Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 23
2. RESEARCH TRAINING IN TEE BIOMEDICAL
BEHAVIORAL SCIENCES-—A NATIONAL OVEF(t1IEW
Research training in the health-related sciences is
accomplished in the United States through a large and
complex array cuff programs. This chapter will outline the
nation purpose, and diversity of these programs. It wit ~
review the system of student support and discuss briefly the
problem of determining the supply and demand for biomedical
and behavioral research personnel.
TElE RESEARCH TRAINING SYSTEM
Formal research training is essentially a postbaccalaureate
activity in the United States. The settings in which
training occurs include university graduate schools; school s
of medicine, dentistry, veterinary science, public health,
and engineering; and associated laboratories, hospitals,
mental health clinics, counseling centers, social agencies,
and other field areas.
Levels of Training
In terms of their level, training programs may be divided
into two broad classes -- predoctoral and postdoctoral.
Predoctoral programs are aimed principally at individuals
who are s ee king a research doctorate defined as the Ph. D. or
equivalent. Postdoctoral train ding is provided for
individua Is who hole either ~ research doctorate or a
professional doctorate such as 14. D., D.~.S., or D.~.M.
Predoctoral. Predoctoral training is generally
carried on within the context of a specific scientific
discipline, such as anatomy, biochemistry, microbiology,
physiological psychology, ethology, psychopharmacology, or
anthropology. The functional unit of predoctoral training
is the disciplinary department or, in some instances, a
structured multidisciplinary program.
The predoctoral student characteristically progresses
through a series of didactic and laboratory courses; learns
to conununicate in small seminar groups; and, through
continuing interaction with peers, postdoctoral f el lows ~ the
23
OCR for page 24
faculty, and visiting lecturers, gradually acquires the
traditions, values, style, and research methodology of the
particular discipline. This experience may be brought into
sharper f ocus through experience as -a teaching assistant in
an undergraduate course under the supervision of an
experienced faculty member and in the pursuit of an original
research pro ject. Although such a pro Sect may be a wholly
independent endeavor, it is common practice in the
biomedical sciences for the graduate student to conduct one
phase of a large research program in the faculty mentor's
laboratory, where he or she works as a member of a research
group, utilizing all the research facilities in the
laboratory and conferring f requently with mentor and
colleagues. The predoctoral student in the behavioral
sciences less often derives a dissertation topic from the
large research pro ject calf the mentor, but this depends on
the particular subfield of behavioral study. Laboratory-
based studies of behavior resemble the biomedical tradition.
More common ly, the di s s ertation i s undertak en v ery much on
the student's own responsibility after a period of limited
app~entices~p, usually as a research assistant, in
connection with one car more research pro jects that are
organized and executed }:y a senior member of the faculty.
In all cases, somewhat more of the research methodology is
learned didactically rather than on the job, although this
again varies across the spectrum of the behavioral science
suLfields.
Less formal predcctoral research training programs are
available to individuals who are pursuing a professional
doctorate and are taking time out for research training.
These individuals are for the most part undergraduate
students in the health professions schools, who, having
shown interest in the biologic phenomena underlying disease,
have been encouraged to develop their res earch potential.
For medical students, this research activity may be carried
on part-time and concurrently with their regular studies, or
full-time (luring an "off" quarter or during a year's release
from medical training. -
Predoctoral training in the basic biomedical and
behavioral sciences generally requires 4 to 7 years of ful I-
time work and study between receipt of the bachelor's degree
and attainment of the Ph. D. degree. Approximately two-
thirds of Ph. D. recipients in the biomedical and behavioral
sciences were engaged in academic employment in ~ 973. The
remainder in the biomedical pock were evenly divided among
industry, the federal government, and other sectors. The
nonacademically employed behavioral scientists worked in
hospitals, clinics, social welfare agencies, or were self-
employed.
Postdoctoral. The Prague of postdoctoral training,
. .
both at the post-Ph. D. and postprofessional levels, has
perhaps been most widely recognized in the biomedical
sciences. In addition, postdoctoral training has become
24
OCR for page 25
almost a requirement for the physician wishing to pursue a
career in academic medicine. The opportunity for M. D. and
Ph.O. graduates to be trained to-tether during the
postdoctoral period is widely regarded as a valuable aspect
of this experience. Further, postdoctoral education is
regarded as important for training graduates from the
physical sciences who wish to utilize their special
knowledge and capabilities for biomedical and behavioral
research.
Post-Ph.D. training is oriented toward specialized
experience in a research topic rather than a discipline.
Each year an increasing proportion of those who receive the
Ph. D. degree have been going on to postdoctoral training to
sharpen their research skills under outstanding mentors. In
}'ioche~nistry, for example, the percentage of Ph. O.
recipients taking immediate postdoctoral appointments rose
from 36~2 to 58. ~ over the period 1962-67.t A large
percentage of these Ph. D. holders seek such experience in
settings other than those of their graduate school mentors
in order to gain new perspectives and to respond to new
opporttlnities, particularly in interdisciplinary an-d
emerging research areas.2 In 1973, the number of biomedical
and behavioral science post-Ph. O. ' s in the United States
totaled ainaost 6, 000. 3
Research training for individuals holding a professional
doctorate must take into account the difference between
their educational background and that of the the post-Ph. D.
Many college seniors with exceptional] y fine academic
records elect to enter Lexical school instead of Ph. D.
programs. To the extent that the 4 years calf medical school
and subsequent years of residency training were designed to
produce clinicians and not medical scientists, residents who
then el ect to become clinical scientists requi re further
scientific training.
Since a proper mix of biomedical, behavioral, and
clinical scientists is essential to initial aiscovery.and
full application to clinical problems, it has become
necessary to decibel ap new programs in school s of medicine to
instill rigorous scientific discipline in the design of
experiments, proper use of methods, and critical evaluation
of data, as well as to provide a inroad background in the
physical and biological sciences. These programs are
demanding on toth the physician-trainee and the faculty,
because of the trainee' s need to acquire in 2 to 3 years an
extraordinary amount of new scientific knowledge, a variety
of skills, and actual research competence. A highl y trained
medical scientist today must have, in addlition to clinical
training, a grasp of the snore basic areas, such as molecular
structure and function of proteins and nucleic acids, the
ultra-structure of various cell types that make up tissues
and organs of the body, and the underlying behavior
mechanisms calf the whole body. Further, although some post-
M.~. Is may do only clinical investigation, others will
25
OCR for page 26
OCR for page 28
OCR for page 29
OCR for page 30
OCR for page 31
OCR for page 32
OCR for page 33
OCR for page 34
OCR for page 35
Representative terms from entire chapter:
student support
engage in research on basic biomedical problems whose
solutions are essential to clinical progress, and many wall
develop competence to do both clinical investigation and
basic laboratory investigation. In ~ 973 the number of
individuals with professional doctorates who were pursuing
postdoctoral research training totaled more than 5, 000.
Interaction among Basic Biomedical, Behavioral,
and Clinical Scientists
At the present time, there are two groups of scientists in
biomedical research. The first consists of those with an
intensive preparation in one of the physical, biological,
basic medical, or behavioral sciences but with relatively
little knowledge of clinical medicines These scientists are
essential to elucidate the fundamental processes that
overlie living systems. Because of the difference in
perspective and training, however, they do not pose the
types of
flexibility to the institutions and departments involved in
the training.
Fellowshins. These are awards made directly to the
individual, largely in the form of a stipend, from a variety
of sources, such as the fevers government. voluntary health
organizations, foundations, and universities. Fellowships
have been used in support of training for many years as a
means of encouraging excellence and reducing financial
barriers to training. Except for local programs, the fellow
may take the appointment at any host institution with
appropriate facilities and where ~ suitable mentor is
willing to supervise the training. Awards are made to both
predoctorals and postdoctorals, selected in separate
competitions. National fellowship programs permit
utilization of a more uniform set of standards in the
selection of fellows, and the awarder has wider latitude in
the selection of a training site. The award, as in the case
of NIH/ADANBA fellowships, may include a modest
institutional allowance to help Be fray costs of training the
fellow such as tuition and fees, research supplies,
equipment, travel to scientific meetings, and related items.
That allowance differs, however' from funds included in
training grants for the specific purpose of strengthening
the scientific milieu of the department in which the
training is pursued.
Training grants. These grants are awarded by the
federal government to institutions for individual
departments or a consortium of departments for training in a
specific field. In addition to providing trainee stipends,--
these grants enhance the quality of training by providing
funds to departments for salaries, special seminars,
courses, supplies, and equipment. In contrast to
departmental training grants, multidisciplinary grants make
it easier for graduate students to anticipate emerging
scientific fields and to select their dissertation topics
accordingly, with access to appropriate faculty in re, ated
sub ject areas across departmental lines. Programs supported
by these grants may I:e devoted exclusively to either
predoctoral or pcst~c~c~toral training levels, or Moth.
:~:ndividual trainees are se:Lected; by faculty participants on
the basis of credentials and letters of recommendation
similar to those in fellowship programs. The grants are
awarded through national competition, with continuing peer
review to ensure that training i s conducted In departments
of the highest quality.
Res earch ass i stantships ~ Many graduate students and
Fostdoctorals receive support for work performed on research
pro jects. Of ten this work fits their scientific interests
and training requirements anti hence serves as useful
educational experience. In some cases, the student may
conduct an original study, as part of the overall project,
that is suitable for a dissertation and that contributes to
the advance of research supported by the pro ject grant. As
27
a major source of support for research training, however,
this mechanism has some drawbacks. Training for research,
particularly at the predoctoral level, requires more than
on-the- jot experience. When the student's support is tied
to a research grant, there exists the possibility of a
mismatch between educational ob jectives and the pro ject
director' s primary interest in maximizing research output.
In addition, competence in research by an individual project
director does not necessarily parallel competence and
interest in training. Experience with the research grant as
a training instrument, in light of its advantages and risks,
suggests the need for a-flexible mixture of the research
assistantship and other types of support for research
. .
. :~n~ ng.
Teaching assistant~h~n`: Graduate students and
postdoctorals may be supported by their institutions as
teaching assi stants because of their contributions to
teaching programs. These teaching assignments are usually
at the undergraduate leered, and, hence, opportunities for
this type of support tend to be available only in
institutions and departments with large undergraduate
teaching responsibilities. The teaching assistant typically
grades papers, sets up experiments for laboratory sessions,
holds classes in which lectures given by senior faculty
members are discussed and in which texts and other written
material relating to the course are reviewed, and guides
students in laboratory work associated with undergraduate
courses. These duties commonly occupy one-half of the
teaching assistant's time, in return for which a modest
stipend i s provided and tuition fees are remitted. Val uable
experience is acquired in the art of teaching, which for
most Ph. D. hollers is an essential part of their future
bobs.
Private means ~ Many students, particularly at the
predoctoral level, support their training through private
resources, including family aid, private loans, part-time
employment, and assistance from a working spouse. Excessive
reliance on this type of support is undesirable, because
many persons of superior potential cannot afford to
undertake yes earch training.
These different forms of support, properly administered.
can be appropriately related to the various purposes and
stages of the training process . For example, graduate
students may be provided with fellowships or traineeships
for the first 2 years, teaching assistantships for the third
year, and research assistantships for the fourth and
succeeding years until the dissertation is completed and the
degree awarded. This pattern allows the student to prepare
as rapidly as possible for teaching and research. Teaching
is postponed until the student knows enough to be able to do
it well f and We research assistantship is utilized at a
time when the student is relatively free and well enough
informed to make a sound choice of mentor whose research
28
program has an appropriate opening. under its auspices and
support, the student can conduct the research needed for the
dissertation while contributing to the n~entor's research
program.
TEE SYSTEM OF SUPPORT IN THE
BIOMEDICAL AND BEHAVIOR SCIENCES
The role of f ederal support in research training can best be
understood when placed in the context of the overall system
of support in the biomedical and behavioral sciences.
Considering the diverse loci of responsibility for
supporting graduate educat~c)n, it should not be surprising
that the federal government supports only a relatively small
fraction of al ~ biomedical and behavioral science graduate
students. In postdoctoral study, the national interest in
targeted research and the absence of ~ ocal university
responsibility have meant that federal support here has
assumed a larger coverall responsible ity. This section will
first consider graduate and postdoctoral support in the
biomedical and behavioral sciences and then discuss briefly
support for clinical research training. Health services
res earc h, which i s b a si call y a mul ti dis cip l i nary f i e ld f e d
by diverse behavioral, biomedical, and clinical fields,
cannot be adequately defined quantitatively at this time and
thus will be omitted from this discussion.
Graduate Student Support
The federal government in ~ 974 provided the primary source
of support for 29.1 percent of fulI-time graduate students
in the biomedical sciences arid 20.9 percent of behavioral
science students.
This compares to 23. 7 percent of graduate students in all
sciences (see Figure 2. ~ and Appendix II, Table lI.2) .
Other primary source s of support were institution~state
sources and personal resources {including self, family. and
spouse}. In the biomedical sciences, institutional support
comprises 4 2. ~ percent and self -support 2 ~ . 5 percent of
primary sources of support. In the behavioral sciences,
where federal funding Is less prominent, a larger portion
(34. 0 percent) comes from self or fancily and 38. ~ percent
from institutional and state sources.
It is important to note that different systems of
support utilize different mechanisms, depending on the
ob jectives of the sponsor. Federal support for graduate
students, which is intended primarily for research purposes,
is made up of fe1 lowships, traineeships, and research
assistantships. In the biomedical and behavioral sciences,
support by the U. S. I,epartment of Bealth, Education. and
29
ALL SCIENCES
24% federal
Fellow! A.:......
Trainee^.......
[22,2,2h222,..
RA
_~ ~ Other
Bellow/
Oth:4
Types of Support:
F:dbwsJliptrrair~h ip
Research Assistantship (RA)
Teaching Assistantship (TA)
Other
2996 Federal
BIOMEDICAL SCIENCES
76% Nonfederal
RA
Fellow/
Trainee
Self
7t~ Nonfederal
Other
I
LEGEND °~ ~ ~lo~ot~r
Inst''ut~onai Fellow/ Train"
a, ~ Other Trainee
t" i';! Nonted~al
BEHAVIORAL SOtENCES
21% Federal
~.~
I .2 .N
by - FslJowl )`
........ Tra'~e
::: .:::::::: ::. ::: ::::::
Vera ~ LA
FelIc~w/ Other
Trainee
79% Nonfederal
,
~ Trainee //~/~ Other ;/4
RAT
Fc''ow/ ~ Seit
Trades \
-
NOTE: See Tables 1 ' .2 and 11 .3 in Appendix I I for supporting data.
SOURCE: National Science Foundation, Graduate Science Student Support and
Postdoctorals Survey, t974
FIGURE 2.1 Primary Source and Type of Support for Full-time Graduate
Students In the Biomedical and Behavioral Sciences, 1974
30
/
Welfare (DREW), primarily through NIR and ADAMEA, is
concentrated in f ellowships and traineeships (over 70
percent}, in contrast to res earch assistantships ~ In
agencies other than DREW federal support for research
training in all sciences is provided primarily through
research assistantships {see Figure 2. ~ and Appendix IT,
Table II . 3) .
Institutional~state support is composed of a wide
diversity of types of support. Fellowships~traineeships, as
well as research assistantships, are utilized, but by far
the largest mechanism is the teaching assistantship, which
reflects the primary interest of the individual institution
and state government. Self-support, as noted above, is a
significant source of support comprised of one' s own
earnings, loans, and spouse and family support.
In recent years, federal support has dropped off for the
biomedical arid behavioral sciences and indeed for all
sciences. Most recently, front 1972 to 1974, federal support
has declined 17. ~ percent in all sciences, 13.6 percent in
the biomedical sciences, and 18. ~ percent in the behavioral
sciences (see Appendix II, Table IT. 2) ~ The mechanism of
federal support most sharply cut back is the
fellowship~traineeship. NIH support here in the biomedical
sciences declined 13.2 percent from 1972 to 1974 (see Table
2. 1} . Federally supported research assistantships, however,
sea rise to offset partially the severe fellow trainee
cutbacks .
Nonfederal sources rose considerably aur~g this time,
at rates far exceeding overall federal declines. In the
biomedical sciences, institution/state support rose 17.0
percent and self-support rose 19.9 percent (see Table 2. 1) .
These increases account for the recent rise in graduate
enrollments in spite of federal cutbacks. s All types of
instituti on/eta te support-- f ellowshi Strain Reships,
research assistantships, ~d teaching ass~stantships--shared
in the overall in-crease. Self- support also rose
substantially during this period.
Postdoctoral Support
The biomedical sciences have a strong tradition in
postdoctoral study as a necessary step in gaining the
knowledge and skills needed to conduct biomedical research.
Because of the highly specialized nature of the training and
its direct relation to federal agency goals for solving
health research problems, the federal government has assumed
a ma jor role in postdoctoral study in this area. The
federal goverrunent provided approximately 75 percent of the
primary sources of support for biomedical science
postdoctorals in 1974. Of those federally supported, about
40 percent were f ellows~trainees and 60 percent were
TABLE 2. ~ Primary So~c~ and Type of Support for E ull-t~me Graduate
Students in the Biomedical Sciences, 1972 and 1974
~ ~ ~ Ch.~ce,
1972 1974 1972 1974 1972- 1974
Total 25205 26663 100.0 100.0 5.8
Fell/Tr 7541 6692 29.9 25.1 -11.3
RA 4925 5435 19.5 20.4 10.4
TA 6600 7376 26.2 27.7 11.8
Other 6139 7160 24.4 26.9 }6.6
Total Federal 8998 7770 35.? 29.1 -13~6
; ~ . . . . ~ . .,
' - Fell/Tr 5830 4508 23.1 16.9 -22.7
~ .
; RA 2671 2845 10.6 10.7 6.5
TA 108 119 .4 .4 10.2
.' Other 389 298 1.5 1.1 -23.4
NTH 5736 5244 22.8 19. ~ -8.6
. .
Fell/Tr 4317 3746 17.1 14.0 -13.2
RA 1264 1364 5.0 5.1 7.9
TA 75 49 .3 .2 -34.7
Other 80 85 .3 .3 6.3
Total Nonfederal 162.07 18893 64.3 7Q.9 16.6
. .
Fell/Tr 1711 2184 6.8 8.2 27.6
RA 2254 2590 8. g 9.7 . 14.9
TA 6492 7257 25.8 27.2 11.8
Other 5750 68G2 22. a 25.7 19.3
Institu~ion/state 97;8 Il41d 38.7 42.8 3~7. O
. . ..
Fell/Tr 1104 1522 4.4 5.7 37.9
BA 1754 2001 7.0 7.5 14.1
TA 6425 7237 25.7 27.1 12.6
Open 475 654 l.9 2.5 37.7
4785
5736
.
19~0 21~5
19 e 9
CODE: Fell/Tr = Fellowship/TraineeshIp' RA = Research Assistantship;
TA = Teaching Assistantship; Other = Other Types of Support.
NOTE: Data include persons enrolled in only those departments which responded
to all three (1972-74) surveys, and hence do not represent population figures.
SOURCE: National Science Foundation, Graduate Science Student Support and
Pos tdoctorals S urvey ~ 1972- 74.
32
research associates, i. e., employed on research grants or
contracts . ~
The behavioral sciences have not utilized postdoctoral
support to the extent that the biomedical sciences have
because their programs were designed primarily to add to the
pool of basic researchers. For example, 40 percent of 1972
Ph.D. 's in the biomedical sciences entered into postdoctoral
work, compared with 22 percent of behavioral science
Ph. O. ' s. of those behavioral science Ph. D. ' s who sea enter
postdoctoral work, approximately 50 percent were federally
supported. NIB and ADANHA data indicate that they supported
2, 798 biomedical science postdoctorals in 1974 {2, 600 of
which were from NIB) and 278 behav-iaral science
postd`~torals 7 .
Clinical Sciences
Clinical science training is unique compared to biomedical
and behavioral science training in three basic respects:
(~) it deals primarily with M. D. ~ s and other professional
degree recipients; (2) ~t is located almost exclusively at
professional schools; and (3) it takes place almost entirely
at the postprofessional ~ ever. Because medical schools are
professional schools with no undergraduate departments an<]
because professional students are generally trained as
practitioners, there are virtually no teaching or research
functions that prcfessional students serve. Clinical
research training is thus concentrated at the
postprofessional level. Typically, over 95 percent of NIH
fellowship,traineesh~p support in the clinical sciences has
been made at the postdoctoral level.~.According to ~
National Science F-oun<~ation survey, approximately 60 percent
of clinical science postprof~ssxonals in ~ 974 were federally
supported, of which 75 percent were fel lows~trainees and the
remainder research associates. 9
:~BrE:RMI GINO THE: SUPPLY AN: D DEMAt1D (NEE:)) FOR BIOMEDICAL
AND BEBA1JIORAI, RESEARCH PERSC)NNhL
The HRSA Act of 1974 calls for assessment of the need for
personnel to perform research in the biomedical and
behavioral fields. The term "need'' can be given various
interpretations, and the Committee has devoted considerable
discussion to the appropriate definition for purposes of
this study. In a general sense, there is a need to reduce
the costs of the various it Inesses that prevail in our
society today. In a very broad sense then, need could be
interpreted as the manpower requirements that would result
from a policy of investment in biomedical research based on
the social costs of disease. This is a fairly unconstrained
33
approach, since it would require that research expenditures
be somehow tied to a social-cost figure without regard to
budgetary ~ imitations.
The Commi tte e ha s a <30pte ~ a somewhat more di sci p li ned
interpretation c>f the task set f orth in the Act. In the
Conunittee's view, need is interpreted as the manpower
requirements dictated by market demand at the prevailing
salary levels. In cipher words, the task is interpreted as
one of determining the number of positions that are expected
to be available in the next f ew years for biomedical and
behavioral scientists assuming that no significant changes
In - their wage structure will occur. The number of available
positions for these scientists in turn is believed to be
governed by the likely future pattern of enrollments in
higher education and biomedical and behavioral research
expenditures. Most researchers perform some combination of
research and teaching. The teaching component of demand is
thus represented by enrollments in higher education, while
the research component is represented by the amount of
research funds available.
Enrollments are fairly predictable, since they depend on
demographic patterns that can be estimated from known birth
rates. For example ~ correct predictions of increased
college enrollment rates some 20 years later were made at
the time of the post-WorId War I] baby boom.
Conversely. research funds are allocated annually and
are sub ject to the normal variations associated with
economic conditions and political processes, making them a
less predictable component.
On the other side of the market picture is the expected
supply of scientists available to perform biomedical and
behavioral research. Demography also plays a role here, for
while the short-run effect of increased enrollments is to
increase the teaching component of demand, the long-run
effect is to increase the supply of scientists. We thus
view the market for research personnel in these fields as a
dynamic system whose el ements are continually changing in
response to demographic and economic factors. Superimposed
on the demo: graphic cyst es are the variations i n research
emphasis and funding that add tc' the difficulties of trying
to assess the future supply and demand pattenas. It seems
clear at this point that Me system has passed through a
sustained period of rapid growth into one in which the
growth is expected to be more moderate. The supply/demand
balance in future years depends, in addition to the factors
mentioned above, on students' reactions to perceived job
opportunities for dc~ctoral-level scientists. The Committee
feels that continual monitoring of trends in enrollments.
research expenditures, and job opportunities provided by
this study can furnish timely information to guide both
individua l deci s ions and fed era l po li cy . A more det al led
assessment calf manpower needs is presented in the next
chapter.
34
FOOTNOTES
1. National Research Council, The Invisible University:
Postdoctoral Education in the United States,
D.C., National Academy of Sciences, 1969.
Washington.
2. National Research Council, Mobility of Ph.~. ' s:
and A f ter the Doctorate , Washington , D. C., National
of Sciences, 1971.
3. National Research Council,
and Engineers, ~ 973.
Before
Academy
Survey of Doctoral Scientists
4. National Science
Student Support and
D.C., U.S. Government Printing Office, 1974.
Foundation, Graduate Science Education:
Postaoctorals. Fall 1973. Washington,
5. National Science Foundation, Graduate Science Student
support and Postdoctorals Survey, 1972-74.
6. Ibid., 1974.
7. National Research Council, Commission on Human
Resources, Personnel Needs and Training for Biomedical
Behavioral Research. The 1975 Report of the Com:~et:ee
study of National Needs for Biomedical and Behaviora
Research Personnel, Washington. O.C., ~~ - ~
Sciences, 1975~ p. 34.
8. Ibid.
and
on a
__ _
National Academy
of
9. National Science Foundation, Graduate Science Student
Support and Postdoctorals Survey, 1974.
35
