "cooperative" rebodies in order to use them where appropriate. There search is usually devoted to overcoming fundamental is a lively argument as to the propriety of this type of problems which hinder the growth of an industry. It research in educational institutions. Such a philosophi- does not encompass the development of a new prodcal argument is beyond the purpose of this paper. It uct. is sufficient to say that if the small businessman wants Product research is directed toward the development

Government Services to Research 84 and commercial introduction of a new item of manu- level, the Department of Commerce and/or suitable facture. Here the businessman is faced with the task Federal research agencies would be invited to take coof creating a market and selling competitively. He operative action. wants patent protection to justify his investment risk A number of state industrial commissions, manuand naturally is uninterested in a cooperative venture facturing organizations, and like groups expressed inwith sharing of results. terest in participating, and we have been moving Today product research, which leads to diversifica- slowly in this direction. tion of industry, the replacement of "slipping" sales A legislative development which will be of interest items with new and better ones, is vital to the small to this group is contained in the Small Business Act manufacturer who aims to continue as an aggressive of 1950. This legislation, S. 3625, is designed to offer competitor. However, he must perform this type of assistance to small business on a variety of fronts. In research in his own plant or hire on a contract basis "Title V" of the bill, section 503, the following lanthe services of a qualified research organization. This guage will be found: latter is common practice, and the ethics of the laboratory personnel in university, private laboratory, re- "The Secretary is hereby authorized to undertake, tory personnel in university, private laboratory, re- through the National Bureau of Standards, other Federal search foundations, or others insures that proper laboratories, nonprofit research foundations and educasecrecy and patent protection will be obtained. tional institutions, or other facilities available to him, Within the Federal Government, the U.S. Depart- engineering and technological research on industrial, ment of Commerce, under Secretary Charles Sawyer, is commercial and related programs of interest to small busithe branch concerned with the fostering of business- ness on nonagricultural commodities and products. No large, medium, and small. In the Department we often such project shall be undertaken unless he finds that it think of ourselves as a "clearing house" of nonconfi- is unlikely that the objective of such project will be dential information to business. This information falls equally well achieved within a reasonable period of time into many categories-statistical, marketing, foreign (1) by private enterprise or (2) by any other research trade, economic, technical, and scientific. In the last- development undertaken or sponsored by the Government or other public authority." named field, my own office is in daily contact, either directly or by letter, with small firms all over the This language is quite general. It would appear that country. As earlier expressed, we find an urgent need the Federal Government participation in research for for education in small business (a) as to the values small business of a financial nature should be exerof research, and (b) how to effectively apply research cised only on projects thoroughly supported by a to their own operations. broad base of business and in which the business firms As one modest step in this direction, we embarked themselves agree to participate, both financially and on an experiment designed to bridge the gap between administratively. The Federal grant, as I see it, should technical problems brought to our attention by small be an incentive to small firms which are not doing business and the many types of research assistance research now to tackle cooperatively some of the basic available if effectively employed. We viewed our role problems which have been holding them back. This would seem as far as the Federal Government could as primarily that of a "catalyst," organizing these as primarily that of a "catalyst organizing these properly go in the field of business research without problems in a form suitable for research, then bring- finding itself in competition with existing private ing together the small businessmen primarily con- firms. cerned and the interested research organization. After Since the type of research to be undertaken would thorough discussion with men in industry, Govern- be cooperative in nature, the development of new ment, and educational institutions, the following prin- products with attendant patent problems would fall ciples for the program emerged: outside of the legislative purpose. Accordingly, I do 1) It should be undertaken at the local level with a mini- not foresee any very difficult administrative problems mum of Federal participation. should the Congress enact the legislation in its pres2) Representative advisory councils of competent and re- ent form. spected men should be enlisted. A typical council I should mention that the international situation might include a financier, a publisher, a small busi- and mobilization planning has brought up for connessman, and a research specialist. sideration an operation modeled on the Office of Pro3) These councils would undertake the task of handling duction Research and Development which was a comresearch proposals, developing suitable programs, in- ponent of the War Production Board. The OPRD cluding financing, and, in general, stimulating research po arct for business in their area. In performing this function, supported research by contract with qualified facilities they would take advantage of all qualified andI inter- on methods and processes to speed war production or ested research facilities, public and private. provide acceptable substitutes. If such an activity is 4) When the problem could not be handled on the local again created, it probably will be within the Depart

85 Government Services to Research ment of Commerce which has the 0 PR D records a normal, peacetime activity designed to assist small and was assigned the responsibility of liquidating that firms which are in difficulties and need incentives to operation when the War Production Board was dis- join together and plan research. The O P R D type solved. of operation is geared to national mobilization; it The research provisions of the Small Business Act seeks information which will strengthen our industrial of 1950, mentioned earlier, and the creation of an capacity for defense. The two might employ similar Office of Production Research and Development should techniques of project planning and administration; not be confused. The pending legislation contemplates however, their basic objectives are distinctly different.

Summary of Addresses and Discussions ERIC A. WALKER Research and Development Board Washington, D.C. I HAVE SAT through four of these conferences, usu- the product in our line? Do we have the manpower ally trying to get somewhere in the back of the room, to carry it through the production stages? Will we have where I can get the feel of the conference rather than the raw materials, can we get the capital, and where the feel of the individual talk, and at the risk of using are the markets, new or old, where we shall put this an analogy which isn't too good, I have a feeling new product?"* that this series of four conferences can be described, if And then an economic evaluation: "What are the you are musically-minded, pretty much as a sym- ratios-and for this we use the ratios of similar comphony, discordant at times, but pleasing in its overall panies in our line-and what is the ratio of sales to effect. investment for companies in our line? What is the The symphony that this conference has reminded return on our investment, ten per cent, twenty per me of most is Dvorak's New World Symphony. Per- cent, or some other per cent? What is the profit on haps most of you know it. It has a theme of six chords sales, eight per cent, six per cent, or minus per cent? which starts in the first movement. You hear it once, And if we are out of line, should we get into line and then you hear it again. Finally, in the fourth and last how should we do it? movement, the theme takes over. The annotators of So the research director makes out a prospectus for this symphony have often called this, "The Call of the a project on research, and in this prospectus he puts Future." himself and his staff on the spot. "What will be the There is in this whole series of conferences, a call of sales per year from the results of this project? What the future. The first time I heard it was in the first investment will it take on the part of my company? conference, in what seemed to be a perfectly casual What profits can we expect to make? How long will discussion on the part of Maurice Holland, when he it take us to get there? What will be our materials said, "The vice president in charge of research is the position?" vice president in charge of the future." From these individual project analyses, we put toThe theme recurred spontaneously a number of gether a composite research picture and a forecast times last year, and this year we heard it again in a of the future of the company: number of different guises, "The vice president in "What projects must we undertake to attain this charge of research is the man who plans the future of position? What per cent return can we expect as a the company," "Not to do military research is to en- result of my research program? What chance have we danger the future of the country," "The director of of achieving technical success? And then, if the direct research should look upwards and not downwards; he returns are low, what is our chance of over-all sucshould look upwards towards the development of his cess?" company." This same theme was taken up by E. Duer Reeves, This note has been struck again and again. As a Executive Vice President, Standard Oil Development matter of fact, Thomas H. Vaughn, Vice President for Corporation. He said many of these things in different Research and Development, Wyandotte Chemicals words. First: "Research, if it is to be undertaken by Corporation, in his talk on Monday morning on cal- industry, must be useful. We must have the supporting culated risk, quickly sounded the theme, risk on re- technical organization to carry through to the end. search is the risk on the future health of the organiza- We must know what is needed, and we must have the tion. The Board decides the areas in which the research courage to take the risks. organization shall operate, but from then on the re- "We should try as research directors to aim for high search director is on his own. usefulness per dollar of cost, in other words, high re"The research director," says Tom Vaughn, "shall turn for dollars put in. However, we should not formake a technical evaluation of every new research * Virtually all the quotations in this summary are paraphrased.project. Can we as a technical organization do it? Is Editor.

87 Summary of Papers get that there are subsidiary results of any research After the military set up these categories of research, program. Sometimes we need to complete a line so that they tell us what the desirable technical objectives our sales organization will be happy; sometimes we are. What do we hope to do in these categories of relook for prestige effects; but in calculating these things, search? And then the three departments must subject let us not mislead ourselves on cost." their programs and budgets for achieving these techni"In chemical industries the cost of a research pro- cal objectives. gram is not just the cost of operating the laboratory. In the Army, the technical services send their proSomeone has to build a pilot plant. Pilot plants cost posals for research and development to a review board money, and pilot plants must be tested, and although for approval. The review board must consider the risk very often we can slough this cost off on the operating of exposure to a loss if the research is not undertaken. department, it is still a real cost." Many of us in the military are concerned about basic "What is the final plant going to cost us in new research. In this area, it is exceedingly difficult to calcapital, and if there is a risk-and there is always a culate our exposure to loss, for the only purpose of risk-what is the equivalent cost of the research proj- basic research is to add to our store of knowledge. Deect? This equivalent cost can be defined as the total velopment, which is the greater part of the military's cost divided by the probability of success." effort, is a careful application of that store of knowlWe have been led to believe, by the figures presented edge. In the military, we do little basic research; it by these two chemical concerns, that the research di- is approximately ten per cent of our total effort. We rectors are exceedingly good; because the probability can only hope that the universities will provide a resof success was always high, sixty or eighty per cent. ervoir of basic research sufficient for the national need. And the desirability of the research project, what is Monday afternoon we talked again about the evaluathat? Is it the chance of return divided by the cost? tion of research. Allen Abrams, Vice President of the If the sum is positive, we should undertake the proj- Marathon Corporation, continued this theme of evaluect. ation, which many of you will remember was started A research director is also charged with making by Olsen of Olin Industries at last year's conference. up a balanced portfolio, just as a banker or trust-in- He said research and development is usually a very vestment officer is. We need some bonds in the port- small percentage of the total outlay of any company. folio, whose return may be low, but sure enough. We "We can measure the products of our engineering need some blue-chip stocks, whose return is good if department, our manufacturing department, our sales they work out our way, and then, if we have any department, but it is exceedingly difficult to measure adventurous blood, we will want a few gold-mine stocks the results of the research department. This is tough -those that will pay off handsomely if only they work. on the research director, and it is tough on the indiAnd finally, we have to provide for this research pro- vidual researcher, because he would like to know gram and for the program of our company, a steady just how much good he is doing. Furthermore, the flow of men, money, and materials. growth of laboratories in dollars and in facilities deDonald H. Loughridge, Senior Scientific Adviser to pends upon the returns of research. Yet in spite of the the Department of the Army, took another definition difficulties of accurately measuring those returns, the of calculated risk. He said that it leads to that action growth of laboratories in industry and in Governwhich is likely to produce a certain exposure to loss. ment, as well as in universities, has been phenomenal It was twenty-four hours before I caught on to this during the past twenty years." and saw how applicable the definition is to military Olsen commented that a questionnaire showed that research. For if you do research, you may expose your- nearly everyone was for research, and that many comself to some degree of loss; but if you don't do re- panies have formal methods of measuring the results search, military loss is inevitable. of research; but it is quite evident that these formal Then Loughridge reviewed the military course of methods are quite arbitrary. Although they produce action leading to the decision as to the categories in a very good base line from which a single research which research must be done-air defense, strategic director in a single company can measure progress from air warfare, land warfare, and so on. I might add as a year to year, these base lines are not applicable to commentary that we have spent a lot of our time on other companies, not even to other companies in the strategic air and air defense operations, and in the same line of endeavor. present war we have not had to use them.* Perhaps "There is only one real measure of the results of we should have stressed land warfare, but who can say research," says Abrams, "one which will stand the what mgthehpndiwhanresearch," says Abrams, "one which will stand the what might have happened if we had not spent our test of history: Will it produce a new industry, new money on air defense?- It might have turned out to I- _ 1 1 — I I'Yr I Rete ik goods, new happiness, and most of all, a better world be what we needed. How does one measure the risk of this sort of decision? ~ than we had fifty years ago?" Rear Admiral W. S. Parsons, of the Weapons Systems * These remarks are as of September 13, 1950.-Editor. Evaluation Group, spoke on operational research

Summary of Papers 88 which began in the military and is rapidly spreading mex costs more per pound than any other enamel for to industry. He quoted Dr. P.M.S. Blackett of England, insulation, yet because we can pack more copper in father of operational research: a given space, and because we can wind it at higher "Many war operations involve considerations with which speeds, we chose Formex instead of enamel. This again scientists are specially trained to compete, and in which is an illustration of the difficulty in calculating the serving officers are in general not trained. This is especially results of more fundamental research." the case with all those aspects of operations into which Then he gave his story on electric blankets, and left probability considerations and the theory of error enters us with the question, "How would you measure the...the scientist can encourage numerical thinking on results of that research?" operational matters, and so can help avoid running the Last of all, Suits talked about Project Cirrus, weather war by gusts of emotion...." modification. It looks today as though they are doing Admiral Parsons gave an historical outline of the an awfully good job on that particular type of regrowth of the Weapons Systems. Evaluation Group, search. Again I would like to quote his closing remarks, sketching how it was first an idea and is now an organi- for they were well chosen. zation in being, and he reiterated that it is very diffi- He said: "I had developed the thesis that measuring cult to estimate the cost of defeat or even the cost of the return from exploratory research is difficult because forestalling an effort on the part of the enemy. of the manifold forms in which the beneficial results "We can calculate the cost of sinking a battleship, appear, and because of the importance and interor we can measure the cost of downing an enemy plane; related contributions of engineering, manufacturing, the method is really very simple. We set up a paper and marketing. I should like to leave no doubt on the problem based on numbers and probabilities, and work one point, however, that although contributions of out the cost of carefully achieving the desired result. many are vital to the practical utilization of a research Then, admitting that this system is a paper system result, it is the scientist himself who produces the vital based on theory only, we look around in our experi- foundation on which the whole structure of research mental laboratories-and for this type of operation rests, namely, the new fact of nature. The scientists our experimental laboratory was World War II-and and the laboratories that have the freedom to follow find out what results our own records and the records their curiosity in search of new facts are among our of the enemy show. By making a comparison of our most important national resources." paper probabilities and paper analysis with the actual Most of you heard the talk by Dr. James P. Adams, facts we have some test of the validity of the method Provost of the University of Michigan, at the dinner before we attempt to use it on new problems." on Monday evening, in which he brought out the fact Admiral Parsons ended on this thought, which I that our geographical frontiers are ended, that we are think is worth repeating: "But in the over-all view, the left now with a desire to achieve the best that can be measure of return from research in this anxious decade accomplished, and that this leads through research will be the degrees to which it increases our national to new knowledge, new materials, new packages of stockpile of flexibility, resourcefulness, and alertness, power. "This," says Professor Adams, "is our new both in the human and the material fields." frontier." Guy Suits, Vice President of the General Electric James Zeder, Director of Engineering and Research Company, took us quickly to the other end of the for Chrysler Corporation, gave us some new slants scale and showed us by object lessons the difficulty in on the problem, the key to which we find in his title, measuring the results of basic research. He started by "Director of Engineering and Research." saying that it is perfectly easy to measure the results He started by saying that in the Chrysler corporaof a piece of research which has as its object the substi- tion there are four types of research activity: (1) prodtution of a die casting for a fabricated part. But look- uct testing, (2) product engineering and developing at research in semi-conductor solids or work in ment, (3) engineering research, and (4) academic resilicons, where they began by trying to replace a car- search. bon molecule by one of silicon-how can one measure Zeder confined his remarks to the third, engineering such research? research. He said: "A research director must be an orSuits continued: "We thought this would be a ganizer and a leader. He must sell but not oversell. good idea, because the silicon bond gives us a higher He must recognize the important facts and not contemperature stability of certain materials, but we never fuse them with toys. He must not hang on too long, thought that it would give us bounceless rubber and and he must'not drop off too early. rubber which would be compatible with pyranol in "From the viewpoint of the company, he must contransformers. sider: (1) how will the company use the products of "At times we do research in the phase of increased this research? (2) have we got the dollars to exploit costs, because there are other benefits to be derived them? (3) is it timely? and (4) can we do it?" from such research. Let us take a look at Formex. For- He said in closing that he thought we should not

89 Summary of Papers spend a given percentage of our budget each year, summary, I could not help but notice the importance but we ought to spend on research when the going is in all these analyses of a pat on the back applied often tough, and go easy when the going is easy. He intro- enough, vigorously enough, and high or low, as the duced one thought that has bothered me a great deal occasion demanded. and which I might bring out here to make sure that Raymond Stevens,. Vice President of Arthur D. LitI understood it. He said that in the early days the tle, asked us to take a new look at the problem of ratio of dollar returns per dollar spent on research was what a research director does. He is not charged, says very high, but in the past few years salaries have been Stevens, with the production of new gadgets. He is rising, the materials and equipment that a research perhaps the vice president in charge of the future, man must use have been going up in cost, and I think but more truly he is charged with the development of he said the easy things that can be done have been the company. done, all of which result in the higher cost of research. It is obvious he should be able to run his shop, to Now, if I heard right, this means the ratio of return determine what research is to be done, how it is to per dollar spent is coming down, and if this figure of be done and who is to do it; but far above this, he $1.35 as called for or quoted in some of our earlier should have a knowledge of economic, human, finanpapers is correct, it isn't far from $1.35 to $1.05 and cial, and Government matters, which will enable him to 95 cents. And when we get below the dollar mark, to meet other officers of the company on their own will industry continue to do research, and why? ground, and to make his wishes heard in the determinaOn Tuesday morning we gathered to talk about what tion of company policy. is required of a research director. This meeting might In successful companies, the man who is responsiwell have taken on the aspect of a whitewash, in which ble for research holds a position more important than several of our prominent research directors, certain is implied by the title "director of research." He helps that they were good research directors, could have de- to manage the company. His training and method of scribed their own characteristics, enhanced by their thought will assist him, for "his industrial corporation own peculiar point of view, and thus reassured them- should be a living thing, recreating itself, adapting selves, and incidentally reassured us that all of us are to changing conditions, growing where growth is desafe in our jobs. sirable, alive and healthy." However, the program committee in its all-seeing The corporation cannot be lively if its functions are wisdom, made sure that this would not happen, be- static. cause it called in a rank outsider, John C. Flanagan, This concern for the future is not the primary conProfessor of Psychology at the University of Pittsburgh, cern of the sales, production, and personnel men. It is to tell us what made a good research director. the job of the company's research officer, who must Flanagan, making no use of a couch, a supine posi- interpret his staff findings honestly and clearly for his tion, or a darkened room, and making no references to lay associates and who must interpret management complex, libido, or prenatal frustrations, proceeded to policy and finance to his staff for guidance and direct describe an objective measurement of the characteris- them toward proper targets. tics of a good and bad research director as judged by He must have breadth, vision, and something of the research directors themselves. urge of a crusader, something of the enthusiasm of the The measurement consisted of an examination of pioneer and promoter, tempered with an awareness critical incidents on the job, in which the observer, of practical possibilities, and with honest and realistic familiar with the subject and the subject's job require- allowance for the limitations of men and money. ments and his objective in the particular incident This is no job for the small or lazy man. being reported, judged why the accomplishment was Albert E. Lombard, Jr., Scientific Assistant, Air good or bad. Force Directorate of Research and Development, conThe test has been used on a group of colonels and tributed his thoughts as to what is needed in a regenerals, and also on scientists, rates P-5 to P-8. The search executive for the Air Force: comparison of a colonel to a P-5 and a general to a 1) He must understand the role of the Government P-8, provoked considerable amusement and interest. and the Air Force in research, as well as the role of The results of these tests do not yet allow us to make industry and of private endowment organizations, and comparisons, except in a preliminary way, but they this is the tightrope he must learn to walk. show that a scientist is expected to treat a new idea 2) He must find the most economic program by more tenderly than a general or a colonel would; and making and using a time table of interlocking developthat usually the scientists are less handicapped by per- ments which will produce no major upheavals. sonnel problems than are the military. However, 3) He must not produce a plane and forget the everyone seems to think that planning, or the ability to armament. He must not produce a reconnaisance sysplan, is an important attribute in both cases. tem and forget the camera. He must understand the If I may be permitted a personal observation in this objectives of the Air Force to provide rapid and effi

Summary of Papers 90 cient transport, to locate and recognize targets, to hit hold for basic research; and basic research is what the the targets, to cause destruction, to operate indepen- universities and the foundations support. dently of weather and darkness, to counter enemy in- The attitude of the Office of Naval Research is that terference, to communicate air-to-ground and air-to- they are helping support a man where he is, doing air, to defend our home territory, and to utilize the what he is doing, because the Department of Defense money and personnel effectively. is interested in basic research. And because that parHe must obviously promote a research program; ticular basic research might be useful to the Departand he must plan the facilities over a rather long time- ment of Defense, they are willing to pay part of the scale, because of the length of time it takes for con- cost. It is important to notice that in the support of struction. basic research and in the support of applied reOne of the most satisfying of the talks was devoted search the objectives are different, and it is reasonable to a subject which we decided a year or two ago we to expect that the cost angle would be treated differought to bring out into the open, namely, this buga- ently. boo of overhead. Wilbur K. Pierpont, Controller of There were four or five conferences in the afternoon. the University of Michigan, pointed out that approxi- I have received the minutes on these conferences and mately one hundred million dollars was being spent would have liked to include them if time had perby the Federal Government in universities for research mitted. I strongly urge the committee in charge of getevery year. This has created a need for new facilities ting out the results of this conference in the form of and has added to the load on existing facilities. proceedings to publish them as part of the minutes. The results of Government research differ slightly They are as important as any of the speeches. from those of other jobs, but the costs of the research The first conference on measuring research returns are just as real as the costs of manufacturing or any- was run in two sessions by Abrams and Brothers. They thing else. These are arbitrarily divided into direct pointed out that one measures the research returns for labor costs, other direct costs such as supplies, com- five reasons: munications, and so forth, and a catch-all known as 1) To be sure that research and development get the indirect costs. credit for the returns of research. I gathered that this The measure of the indirect costs, or overhead, is was desirable in order to get funds for research. determined by what the accountant, comptroller, or 2) Such measurement is necessary in order for the other fiscal officer decides can be juggled and can differ directors of the company to justify research and dein any way you wish to juggle it. This is not the point velopment to the stockholders. of overhead at all. 3) It is necessary in order to get capital for buildIf the Government wants the work done, and done ing new laboratories. at a given place by a given man, then the overhead 4) It is desirable in selling research to companies should not figure in the argument. The only question who do not have research programs. is: Is the requesting agency willing to pay the proper 5) It is desirable because such a measurement is a costs? check which the research and development director Some schools, especially in the field of medicine and can use on his own activities. public health, fool themselves in this manner, and They then went on to discuss the procedures and they will fool themselves into bankruptcy; or else such hidden gains in a research program as not havthey will have to rob the student who pays the fees, ing to expand the plant because the yield can be inor the state which gives the appropriations, to make up creased as a result of a research program. the difference. This particular conference succeeded in listing sixA uniform overhead figure for all universities and teen different kinds of gains which could accrue from all governmental agencies does not appear feasible a research program. and probably is not desirable; but we might agree more The second session, under LeRoy Brothers, attacked generally on what is direct and what is indirect cost. this problem from an entirely different point of view. Doctor C. C. Furnas, Executive Vice President and They began by pointing out that you cannot apply Director of the Cornell Aeronautical Laboratory, in to fundamental or basic research any indices discussed the discussion of this problem, concurred in many of in the meetings, because worker satisfaction and worker these arguments, and added that some schools, by di- morale is probably the biggest reason for undertaking verting their best teachers to research, are facing not that type of research. only economic bankruptcy but academic bankruptcy In applied research, one deals with many things, the as well. dollar return, the saving of manpower, the saving of Doctor Seeger raised some points that I think are casualties, the use of strategic materials, capital rewell worth bringing out here. He pointed out that the quirements, human relations, etc. Any attempt to conditions for applied research, on which this whole find an efficiency figure will be a difficult one and one discussion seems to be based, might not necessarily which has to be done on the long-term basis.

91 Summary of Papers I am going to read in toto another set of minutes Tom Killian, Director of Research for the Office of on what is needed in a research executive, because Naval Research, spoke on the National Science Founit is a charge to the conference. dation, making three points. He described the gestation "It is generally agreed there is no valid concept period of the National Science Foundation, during of an ideal research executive, inasmuch as such an which the O N R and the State Department Scientific individual may be called upon to perform many func- Office have both been born. If the National Science tions, the exact pattern depending upon the specific Foundation is established, O N R will have to decide job and its place in a particular organization. on some meeting ground with it. Is is obvious that "The essential responsibilities seem to be: (1) Ob- the applied research program would not be transferred taining ideas, that is, planning; (2) selling ideas up to the Foundation, and those things which the Navy and down, that is, promoting; and (3) administering considers basic research should not be transferred, nor the resulting program, including supervision, training, should those programs which are used by the Departand so on. ment of Defense as "scientist catchers," to interest good "A research background is regarded as essential for scientists in the military program. such an executive. Most individuals were of the opin- Richard U. Ratcliff, Director of Housing Research ion that agreement could be reached as to the primary of the Housing and Home Finance Agency, said (and elements listed in the research executive. The ability I am liberally paraphrasing his remarks): and desire to handle people was stressed. It has been said that one of the criteria by which "The primary requisite at present is the unavaila- we decide if the Government should do research is bility of a reliable list of items in a usable form for the portion of the population that it benefits directly. evaluating and selecting research executives. The con- The higher the proportion of the population to benecepts must be defined precisely and in as quantitative fit, the more responsibility there is for the Government terms as possible. Records of specific instances of good to support that type of research. and bad performance were recommended as an initial We all use houses. It is therefore comforting to see procedure." a program of research on the physical aspects of housThe Round-Table group recommends that the con- ing getting under way. The problem of publication of ference request Dr. Flanagan to conduct an investiga- the results, I can only say, was left unresolved. tion to ascertain a short list of critical requirements John C. Green, Director, Office of Technical Serfor a research executive, on the basis of information vices, Department of Commerce, pointed out that aid to be solicited from members of the conference. to small industry and business is a catch phrase, almost At another conference, on overhead, five points were as research is; and so the two can be linked together brought up: in aid or research for small industry, but research is (1) What is sponsored research? only one factor in the success of a business. Sales, pro(2) What items should be included in overhead? motion, skill, and so forth, are also factors. (3) How can we distinguish clearly between costs Most small firms operate around the level of one determined by accounting procedures and the amount hundred to five hundred thousand dollars a year, and paid? if we use the usual percentage of gross sales for re(4) If contracts are accepted for less than the total search, we end up with about ten thousand dollars cost, then a university must recognize its monetary per year available for research in many of these small contribution; and industries. All research directors know that ten thous(5) A university also should recognize the nonmon- and dollars per year buys about one-half a senior scienetary rewards of a research program, such as publicity, tific man. patents, contributions to teaching, etc. Many of the small companies depend upon the reThe conference on calculated risk agreed that charts suits of research done by the larger companies. Many are primarily for the use of management. Rising costs small companies want research because other small bring pressures on research directors to find ways of companies have it and seem to be successful, while cutting costs. Very few industrial companies are doing they don't have it and are not successful. Many of long-term research but seem to be concentrating on them will be able to get the benefits of research through the near future. Is this where the Government and the use of one man, or through the use of part of one universities may have to step in to fill a gap? man whom they can get through universities, research Calculated risk in some industries, such as the chemi- organizations, and science foundations. Most small incal industry, seems to be low, but here the cost of dustries need access to the results of research, and pilot plants is high. Is the converse true in other in- therefore it is probable that the Department of Comdustries? merce will take every step to put those who need reThen we came to this morning, which I will sum- search in touch with the research that has been done marize quickly, because most of you were present. and with the people who can do research for them.

UNIVERSITY OF MICHIGAN 3 901 5 03483 42941111111 3 9015 03483 4294