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 37
CHAPTER FOUR
BEHAVIORAL SCIENCES PERSONNEL
The United States leads the world in behavioral research,
and there is now increasing recognition, especially among
policymakers, Hat the solution to many of He worst prob-
lems facing the country are primarily behavioral in charac-
ter. The behavioral sciences yield knowledge to deal win
many of these problems.
Behavior ranges from individual skills (for example, the
lack of self-control implicated in violent behavior) to cogni-
tive mediators of risky behavior (e.g., an individual's per-
ceptions of invulnerability relative to risk-taking behaviors
that affect health) to community-level phenomena (e.g., an
individual's belief in the legitimacy of medical prescrip-
tions derived from beliefs shared with a group). Every area
of major health risk in this country can be informed by
behavioral science research. Moreover, in some areas of
great risk such as child abuse, substance abuse and crime,
and acquired immune deficiency syndrome (AIDS) large
scale behavioral programs are being developed, and ad-
vances in demographic and population studies have en-
hanced our understanding of "longevity".
Since issuing its first report on the National Research
Service Awards CURSE) program in 1975, the National Re-
search Council ARCS has strongly endorsed He continua-
tion of support for research training in the behavioral sci-
ences through the NRSA program. This occurred in
recognition of the inclusion of the Alcohol, Drug Abuse,
and Mental Health Administration (ADAMHA) in the re-
structured research training authority (P.L. 93-348~. The
behavioral sciences also received attention because of their
critical role in exploring the underlying processes of normal
human development and aging; for their role in outlining
the causes, etiology, and treatment of diseases involving
neurological and sensory processing; and for their role in
prevention research. In October 1992 He research compo-
nents of the three former institutes of ADAMHA joined the
37
National Institutes of Heals (NIH) signaling an historic
shift (Goodwin, 1993) in the nation's view of the subtle
interactions between changes in behavior and health and
dysfunction.
Recommended levels of support in the behavioral sci-
ences have been predicated on the belief that NRSA awards
leverage Be production of small numbers of highly skilled
workers whose research is conducted in the national inter-
est. Indeed, follow-up studies have found most former
NRSA awardees actively engaged in research (see Appen-
dix A).
The number of degree recipients having had NRSA sup-
port remains low and, therefore, does not drive the supply
of behavioral science doctorates. Nonetheless, He commit-
tee this year has paid particular attention to the market for
behavioral scientists for reasons related to the health of the
overall behavioral research enterprise. As noted in previ-
ous chapters (2 and 3), the committee has replaced earlier
committees' analyses of supply-and-demand models win
the use of new techniques which involve multistate life table
analysis of future supply and a separate assessment of short-
term indicators of current and past market conditions.
The committee was gratified to learn, for example, that
despite severe funding cuts in He early 1980s, behavioral
scientists continue to find productive employment although
overall Ph.D. production in these areas has slowed some-
what in recent years. Based on this information, and infor-
mation gathered from a variety of other sources, we con-
clude that the demand for behavioral science personnel will
grow slowly as a whole but that He few behavioral scien-
tists receiving NRSA support can expect to enter into pro-
ductive employment upon completion of their doctoral or
postdoctoral studies.
While these analyses provide an important backdrop
against which the committee's judgment about future needs
OCR for page 38
MEETING THE NATION'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
for behavioral research personnel can be assessed, they rep-
resent but one of the several dimensions of "needs." Cogni-
zant of the important advances in behavioral research that
promise to enhance the health and well-being of all citizens
and confident that NRSA awarders will have the opportu-
nity to contribute to the national health effort, the commit-
tee has concluded that the present modest program of NRSA
support in the behavioral sciences should essentially double
in the next few years. However, because of our concern
with current low levels of stipend support, we have adjusted
our goals to permit immediate expansion of training sti-
pends throughout the NRSA program. We believe, how-
ever, that the national call for skilled behavioral science
investigators should be answered swiftly and strongly
through an increase in training support in this area.
ADVANCES IN RESEARCH IN
THE BEHAVIORAL SCIENCES
The knowledge base in the behavioral sciences has
reached the point where effective utilization of its findings
by physicians, clinical psychologists, nurses, and social
workers can have a very significant impact on health-re-
lated problems in our society. Advances in basic research
in behavioral science dovetail well with national needs.
This is a most exciting and intellectually stimulating time
in behavioral science. Work on both basic and applied prob-
lems and work done at many levels of analysis, from
brain-behavior relations to the study of disease processes in
human populations is making great progress. We cite
three examples in more detail to illustrate the intellectual
excitement in the field.
Health and Behavior
Much of the social and behavioral research supported by
the National Institutes of Health represents research on
"health and behavior". Current estimates place NIH sup-
port for research in this area at about 8 percent of total NIH
R&D support (COSSA, 1994~. The "NIH Implementation
Plan for Health and Behavior Research" (NIH, 1993) out-
lines, furthermore, what the Institutes and Centers consider
to be optimal spending levels in this area during the next 5
to 10 years.
Expanded support for research on health and behavior
reflects an increasing recognition by the health research
community that social and psychological factors play a sig-
nif~cant role in the natural history of disease, prevention of
disability and illness, and promotion of recovery. As the
Director of NIH stated in a 1991 report on the same subject
(NIH, 1991):
Our research is teaching us that many common diseases
can be prevented, and others can be postponed or well
38
controlled, simply by malting positive life style changes.
For these reasons, intensifying such research and encour-
aging all Americans to make health-enhancing behav-
iors a part of their daily lives has taken on more and
more importance in our efforts to conquer disease.
The domain of research on health and behavior is broad
and the enormity of its knowledge base is daunting, but as
Adler and Matthews (1994) suggest, many of the concepts
in recent years pertain to three essential questions:
1. Who becomes sick and why?
2. Among the sick, who recovers and why?
3. How can illness be prevented or recovery promoted?
To facilitate the answers to these and other related ques-
tions, the NIH Reauthorization Act of 1993 established the
Office of Behavioral and Social Sciences Research and
called for a report to the U.S. Congress on the extent to
which the Institutes of Health conduct and support research
in the component disciplines. Because the Office is still
being organized within the National Institutes of Health,
Howard Silver and the staff of the Consortium of Social
Science Associations (COSSA) recently prepared a detailed,
although preliminary, summary illustrating NIH funding pri-
orities in the area of health and behavior. For example, the
COSSA summary notes that the National Heart, Lung, and
Blood Institute (NHLBI) supports a variety of activities
primarily through the Behavioral Medicine Branch-on dis-
ease prevention, etiology, diagnosis and treatment of car-
diovascular diseases. The National Institute of Allergy and
Infectious Diseases (NIAID), which is concerned increas-
ingly with the transmission of AIDS, supports only a mod-
est amount of social and behavioral research but has ex-
pressed interest in psychosocial factors affecting medical
treatment of compliance. The National Institute of Envi-
ronmental Health Sciences, to give another example, sup-
ports research on the effects of environmental agents on
human heals and well-being with particular attention to
behavioral and neurological effects of exposure to toxic sub-
stances.
Variables that have been explored by research scientists
over He years can be categorized in any number of ways,
but include research on factors Hat arise from the social
environment which contribute to disease. Such factors in-
clude: stress (Cox and Gonder-Frederick, 1992; Beardsley
and Goldstein, 1993) or social isolation/social connected-
ness(Cohen, 1988;Reynolds and Kaplan, 1990~. Individual
dispositional factors are also Bought to contribute to dis-
ease onset and recovery, such as: hostility/Type A personal-
ity (Siegrist et al, 1990; Mathews et al, 1992~; depression/
exhaustion (Hahn and Pettiti, 1988; Markovitz et al, 1991~;
neuroticism and negative affect (Costa and McCrae, 1987;
Salovey and Birnbaum, 1989~; and optimism/self-esteem
(Schreier and Carver, 1992; Brown and McGill, 1989~.
OCR for page 39
BEHAVIORAL SCIENCES PERSONNEL
Advances in research on health and behavior often
emerge from multidisciplinary studies involving social, be-
havioral and biomedical scientists including clinicians. As
federal funding priorities increasingly emphasize research
in this area, we can anticipate that new research training
opportunities will emerge, perhaps requiring consideration
of new research training arrangements which emphasize
cross-disciplinary connections.
Learning and Memory
Memory is dissociated into processes or systems that are
fundamentally different. For example, amnesic patients
with brain injury or disease exhibit severe inabilities to re-
call and recognize recent events and have difficulty learn-
ing new facts. However, these patients possess some rela-
tively intact learning and memory on tasks such as
manual-dexterity learning, which they perform as well as
do healthy and uninjured people, even though they may
have no conscious memory of having performed the task
before. This evidence-that some kinds of learning can
proceed normally even when the brain structures that medi-
ate conscious remembering are damaged supports the gen-
eral proposition that there are distinct, dissociated types of
memory.
Memory is an active process of seeking and recons~uct-
ing information, not a passive recording and reproducing of
events. Thus, expectations of what things should look like
or the way events should happen influence what people no-
tice and remember. For example, after listening to a story
presented in jumbled order, people still tend to remember it
as being told in proper sequence, following certain widely
accepted scenarios for what constitutes a story. People also
tend to pay little attention to the details of routine situa-
tions. Consequently, people often remember Hat He most
probable things happened even when they did not. This
phenomenon has been demonstrated in He context of eye-
witness court testimony. Memory of an event can be modi-
f~ed or distorted by how questions about the event are posed.
Such experimental findings have important theoretical im-
plications for understanding the formal structure of memory
and they have practical implications for legal proceedings.
Recent progress in the broad field concerned win brain
substrates of learning and memory (behavioral and cogni-
tive neuroscience) has been impressive. The essential cir-
cuits and probable loci of memory traces have been largely
identified and mechanisms are being characterized for sev-
eral invertebrates (e.g., Clark and Schuman 1992; Kandel,
1976~; for learned fear in mammals, including both behav-
ioral and autonomic measures (amygdala) (Hitchcock and
Davis, 1986; McGaugh, 19894; and for classical condition-
ing of discrete behavioral responses (skilled movements) in
mammals, including humans (cerebellum) (Daunt et al.,
39
1993; Thompson, 19864. Progress is rapid in determining
the role of the hippocampus and associated cortical areas in
contextual-spatial and declarative memory (Meuner et al.,
1993; O'Keefe and Nadel, 1978; Squire, 19924.
A particularly important application of this basic research
is in aging and Alzheimer's disease. The procedural
memory system concerned win motor skill learning de-
clines with normal aging, with concomitant loss of neurons
in the cerebellum. Declarative or experiential memory
shows only a modest decline in normal aging but a dramatic
and profound impairment in Alzheimer' s disease. The brain
substrates of this declarative memory system (hippocampus
and associated cortical areas) are particularly vulnerable and
show progressive and profound deterioration in the disease.
Thanks to the well-developed fields of learning and memory
and abilities testing, we now have very sensitive diagnostic
tools to identify the earliest stages of putative Alzheimer's
and other dementias.
Early childhood is a period rich in experience in which
young children are capable of acquiring, registering, and
recalling events and episodes as well as recognizing places
and people. They can also vividly react to impressions,
manifest pain and pleasure, and express love, jealousy, and
other passions. It is Gus astonishing that this period is
usually lost to memory entirely except for a few fragments.
Recent work in developmental-cognitive psychology and
cognitive neuroscience suggests that infantile amnesia is not
due to repression, as Freud hypothesized but rather occurs
because of the length of time necessary to develop He req-
uisite cognitive skills and the concomitant developmental
schedule of He brain substrate of declarative (experiential)
memory. The hippocampal and cortical memory system is
not fully mature until about He age of 3 (Bachevalier, 1991~.
These data raise serious questions concerning the current
rash of therapy- and hypnosis-related instances of patient
accusations of infantile assault, based on presumed recov-
ery of infantile memories (for example, Time Magazine,
November 1993~. Over and above the issue of suggestibil-
ity is the well-documented evidence Hat the original
memory of an experience can be modified substantially long
after He fact by how questions about the experience are
phrased (Loftus and Loftus, 1976~.
Signal Detection and Medical Imaging
Signal-detection theory is an outgrown of behavioral sci-
ence, first noted in the field of hearing. A fundamental
problem in hearing is how to detect the presence of a signal
in noise. Two extreme listener strategies are to respond
positively with great uncertainty, leading to many false posi-
tive responses, and to respond positively only with great
certainty, leading to many false-negative responses. In a
brilliant mathematical and behavioral analysis, Swets,
OCR for page 40
MEETING THE NATION'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
Green, and many other behavioral scientists developed a
comprehensive theory of signal detection that makes it pos-
sible to establish probability levels of these two types of
errors and to manipulate them by establishment of criterion
levels of judgment (see, Green and Swets, 1966~.
The practical applications growing from this basic re-
search have gone beyond sensory problems. Modern elec-
tronic and computer technologies for image enhancement
have wide applications today, ranging from enhancement of
space-probe images to brain imaging to computer-reading
of X-ray films. Computational techniques combined with
various medical imaging procedures are making it possible
to see specific organs and internal body parts volumetri-
cally and in ever-increasing detail. Computer-based imag-
ing techniques include computed axial tomography, nuclear
magnetic resonance imaging, positron emission
tomography, and digital subtraction angiography. Some
techniques, by assigning different degrees of transparency
to different layers of bones and tissue, let the observer not
only see them but look through them and see what lies
within or behind them (Fitzgerald, 1989~.
Signal-detection theory has addressed the diagnostic
power and cost effectiveness of these imaging-techniques,
relative to more conventional and less expensive X-ray tech-
nology (Swets, 1979 and 1988; S wets and Pickett, 1982;
Metz, 19864. As medical imaging technology continues to
develop, the need for studies evaluating the effectiveness of
specific instruments and procedures will increase. This is
so especially in view of the considerable cost of the equip-
ment involved and because diagnostic images produced will
not always be interpreted by image interpreting specialists
but by physicians who view certain types of images only
occasionally. There is a need also for He development and
evaluation of procedures (including computer-aided proce-
dures) to help clinicians, especially those with limited expe-
rience in image interpretation, to extract from medical im-
ages the diagnostically relevant information they contain
(Getty et al., 1988; S wets et al., 1991~.
Another example of current practical application is the
cochlear implant. This application, like many from behav-
ioral science, is interdisciplinary in that it involves neuro-
science, orology, physics, engineering, and behavioral sci-
ence, specifically the psychophysics of hearing and methods
of training.
Signal-detection theory is one aspect of the broader field
of decision theory, a set of well-developed methods to
evaluate decision-making in situations where uncertainty is
inherent. These methods have been applied, with consider-
able success, in a variety of medically related situations and
have proved to be extremely helpful in improving training
procedures that both accelerate training and increase the
overall quality of medical diagnosis (Luce et al., 1989~. A
dramatic illustration of such an effect is given by McNeil et
40
al. (1982), who asked people to imagine that they had lung
cancer and had to choose between two therapies: surgery or
radiation. Each therapy was describedin some detail. Then,
some subjects were presented with the cumulative probabili-
ties of surviving for various lengths of time after each type
of treatment. Other subjects received the same cumulative
probabilities expressed in terms of dying rather than surviv-
ing (e.g., instead of being told that 68 percent of those hav-
ing surgery will have survived after 1 year, they were told
that 32 percent will have died). Framing He statistics in
terms of dying dropped the overall percentage of subjects
choosing radiation therapy over surgery from 44 percent to
18 percent. Of the three groups of subjects used in the
study patients, students, and physicians this effect was
found to be the strongest among physicians.
ASSESSMENT OF TO CURRENT MARKET
FOR BEHAVIORAL SCIENTISTS
The labor market for researchers in the behavioral sci-
ences was robust in the 1980s. The behavioral sciences
work force increased by 35 percent between 1981 and 1991,
climbing from approximately 49,000 to nearly 67,000 (Fig-
ure 4 1~.2 However, over half of this growth was accounted
for by clinical psychologists, who are less involved in re-
search and development (R&D) than are their nonclinical
counterparts (Pion, 19934.3 Although psychologists have
been fully employed and utilizing Heir research skills in
recent years, there has been a shift in employment prospects
from nonclinical to clinical fields.
Accompanying this grown were notable changes in the
composition and employment distribution of these workers.
Among these changes were the growing prominence of
women and members of racial and ethnic minority groups
and the increasing importance of the nonacademic sectors
in providing employment opportunities. The latter change
80000
75000
70000
65000
60000
55000
50000
45000 - .
40000 ~
/
-
-
_'
/
-
-
1981 1983 1985 1987 1989 1991
FIGURE 4-1 U.S. behavioral science employment, 1981-1991.
See Appendix Table F-13.
OCR for page 41
BEHAVIORAL SCIENCES PERSONNEL
reflects in part We rapid employment growth of clinical psy-
chologists, who are primarily employed in nonacademic
positions.
Women have been becoming an increasingly important
component of We labor market for behavioral scientists.
Nearly 40 percent of this work force was female in 1991
compared with 27 percent in 1981 (Figure 4-2~. The fe-
males tend to be concentrated in the younger age groups.
About two-fif~s of the female behavioral scientists were
younger than 40 in 1991 compared win, et moss, one-four~
of He comparable males.4
The U.S. work force has become more racially diverse
over the years and the behavioral science work force also
reflects these changes. In 1991 nearly 8 percent of em-
ployed behavioral science Ph.D.s represented individuals
from a racial minority group (Table 4-1~. In 1979 minori-
ties represented about 4 percent of these Ph.D.s. Most of
He grown occurred for blacks. The progress in ethnic di
40
35
30
~ 25
c 20
L)
1 5
10
o
1 981
1991
FIGURE 4-2 Fraction of the U.S. behavioral science work force
who are women, 1981 and 1991. See Appends Table F-13.
versity is less dramatic. In 1991 Hispanics represented
about 2 percent of the behavioral scientists. In 1979 the
comparable statistic was roughly 1 percent.
The behavioral work force is aging. The median age of
behavioral scientists has increased from almost 40 in 1981
TABI,E 4-1 Racial/Ethnic Composition of Employed Behavioral Science Ph.Dos° 1981 and
1991
,
198 1a
Number Percent
199 lb
. . .
Number Percent
Race
TOTAL
White
Black
Asian/Pacific Islander
Other (Incl. Native Amencan)
Ethnicity
TOTAL
Hispanic
Non-Hispanic
55,821
53,506
1,102
1,013
200
100.0
95.9
2.0
1.8
0.4
54,742100.00
8591.6
53,88398.4
88,340
83,533
2,590
1,924
293
87,686
1,931
85,755
100.0
94.6
2.9
2.2
0.3
100.0
2.2
97.8
~ ., .
aFor those who responded in 1981. Race nonresponse was 222 in 1981 and ethnic nonresponse was 1,301.
bFor those who respondedin 1991. Race nonresponsewas229in 1991 end ethnic uonresponse was 8830
NOTE: Employed behavioral science Ph.D.s are those with a behavioral science Ph.D. in behavioral science
fields, regardless of employment field. Estimates are subject to sampling error. Comparisons between 1991
estimates and those of earlier years should be made with caution due to changes in survey methodology. Prior
to 1991, the SDR collected data by mail methods only. In 1991, the survey had both a mail component and a
telephone follow-up component. In this table, 1991 estimtes are based on "mail-only" data to maintain greater
comparability with earlier years. Totals may not add up to 100 due to rounding.
SOURCE: NRC, Sunrey of Doctorate Recipients. (Biennial)
41
OCR for page 42
MEETING THE NATION'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
45
44
43
42
41
40
39
38
37
36
__ 1 1 ~ i ;
1981 1983 1985 1987 1989 1991
FIGURE 4-3 Median age of the U.S. behavioral science work
force by gender, 1981-1991. See Appendix Table F-13.
to a little over 43 in 1991 (Figure 4-3~. In part, this reflects
trends in degree production. The number of degrees
awarded in the behavioral sciences actually fell for a period
of time and showed no strong trend in the latter part of the
decade.
Over 9S percent of employed behavioral science Ph.D.s
were U.S. citizens in 1991, unchanged from comparable
numbers for 1981 (Figure 4-44. Almost all U.S. citizens in
the behavioral sciences were native born, rather than natu-
ralized (94 percent). On the basis of these data, we can
conclude that immigration is not as important a source of
supply to the behavioral sciences as it is to other fields,
particularly the mathematical and physical sciences and en-
gineering. The comparable 1991 numbers for all science
and engineering fields were 93 percent for U.S. citizens and
83 percent for native-born citizens.s
Postdoctoral appointments are less important in the labor
market for behavioral scientists than for biomedical scien-
tists. Fewer than 1 percent of the behavioral science work
force were employed in postdoctoral positions (in contrast
to 7 percent in the biomedical sciences). Proportionately
loo
0
cat
L,
-
95
90
85
· U.S. Citizen
-
, ~1
[} U.S. Native-born
citizen
1 1 1
1981 1983 1985 1 987 1989 1 991
FIGURE 4-4 Citizenship status of employed behavioral science
Ph.D.s, 1981-1991. See Appendix Table F-14.
more women than men held postdoctoral appointments
(1 percent vs. 0.4 percent).6
The academic sector has not been as important as a
source of employment for behavioral scientists in recent
years as it has been for most other science fields. And it is
becoming even less important. Most behavioral scientists
were employed in settings outside the academic sector, with
strong representation in industry and in hospitals and clin-
ics.7 As noted earlier, this reflects in part the inclusion of
clinical psychologists in these statistics. Figure 4-S details
the trends. In 1981 roughly 55 percent of the behavioral
scientists were employed in the academic sector. By 1991
{} Academe
X IndustrY
Government
Hosp/Clinic
60
50
40
w
' 30
a}
20
10
o
Ax
x
X
-
I Other
~-C
-
X
WAX
C_
198 1 1983 1985 1987 1989 199 1
FIGURE 4-5 Employment sector of the U.S. behavioral science
work force, 1981-1991. See Appendix Table F-15.
this percentage had fallen to 42 percent. The shift from the
academic sector was almost entirely absorbed by the indus-
trial and clinical sectors, which rose from 19 percent in 1981
to 32 percent in 1991.
Degree Production and Career Patterns
There are many sources of talent available to the market
for behavioral sciences, and they operate in rather complex
ways. The distinctive role of clinical psychologists further
complicates interpretation of labor market data. The major
source of new behavioral science talent has traditionally
been our nation's university system, but some jobs in this
42
OCR for page 43
BEHAVIORAL SCIENCES PERSONNEL
market are filled by workers with degrees in closely related
fields or in special Ph.D. programs of research in social
work, nursing, and health services. Unlike the biomedical
labor market, however, immigration does not appear to be a
major supplier of talent.
Degree Production
The most readily available source of information de-
scribes degree production from U.S. universities, and the
committee summarizes this information below. In contrast
to the trends observed for the biomedical sciences, He data
reveal a significant downward trend in degree production.
The annual number of degrees produced in the behavioral
sciences fell between 1981 and 1989, rose between 1989
and 1991, and declined slightly between 1991 and 1992.
The net result was Hat degree production declined by about
12 percent (from 4,149 to 3,647) between 1981 and 1992
(Figure 4-6~. This downward trend in degree production
starkly contrasts with the increases experienced during this
5000
4800
4600
4400
4200 j
4000-
3800 .
3600 - .
3400 - .
3200
3000 1 , , , , ~, , ! 1
1 g81 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
FIGURE 4-6 Behavioral science Ph.D. production, 1981-1992.
NOTE: Data limited to U.S. citizens and permanent residents.
See Appendix Table F-16.
period by the biomedical sciences and by all fields of sci-
ence and engineering.8
Significant progress has been made in achieving gender
diversity, and the behavioral sciences have been in the fore-
front of this movement. Women have represented more
than half of this degree production since 1984 (Figure 4-7~.
The share of degrees granted to women increased from 44.2
percent to 58.7 percent between 1981 and 1992. Although
total degree production in the behavioral sciences remained
essentially unchanged during this period, the number of de-
grees granted to women rose 17 percent, from roughly 1,832
to about 2,142.
Less progress has been made with respect to race and
ethnic diversity, however. Whites constituted about 90 per-
cent of degree production in 1981-1992 (Table 4-2~. There
were small increases in the shares awarded to Asians and
Hispanics. The black share remained virtually unchanged.
70
60
50
1 ~-
q)
40
30
20
10
O
~J-
_
-I
1 1 1 1 1 1 1 1 1 1 1
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
FIGURE 4-7 Fraction of behavioral science Ph.D. degrees
earned each year by women, 1981-1992. NOTE: Data limited to
U.S. citizens and permanent residents. See Appendix Table F-16.
TABLE 4-2 Behavioral Ph.D. Production Over Time, by Race and Ethnicity
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
Total N 4022 3765 3896 3743 3586 3572 3477 3383 3353 3610 3681 3594
White % 92.0 9103 91.5 90.9 90.9 - 90.6 90c2 90.1 8907 89.2 88.5 89.1
Black 3.9 4.2 3.8 4.3 4.1 4.0 3.4 3.9 4.1 4.0 4.6 3.9
Hispanic 2.3 2.6 2.9 2.9 2.9 3.3 3.5 3.6 3.7 1.4 4.2 4.0
Asian 1.6 1.4 1.5 1.6 1.8 1.7 2.3 2.0 2.2 2.1 2.3 2.4
Native 0.3 0.5 0.3 0.2 0.4 0.4 0.5 0.4 0.4 0.6 0.4 0.6
American
NOTE: Cases with missing data are excluded. Data limited to U.S. citizens and permanent residents.
SOURCE: NRC, Survey of Earned Doctorates. (Annual)
43
OCR for page 44
MEETING THE NATION S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
Unlike biomedical degree recipients, very few new doc-
torates are not U.S. citizens and there is no strong trend
suggesting that this situation is changing. Roughly 94 per-
cent of all new doctorates in the behavioral sciences were
U.S. citizens in 1981. This figure had fallen to approxi-
mately 89 percent in 1992. As noted earlier, these statistics
do not imply any large changes in the future composition of
the behavioral science work force with respect to citizen-
ship status (Figure 4-8~.
Career Patterns
Given the objective of the NRSA awards to produce
research scientists it is useful to have some notion of the
number of years over the course of a career that these scien-
tists remain engaged in R&D. The effectiveness of the pro-
gram will vary with this number. As noted in Chapter 3,
the Survey of Doctoral Recipients a longitudinal survey
that tracks doctorates in the sciences, engineering and hu
'°°T
a, 901
c ~
of 85 t
80
75,
~ _ ~ ~
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
FIGURE 4-8 Fraction of behavioral science Ph.D. degrees
earned each year by U.S. citizens, 1981-1992. See Appendix Table
F-17.
inanities biennially-provides useful information on em-
ployment patterns, including postdoctoral work. This sur-
vey has the potential for illuminating career patterns of be-
havioral scientists. Thus, the Panel on Estimation
Procedures will examine more closely the feasibility of esti-
mating such patterns.
Market Conditions
This section presents selected indicators of short-term
market conditions, which include unemployment and un-
deremployment rates, postgraduation commitments of new
doctorates, and relative salaries.9
Ur~employmer't and Underemployment
As noted in Chapter 3, the most commonly used short-
term indicator of labor market conditions is the unemploy
44
1 2
0.8
0.6 t
0.4
0 2
\\~,/''J;" /~" \ ~
1973 1975 1977 1979 1981 1983 1985 1987 1989 1991
Or: · Behavioral Sciences I Physical Sciences |
FIGURE 4-9 Unemployment rates for behavioral and physical
sciences Ph.D.s, 1973-1991. See Appendix Table F-18.
ment rate. It is not as meaningful as an indicator of these
conditions for highly skilled workers, however. Because
such workers are usually able to find jobs even in times of
weak demand the issue is not whether the worker has a
job, but whether the job is fully utilizing his or her skills.
For this reason, the committee has also compiled informa-
tion on underemployment: i.e., workers who are working
part time but would prefer full-time jobs and workers who
have jobs that are outside of science and engineering and
who indicate they took these jobs because they could not
find work in science and engineering.
Figure 4-9 summarizes the unemployment rates. Given
the recent publicity about the weak state of demand in the
physical sciences, comparable rates for physical scientists
are also included so Cat the reader can assess conditions in
behavioral science labor markets relative to Dose in We
physical sciences.
Rates of unemployment have generally been in the range
of 1 percent, with little variability. The rate of underem-
ployment is also relatively low (Figure 4-10~.
3T
2.5
1 5 ~_~
1
0.5
1973 1975 1977 1979 1981 1983 1985 1987 1989 1991
· Behavioral Sciences {I Physical Sciences ~
FIGURE 4-10 Underemployment rates for behavioral and physi-
cal sciences Ph.D.s, 1973-1991. See Appendix Table F-l9.
OCR for page 45
BEHAVIORAL SCIENCES PERSONNEL
Postgraduation Commitments
Postgraduation plans of new doctorates may also reflect
market conditions. In particular, the percentage of new
Ph.D.s who indicate that they have definite commitments at
the time they are completing their requirements for the de-
gree can reflect the strength of demand. When demand is
weak, this percentage will fall; when demand is strong, this
percentage will rise.
Figure 4-11 summarizes these plans for 1975-1992. To
provide a comparative base, similar information is provided
for degree recipients in the physical sciences, a field thought
to be suffering currently from weak demand. The data re-
veal a declining trend in the percentage of new behavioral
science graduates who report definite employment or
postdoctorate commitments when they receive their de-
grees. In contrast the physical sciences display an upward
trend for 1975-1981 and a slow decline beginning in 1981
~ ~ 0 - ~
i05
Coo
.-
95 ~ ~
90
85
80
p cap
'I ~
1 1 1 1 ' 1 1 ; ; : 1
~CD ~CO O) O - ~US Cal ret oo ~0 ~
rat ~ret ~Go a, so GO 00 00 CO CO 00 00 0} a)
or a) a) o) a) Al or Al as ~Al as as as ~or en
_ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _
=~
-
a)
a,
· Behavioral Sciences ~ Physical Sciences
FIGURE 4-11 Fraction of new behavioral arid physical sciences
Ph.D.s with definite commitments, 1975-1992. See Appends
Table F-20.
that becomes more pronounced from 1989 to 1992 than the
decline for behavioral sciences. These data suggest that the
market was relatively stronger for behavioral scientists in
the early part of this period but deteriorated and became
relatively weak in more recent times.
Starting Salaries
Salaries are also considered to be a valid indicator of
market conditions. The commonly accepted economic
analysis of labor markets postulates a positive relationship
between relative salaries and relative demand. This rela-
tionship is expected to be strongest at the entry point of a
career, which typically occurs at around age 30 for behav-
ioral scientists. Median salaries of employed behavioral
science doctorates, age 30-34, who currently hold full-time
45
positions (excluding postdoctoral positions) relative to com-
parable salaries of all employed science and engineering
doctorates, age 30-34, are presented in Figure 4-12.~° Since
1983 these salaries have been declining relative to those of
comparable scientists and engineers in all fields. This sup-
ports the hypothesis that the relative demand for behavioral
scientists has been declining. However, the behavioral sci-
ences represent a heterogeneous collection of disciplines
with quite diverse employment paths. Clinical psycholo-
gists, for example, have increased their numbers and typi-
cally pursue career paths in sectors other than academia.
Also, social scientists have for many years found research
opportunities in business and industry, including self-em-
ployment. As a result of these differences in employment
opportunities, starting salaries are likely to be influenced by
"compositional" effects. Thus, these data must be inter-
preted with some caution.
96 ~_
94
92
90
-
~6
84
82
80
78 - 1 ~1 1 1 1
1973 1975 1977 1979 1981 1983 1985 1
\
\
\
\
an_
-
\
1 1 1
1987 1989 1991
FIGURE 4-12 Salaries of behavioral science Ph.D.s (ages 30-
34) who currently hold full-6me employment positions (excluding
postdoctoral positions) as a percentage of comparable salaries for
all scientists and engineers, 1973-1991. See Appendix Table F-
21.
OUTLOOK FOR BEHAVIORAL SCIENTISTS
A major goal of this committee is to estimate future
needs for behavioral scientists. As noted earlier in this re-
port, need can be defined in a variety of ways. In the con-
text of the labor market, need has often been expressed in
terms of job openings that must be filled to attain a particu-
lar employment objective. Selection of the specific em-
ployment objective to be reached is a policy decision usu-
ally made on normative grounds. Some job openings arise
from deaths, retirements, and other forms of separation from
the behavioral science work force. Other job openings are
generated by growth in employment opportunities. All of
these job openings may be filled by recruitment from many
talent pools: new doctorates, experienced doctorates from
other work forces or from outside the labor market (includ-
ing doctorates from abroad), nondoctorates, etc.
OCR for page 46
MEETING THE NAnON' S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
Given this broad context, the committee examines future
employment conditions in an effort to estimate need (ap-
proximated by job openings) and our ability to meet this
need (measured by new Ph.D.' s entering the behavioral sci-
ences workforce). Because job openings can be filled by
recruitment from a variety of talent pools, the reader is cau-
tioned that the committee's indicator of our ability to meet
this need represents a lower-bound estimate of this ability.
Table 4-3 shows the future number of job openings in
behavioral sciences to be filled under alternative scenarios
about employment growth. Three scenarios are examined:
zero growth; 3.0 percent per year~the 1981-1991 compound
growth rate for the behavioral science work force); and 1.5
percent per year (one-half the 1981-1991 compound rate).
As noted in Chapter 3, the method used to generate these
estimates is a variant of demographic cohort-survival mod-
els. It generates flows of workers into and out of this work
force and among the various employment states within this
work force. On the basis of these flows, it generates esti-
mates of changes in the size and composition of this work
force. There are, of course, many ways to do multistate
life table analysis. The data presented below should be
viewed as preliminary work by the committee, which will
be explored further by the Panel on Estimation Procedures
in the coming months.
Average annual estimates are developed for three time
periods: 1996-1997, 1998-1999, and 2000-2001. The esti-
mates are quite sensitive to the growth rate assumption,
varying from 1,404-1,512 for zero growth to 3,937-4,422
for 3.0 percent per year growth. The range is substantially
narrower for a given growth rate scenario. The modest in-
creases observed over time for a given rate of growth reflect
the widely anticipated increases in deaths and retirements in
the late 1990s as this work force ages. Except for the zero-
growth scenario, these increases also reflect Me expected
overall growth in the behavioral work force.~3
For comparison, Table 4-4 shows We number of new
behavioral Ph.D.s entering Me behavioral work force in pre-
vious years, estimated from Me longitudinal SDR (a sample
survey). Note that these numbers represent only a fraction
of the degree production Mat occurred in these fields, be-
cause a significant number of new graduates found employ-
ment in other fields or delayed entry into the work force.
An estimated 73 percent of the behavioral Ph.D.s entered
this work force during the period 1985-1990.
This level of work force entry, if maintained, could more
than meet the need for zero growth, but it will fall consider-
ably short of the number needed to maintain recent grown
rates.~4 It can be argued that maintenance of these recent
growth rates for behavioral scientists is an unrealistic sce-
nario. In the near future, universities are unlikely to in-
crease faculty size dramatically, federal spending on behav-
ioral research is not likely to increase in real terms, and
private sector demand (viz., industry) is not likely to in-
crease rapidly.
The best predictions for economic activity and R&D
funding in the near future suggest Mat demand for behav-
ioral scientists will grow slowly at best. Under these cir-
cumstances, maintenance of the current rate of entry of
Ph.D.s in the behavioral sciences should provide an ad-
equate supply for the years 1996-2001.
The NRSA program provides predoctoral support for
nearly 600 individuals in the behavioral sciences, although
only a fraction complete doctoral degrees in Me same year
TABLE 4-3 Committee Estimates of the Average Annual Number of lob Openings
Needed to Sustain Various Growth Rates of the Behavioral Science Work Forces b
Zero Growth Half the Average Average Growth Rate
Rate Scenario Growth Rate Scenarios Scenarios
. , .
Year Numbers Needed Numbers Needed Numbers Needed
1996-1997 1404 2592 3937
1998-1999 1441 2626 4076
2000 2001 1512 2780 4422
Behavioral science work force consists of those employed or on postdoctoral appointments In a behavioral field.
Data derived from the NRC Survey of Doctorate Recipients, a sample survey.
bBased on multistate life table methods. See Appendix G for methodology.
CHalf the average referred to in footnote d or 1~5 percent.
Refers to behavioral work force's average annual compound growth rate over the past decade or 3.0 percent
(3.5 percent, uncompounded).
46
OCR for page 47
BEHAVIORAL SCIENCES PERSONNEL
TABLE 4-4 Estimated Number of New Behavioral Ph.D.s
Entering the Behavioral Work Force in Selected Years.
Year
Numbers
1985-1986
1987-1988
1989-1990
2932
2797
2973
Manual averages.
NOTE: The Survey of Doctoral Recipients is a sample survey and subject to sampling error.
_ ~ _ A~ _ _r ,
SOURCE: NRC, Survey of Doctoral Recipients. (Biennial)
as receiving NRSA support. The number of behavioral de-
gree recipients in any year having had NRSA support is
unknown but presumed small.~5 Even if current levels of
predoctoral NRSA support are maintained, then the NRSA
program will most likely produce behavioral scientists at a
rate which future markets will absorb.
Relationship Between Market Outlook arid the Need
for Behavioral Research Personnel
The preceding analysis provides background information
with which we can anticipate the experiences of behavioral
scientists who will join the U.S. labor force in the coming
years. NRSA support although important in the behavioral
sciences represents, however, a small source of research
training support for new Ph.D.s in this area. Because NRSA
is such a small source of support, its effect on the aggregate
supply of behavioral scientists will be small. Thus, the jus-
tif~cation for the NRSA program in behavioral sciences must
be based heavily on need factors other than those generated
by aggregate market conditions.
In this report, we have taken an important first step in
adopting a new strategy for understanding aggregate
changes in the market for behavioral scientists. We have
concluded ~at, in He aggregate, behavioral scientists are
experiencing only modest employment problems. From our
global analysis of need associated with the growth in the
national research agenda, we expect that NRSA award re-
cipients in the behavioral sciences will continue to find em-
ployment. Further work is needed, however, before we
reach a more satisfactory understanding of He specific
needs and outlook for behavioral scientists engaged in
health-related research in areas of special interest the tar-
get group of the NRSA program. Moreover, in order to
assess the impact of the program on NRSA awarders, ca-
reer outcomes studies are needed.
47
Priority Fields
The national environment for behavioral and social sci-
ences has been variable during the past two decades (see,
for example, Gerstein et al., 1988~. For example, federal
obligations for basic and applied research in psychology,
sociology, and anthropology consistently declined through-
out most of the 1970s, reaching a low of $286 million in
1982.~7 After that time it began to slowly inch back up-
wards until 8 years later, when it totaled $445 million (NSB,
19914. From 1985 to 1990, funds for basic research in-
creased, on average, by $8.4 million per year, rising from
$162 million to $204 million. The average annual increase
in the applied research during that time was another $7 mil-
lion, increasing its worth from $206 million to $241 mil-
lion. Psychology has remained the primary beneficiary of
federal support, accounting for 76 percent ($340 million) of
the $445 million obligated in 1990, with commitments to
sociology and anthropology of $93 million and $12 million,
respectively.
Societal Problems
Despite the severe funding setbacks in the early 1980s,
there is now increasing recognition Hat many of the worst
problems facing this country are primarily behavioral in
character and that these sciences possess important infor-
mation to address those problems. The information that
has accumulated over years of careful basic research is ad-
equate to design developmental programs that have reason-
able prospects for success.
A key link in relating basic findings from the behavioral
sciences to a wide variety of clinical and applied settings is
the recruitment and training of high-quality people to carry
out this linkage. Perhaps the single largest potential for
doing this lies in the NIH pre- and postdoctoral fellowship
OCR for page 48
MEETING THE NAT10N'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
program. Were these to be viewed by NIH as a major activ-
ity, surely university-medical school consortiums would de-
sign fellowship opportunities which would produce scien-
tists well-trained both in aspects of the behavioral sciences
and in the problems encountered in medical settings.
NIH (as reconfigured in October 1992) has for many
years offered multidisciplinary training opportunities in re-
search related to the solution of such pressing social prob-
lems as substance abuse, violence, and the prevention of
infectious diseases. However, because of competing but
important developments in the biological bases of behavior,
NIH has directed less attention to research training in the
behavioral and social dimensions of physical and mental
health than might be desired. It is the hope of the commit-
tee that expanded NRSA support in the behavioral sciences
in the coming years will result in significant growth in the
number of awards made for research training in the solution
of social problems related to the health and well-being of
all Americans.
A Special Note on Clinical and Clinical
Services Research
Consideration should be given to enhancing federal sup-
port for training programs to produce clinical investigators
(see Kraut, 1993, for suggestions). Currently, some monies
are available to support mental health services research
training, but few opportunities exist for supporting He pro-
duction of behavioral science investigators in other problem
areas (e.g., alcohol and drug abuse interventions and psy-
chotherapy outcomes). Most clinical research Gaining be-
ing sponsored by the National Institute of Mental Health
(NIMH), for example, is primarily oriented toward psychia-
try; of the 35 clinical research training grants this year, only
5 are in psychology departments, with He remainder in psy-
chiatry units. As a result of NIMH's reorganization in 1985,
clinical training programs were relocated from the research
divisions to the services component and now, with the
movement of NIMH to NIH, are housed in the Center for
Mental Health Services (CMHS) in SAMHSA. Although
subsequent engagement in research fulfills the payback re-
quirement for clinical training, only a small proportion of
previously supported trainees actually utilized this option
(13.2 percent of all individuals engaged in payback require-
ments between 1981 and 1993), and a sample survey of
former trainees indicated that only a small fraction (1.8 per-
cent) cited research as Heir current primary activity (CMHS,
1993~. Furthermore, He clinical training budget totaled only
$2.9 million in 1992 and is a candidate for elimination in
congressional budget hearings.
Previous recommendations regarding the interdiscipli-
nary nature of conducting research in these areas (for both
behavioral and clinical scientists), exposure to applied re
48
search settings during the training process, and method-
ological training are critical to promoting these areas. Clini-
cal training programs are not always directed at research
competence per se (understanding research so as to keep up
with the literature vs. knowing how to do it). This is re-
flected by nearly half (48.6 percent) of all Ph.D.s and
Psy.D.s awarded in clinical psychology in 1989 being from
programs that emphasized the practitioner model rather than
the scientist-practitioner model of training, up from 28.7
percent in 1979 (Pion, 1993~. Even in programs that train
students for research careers, the clinical training process
itself, similar to medicine, is structured (of necessity) to
meet the increasing demands of accreditation agencies
whose aim is to ensure practice competencies but not mini-
mal standards of research conduct. For example, He re-
quirement of a predoctoral internship coupled with the need
of postdoctoral clinical training to meet licensing require-
ments affects both interest in and ability to carry out re-
search for both future clinicians and clinical faculty (many
clinical training programs require their faculty to be li-
censed). Expanding the pool of clinical researchers can be
accomplished in other ways by establishing specific pro-
grams (e.g., the program in clinical sociology at Yale) and
by exposing nonclinical students in traditional disciplines to
key substantive courses in prevention and treatment and re-
search opportunities in settings where they work with clini-
ciansonprojects. Forexample,certainproblemssurround-
ing diagnostic accuracy and clinical decision-making can
pose interesting research questions for cognitive psycholo-
gists and also lead to improvements in the application and
teaching of these skills. Transmission of He human immu-
nodeficiency virus (HIV) can be better understood by the
application of social network models and ethnographic stud-
ies, and models for interagency collaboration (e.g., resource
dependency theory) can be examined by organizational psy-
chologists and sociologists. In general, the issue of how
best to produce clinical investigators has recently generated
much attention, along with serious examination of how ac-
creditation practices have adversely affected He production
of clinician scientists (see the Accreditation Summit held in
1992 by the American Psychological Society). Aside from
the question of exactly how many clinical researchers are
needed, the data suggest that current programs are less suc-
cessful than expected in producing researchers who can ex-
amine key clinical problems and clinicians who can incor-
porate the logic of research into their service-delivery
activities. Concerted efforts are needed to identify the most
appropriate strategies for the development of clinical re-
searchers (perhaps through retrospective examination of
previous trainees and pilot tests of innovative programs.
Depending on the types of changes resulting from national
health care reform (e.g., the changes in service delivery set-
tings, practitioner responsibilities, and practitioner au
OCR for page 49
BEHAVIORAL SCIENCES PERSONNEL
tonomy as a result of managed care), opportunities for clini-
cal and clinical services research in these settings may ex-
pand along with the attractiveness of research careers.
ENSURING THE DIVERSITY
OF HUMAN RESOURCES
The characteristics of the graduate student population
have changed significantly since the 1960s. Thirty years
ago the flow of women into and through doctoral programs,
regardless of discipline, was exceedingly modest, with only
10.7 percent of the 9,733 doctorates awarded in the United
States earned by women. Beginning in the early 1970s,
however, the number and proportion of women doctorate
recipients rose dramatically. By 1991 it had more than
tripled, to 13,765, accounting for 36.8 percent of all new
doctorate recipients (Ries and Thurgood, 1993~. This
marked growth occurred in nearly every broad field of in-
quiry, particularly the behavioral sciences.
This pattern, however, has been a product of differential
trends in gender participation across disciplines. First, there
was a dramatic upsurge in the number of women earning
doctoral degrees, which was a product of the substantial
growth in the pool of women eligible to earn the doctorate
in all fields. Second, the pool of eligible men has remained
relatively stable or undergone some erosion, and this is re-
flected in the number being awarded doctoral degrees,
which has been steadily shrinking. Whereas the same per-
centage of women baccalaureates (approximately 3 percent)
has gone on successfully to earn the doctorate, the propor-
tion of men who enter and complete doctoral programs
dropped from 10 percent to 4 percent during this same pe-
riod. Consequently, the flow of women into the graduate
training pipeline has increased, but the stream of men sub-
sided from its early-1970s level.
Other factors aside (e.g., the quality of baccalaureate
training), the recent upsurge in the awarding of baccalaure-
ate degrees suggests a larger pool of candidates who may be
potentially interested in pursuing advanced degrees in these
areas. It also suggests that if current patterns persist, the
composition of the graduate student population will change
accordingly (see following section). Among current fresh-
men, interest in the social and behavioral sciences has in-
creased, with much of the growth attributable to females.
In 1990, 55 percent of all freshmen intending to major in a
science or engineering field chose psychology or one of the
social sciences as compared with 39 percent in 1983 (NSF,
1991~. Given that freshmen intentions have signaled trends
in baccalaureate production, graduate training programs can
expect the pool of prospective applicants to have, at least in
the short term, a large proportion of women, and although
more ethnically diverse than many other science and engi-
neering fields, to remain sparsely populated by people of
49
color. Ethnic minority representation has remained rela-
tively stable, with about 1 of every 10 graduate students
being Black, Asian, Native American, or Hispanic. The
recruitment and retention of ethnic minorities has been and
remains an important concern.
In addition, the long-term implications of shifting gender
distributions have come under review (e.g., APA Task Force
on the Changing Gender Composition of Psychology), with
available research suggesting that differences exist in career
tracks and advancement. Although often each difference is
reasonably small, the results are cumulative advantages and
disadvantages that appear gender-specific (e.g., Hornig,
1987; Long, 1990; Zuckerman, 19874. Such disparities do
not work toward maintaining the health and vitality of the
investigator pool, and well-designed research studies inves-
tigating the antecedent factors for choosing and pursuing a
research career are urgently needed.
TEW NRSA PROGRAM IN
THE BEHAVIORAL SCIENCES
For operational purposes, the committee created and
first reported in 1975 a tentative taxonomy of behavioral
science research fields using the departmental classification
scheme of the Doctorate Records File maintained by the
NRC. Throughout its tenure the committee continued to
define the behavioral sciences as psychology, anthropology,
sociology, and speech and hearing sciences because these
were the fields most closely involved in investigating health
problems.
In its earliest reports, the committee presented the labor
market outlook for Ph.D.s in the behavioral sciences as a
whole, without distinguishing fields. In 1978 the commit-
tee, realizing that the analysis of He labor market was hin-
dered by treating He behavioral sciences as a single entity,
separated the data into clinical (clinical psychology, coun-
seling and guidance, and school psychology) and nonclinical
(nonclinical psychology, anthropology, sociology, and
speech and hearing sciences) fields. This disaggregation
enabled the identification of divergent market trends within
the behavioral sciences.
For its 1985 report the committee decided to carry the
disaggregation one step further and divided the nonclinical
fields into nonclinical psychology and other behavioral sci-
ences. This step yielded Tree behavioral science subdivi-
sions: clinical psychology, nonclinical psychology, and
other behavioral science fields (sociology, anthropology,
and speech pathology and audiology).
With this additional level of disaggregation, substantial
differences among disciplines and education levels (gradu-
ate and undergraduate) began to emerge in the 1980s. For
example:
OCR for page 50
MEETING THE NATION'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
· undergraduate enrollments in psychology were not
subject to the decline experienced by He other behavioral
fields;
· graduate enrollments in psychology declined less than
those in other behavioral science fields;
· the decline in R&D funding occurred in behavioral
fields outside of psychology;
· academic employment continued to grow, but the in-
creases were concentrated in the fields outside of psychol-
ogy: sociology, anthropology, and speech pathology and
audiology; and
· although the number of behavioral scientists on
postdoctoral appointnents rose, the number of nonclinical
psychologists with postdoctoral appointments fell to its low-
est level since the committee began monitoring these data.
In general, the committee' s recommendations for behav-
ioral sciences personnel were consistent throughout the
1970s with respect to the numbers of trainees and fellows
supported, level of dollars expended, and distribution of
awards between predoctoral levels. The core of the recom-
mendations was, within the framework of a constant dollar
level of support, to redistribute awards in the behavioral
sciences from the traditional predoctoral emphasis (70 per-
cent/30 percent, predoctoral to postdoctoral) to one of
postdoctoral emphasis (70 percent/30 percent, postdoctoral
to predoctoral). The committee recommended that this tran-
sition be completed by fiscal year 1981. The rationale for
the recommendation was that the less favorable labor mar-
ket for behavioral scientists, especially in academia, dic-
tated a decrease in predoctoral support whereas the growing
sophistication of behavioral research in the area of health
warranted an increase in postdoctoral support.
During the early 1980s the committee reported that a
shift in emphasis to postdoctoral training awards had oc-
curred, but not in the gradual and orderly manner recom-
mended. Instead, the ratio had moved slowly toward a
higher postdoctoral concentration because of a sharp de-
cline in predoctoral awards and a level of postdoctoral sup-
port Hat had risen between 1975 and 1977 and then re-
mained basically constant through 1981.
Although He ADAMHA had supported the committee's
recommendations and made strong efforts to implement
them, significant external impediments had prevented
ADAMHA's ability to respond more rapidly: overall funds
available for research training grants and fellowships had
lost ground to inflation, mandated increases in stipend lev-
els in fiscal year 1980 meant that fewer pro- and
postdoctoral students could be supported, and professional
institutions had been slow in broadening Heir emphasis to
include more postdoctoral training positions.
The committee in its 1983 report recommended that de-
spite some necessary budgetary reductions, some training
50
support must be preserved in the health-related behavioral
and social sciences and directed to the highest quality pro-
grams. The recommended shift in emphasis from
predoctoral to postdoctoral awards was reaffixed, but some
numerical modifications were necessary because of budget-
ary constraints. The committee recommended that
predoctoral training should remain at the 1981 level (about
650 awards) whereas postdoctoral support should increase
only modestly from the 1981 level (about 350 awards) to
about 540 awards by 1987.
By the mid 1980s the committee was reporting an aca-
demic labor market for behavioral science Ph.D.s that was
expanding at a moderate rate of growth. However, most of
the growth had occurred in the field of clinical psychology.
This finding caused the committee concern from the stand-
point of the federal government's research program because
most clinical psychologists work outside He academic sec-
tor and do not contribute to the research effort. Also of
concern to the committee was the continuing decline of
R&D funding and the observation that graduate enrollments
and Ph.D. production in nonclinical fields were beginning
to decline whereas those in the clinical fields were continu-
ing to increase. However, the committee noted that a sub-
stantial number of behavioral science courses were being
taught to graduate students at professional schools public
health, law, medicine, nursing, social work, and business
which would tend to increase the demand for behavioral
scientists.
Recommendations in 1985 were for 450-790 predoctoral
trainees per year and 460-800 postdoctoral trainees per year.
The committee also strongly endorsed the continuation of
the training grant mechanisms as a way of improving the
quality of graduate education in the behavioral sciences at
the predoctoral and postdoctoral levels. It was recom-
mended that training grants should be the predominant
mechanism of support, with an 80 percent/20 percent ratio
of traineeships to fellowships.
Finally, in its 1989 report, the committee projected that
the labor market for behavioral scientists would be fairly
stable. It recommended that the level of predoctoral and
postdoctoral support be kept at their current levels of ap-
proximately 500 and 420 positions, respectively. Given the
low level of research involvement by clinical psychologists,
the committee recommended moving support away from
clinical psychology and toward nonclinical psychology and
other behavioral sciences.
RECOMMENDATIONS
Total support for research training in the behavioral sci-
ences increased slightly from about 902 awards in fiscal
1991 to an estimated 1,069 awards in fiscal 1993 (Table 4-
5~. Most awards are offered as institutional training grants
OCR for page 51
BEHAVIORAL SCIENCES PERSONNEL
TABLE 4-5 Aggregated Numbers of NRSA Supported Trainees and Fellows in
Behavioral Sciences for FY 1991 through FY 1993
Fiscal
Year
1991
Level of Training
Number of awards
Predoctoral
Pool
MARC Undergraduate
Type of Support
TOTAL Traineeship Fellowship
902
519
338
45
775
472
258
45
127
47
80
1992 Number of awards 908 790 118
Predoctoral 534 481 53
Postdoctoral 323 258 65
MARC Undergraduate 51 51
1993 Number of awards 1,069 930 139
Predoctoral 672 604 68
Postdoctoral 349 278 71
MARC Undergraduate 48 48
NOTE: Based on estimates provided by the National Institutes of Health. See Summary Table 1.
(traineeships), which account for about 85 percent of total
NRSA support in the behavioral sciences. Emphasis has
been given to Predoctoral support, although institutional
training grants permit the mix of Predoctoral and postdoctoral
trainees. Individual fellowships have been made available at
both the Predoctoral and postdoctoral levels.
Predoctoral Training
On the basis of continuing gains being made by behav-
ioral scientists in areas of national interest and of antici-
pated employment opportunities for highly skilled research-
ers, the committee urges the continued expansion of federal
support through Predoctoral awards in the behavioral sci-
ences. Predoctoral awards permit the preparation of inves-
tigators familiar with the broad range of research techniques
and theories that characterize doctoral preparation in the
behavioral sciences. Many graduates are ready to assume
research positions on completion of the doctoral degree, al-
though postdoctoral training has gained some momentum in
certain component subfields.
As is the case in other areas, the committee is concerned
that current low stipend levels for NRSA awardees serve as
a disincentive to attract the most able scientists to research
careers in health-related fields. Thus, the committee has
tempered its call for expansion in total support in the be-
havioral sciences in recognition of the competing need to
increase stipend support (Table 4-6~.
51
RECOMMENDATION: The committee recommends
that the number of Predoctoral trainees and fellows sup-
ported annually in the behavioral sciences increase frown
an estimated 672 in fiscal 1993 to 900 by fiscal 1996.
The expansion in support should maintain the same ratio
of trainees to fellows.
Postdoctoral Training
Postdoctoral research training Trough the NRSA award
provides the nation with a mechanism to attract the most
skilled scientists to address areas of national need. Because
of differences in the evolution of research careers,
postdoctoral research training plays a greater role in some
behavioral sciences than others. Nonetheless, postdoctoral
studies increase the technical skills of the investigator and
strengthen the pool of talent available to the nation for re-
search.
RECOMMENDATION: The committee recommends
that the number of postdoctoral trainees and fellows sup-
ported annually in the behavioral sciences increase from
approximately 349 awardees in fiscal 1993 to 500 in fis-
cal 1996.
Minority Access to Research Careers
Since its inception in the late 1970s, the special program
of undergraduate support for Minority Access to Research
OCR for page 52
MEETING THE NATION'S NEEDS FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS
TABLE 4-6 Committee Recommendations for Relative Distribution of Predoctoral
and Postdoctoral Tra~neeship and Fellowship Awards in Behavioral Sciences
for FY 1994 through FY 1999
Fiscal
Year
1994
Type of Support
TOTAL Traineeship Fellowship
Recon~rnended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1,195
745
400
50
1995 Recommended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1996 Recommended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1997 Recommended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1998 Recommended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1999 Recommended number of awards
Predoctoral
Postdoctoral
MARC Undergraduate
1,325
825
450
50
900
500
50
1,450
900
500
50
1,450
900
500
50
1,450
900
500
50
1,040
670
320
50
1,150
740
360
50
1,260
810
40()
50
1,260
810
400
50
1,260
810
400
50
1,260
810
400
50
155
75
80
175
85
90
190
90
100
190
90
100
190
90
100
-
190
90
100
-
Careers (MARC) has included a small number of awards
for research Raining in He behavioral sciences. In recent
years, about 50 awards were made for Gaining in this area
(Table 4-5~.
The MARC program represents a unique and special pro-
gram of support for the recruitment of minorities into re-
search careers in the behavioral sciences. We look forward
to reviewing the outcome of the review of the MARC pro-
gram presently being conducted under the auspices of NIH.
We endorse the continuation of MARC support for research
Raining in He behavioral sciences at current levels until
that study is complete.
RECOMMENDATION: The committee recommends
Hat the number of NRSA awards for research Raining in
52
the behavioral sciences through the MARC program be
maintained at about 50 per year, pending completion of
the NIH review.
NOTES
1. Broadly construed to include psychology, sociology, anthropology,
and speech and hearing sciences. See Appendix B for a list of disciplines
included in this area.
2. There was a significant drop in the academic sector work force
between 1989 and 1991. In part, this decline may reflect methodological
changes that were introduced at that time to the Survey from which these
data were generated. However, the decline may be reflecting a weakening
of demand in the academic sector.
3. Approximately 30,000 thousand of the 67,000 workers employed
in the behavioral sciences in 1991 were clinical psychologists whose par-
ticination in health-related research varies from year to year.
OCR for page 53
BEHAVIORAL SCIENCES PERSONNEL
4. See Appendix F.
5. Survey of Doctoral Recipients (SDR) data, special tabulations.
6. See Appendix F.
7. In 1989 hospitals and other nonprofit institutions employed 23 per-
cent of the clinical psychologists, seven percent of the nonclinical psy-
chologists, and five percent of the sociologists and anthropologists (Pion,
1993). The numbers in Pion's paper come from Quantum Research
Corporation's analysis of the SDR data. Those numbers are not strictly
comparable with labor force numbers used elsewhere in this report, due to
differences in definition of variables.
8. See Chapter 3, Reis, P. and D.H. Thurgood, Summary Report 1992:
Doctorate Recipients from United States Universities, Appendix Table
B-1.
Clark, G.A. and E.M. Schuman
1992 Snails' tales: Initial comparisons of synaptic plasticity under-
lying learning in Hermissenda and Aplysia. In
Neuropsychology of Memory, 2nd Edition, ed. L.W. Squire
and N. Butters. New York: Guilford Press.
Consortium of Social Science Associations (COSSA)
1994 Washington Update. 13(4):18. March 7.
Costa, P.T., Jr. and R.R. McCrae
1987 Neuroticism, somatic complaints, and disease: Is the bark
worse than the bite? Journal of Personality. 55:299-316.
Cox, D.J. and L. Gonder-Frederick
9. Although the number of postdoctoral appointments are also thought
to be strongly influenced by current job opportunities, we do not include
them as an indicator for behavioral scientists because, as shown earlier,
they are a relatively unimportant part of this labor market.
10. Starting salaries are defined as the salaries of workers in the age
group 30-34.
11. See IOM, 1994 for further discussion of this phenomenon.
12. For more detail, Chapter 3, note 11 and Appendix G.
13. The increases could also be reflecting the impact of the expected Green, D.M. and J.A. Swets
aging of this work force on separations other than deaths and retirements. 1966
See supra, Chapter 3, note 9.
14. The estimate is presented as a minimum value because these job Goodwin, F.K.
Openings could also be filled by recruiting workers with degrees and kain
ing in closely related fields or workers from abroad.
15. On the basis of information gathered from the National Science
Foundation the committee estimates that fewer than 5 percent of graduate
students in the behavioral sciences received NRSA support in FY 1990.
16. See Appendix A for a brief review of earlier NRC studies in this
area.
17. These figures are in constant 1982 dollars.
18. Some lessons can be learned from the training of physician-scien
tists and the performance of trainees who are supported during predoctoral Hornig, L.
years (M.D.-Ph.D. programs) vs. postdoctoral support on research-related 1987
outcomes.
REFERENCES
Adler, N. and K. Matthews
1994 Health Psychology: Why do some people get sick and some Kandel, E.R.
stay well? In Annual Review of Psychology, Volume 45, ed. 1976
L. Porter and M. Rosenzweig, pp. 229-259. Palo Alto, CA:
Annual Reviews Inc.
Bachevalier, J.
1991 Cortical versus limbic immaturity: Relationship to infantile
amnesia. In Developmental Behavioral Neuroscience, ed.
M.R. Gunner and C.A. Nelson, pp. 129-153. Hillsdale, 1~J:
Lawrence Erlbaum Associates, Publishers.
Beardsley, G. and M.G. Goldstein
1993 Psychological factors affecting physical condition. Endocrine
disease literature review. Psychosomatics. 34(1):12-19.
Brown, J.D. and K.L. McGill
1989 The cost of good fortune: When positive life events produce
negative health consequences. Joun~l of Personality and So
cial Psychology. 57 :1103-11 10.
Center for Mental Health Services (CMHS)
1993 Report to Congress on Mental Health Clinical Training Pro
gram Payback Requirements. June. Rockville, MD: Sub
stance Abuse and Mental Health Services Administration 1991
(SAMHSA).
53
Institute of Medicine
1992 Major developments in behavioral diabetes research. Joumal
of Consulting and Clinical Psychology. 60:628-638.
Fitzgerald, K.
1989 Medical electronics. IEEE Spectrum. 26:67-69.
Gerstein, D., R.D. Luce, and N.J. Smelser
1988 The Behavioral and Social Sciences: Achievements and Op-
portunities. V\Tashington, D.C.: National Academy Press.
Getty, D.J., R.M. Pickett, C.J. D'Orsi, and J.A. Swets
1988 Enhanced interpretation of diagnostic images. Investigative
Radiology. 23:240-252.
1966 Signal Detection Theory and Psychophysics. New York: Rob
ert E. Krieger Publishing Company.
1993 NIH adds institutes on mental and addictive disorders. NIH
News and Features. Summer: 12-25.
Hahn, R.C. and D.B. Petitti
1988 Minnesota Multiphasic Personality Inventory-rated depression
and the incidence of breast cancer. Cancer. 61:845-848.
Hitchcock, J.M. and M. Davis
1986 Lesions of the amygdala, but not of the cerebellum or red
nucleus, block conditioned fear as measured with the potenti
ated startle paradigm. Beh~ioral Neuroscience. 100:1 1-22.
Women Graduate Students. In Women: Their
Underrepresentation and Career Differentials in Science and
Engineering, ed. L.S. Dix, pp. 103-122. Washington, D.C.:
National Academy Press.
1994 Careers in Clinical Research: Obstacles and Opportunities.
Washington, D.C.: National Academy Press.
Cellular Basis of Behavior. San Francisco: W.H. Freeman
and Co~npany.
Kraut, A.
In Press Presentatiol1 at the Committee Public Hearing, May 3, 1993.
In Meeting the Nation's Needs for Biomedical and Behav-
ioral Scient~sts: Summary of the 1993 Public Hearing. Wash-
ington, D.C.: National Academy Press.
Loftus, G.R. and E.F. Loftus
1976 Huw~n Memory: The Processing of Information. Hillsdale,
NJ: Lawrence Erlbaum Associates, Publishers.
Long, J.S.
1990 The Origins of Sex Differences in Science. Social Forces.
68(4):1297-1315.
Luce, R.D., N.J. Smelser, and D.R. Gerstein
1989 Leading Edges in Social and Behavioral Science. New York:
Russell Sage Foundation.
Markovitz, J. H., K.A. Matthews, R.R. Wing, L.H. Kuller, and E.N.
Meilahn
Psychological, biological, and health behavior predictors of
blood pressure change in middle-aged women. Joumal of
Hypertension. 9:399-406.
OCR for page 54
MEETING THE NAT10N'S NEEDS FOR BIOMEDICAL AtID BEHAVIORAL SCIENTISTS
Matthews, K. A., K.L. Woodall, T.O. Engebretson, B.S. McCann, and
C.M. Stoney
1992
McGaugh, J.
1989
Influence of age, sex, and family on Type A and hostile atti-
tudesandbehaviors. Health Psychology. 11:317-323.
Involvement of hormone and neuromodulatory systems in the
regulation of memory storage. Annual Review of Neuro-
science. 12:255-287.
McNeil, B.J., S.G. Pauker, H.C. Sox, Jr., and A. Tversky
1982 On the elicitation of preference for alternative therapies. New
England Journal of Medicine. 306:1259-1262.
Metz, C.E.
1986
Squire, L.R.
ROC methodology in radiologic imaging. Investigative Radi- 1992
ology. 21:720-733.
Meuner, M., J. Bachevalier, M. Mishkin, and E.A. Murray
1993 Effects on visual recognition of combined and separate abla- Swets, J.A.
lions of the entorhinal perirhinal cortex in rhesus monkeys. 1979
Journal of Neuroscience. 13(12):5418-5432.
National Institutes of Health (NIH)
1991 Health and Behavior Research: NIH Report to Congress.
Bethesda, MD: National Institutes of Health.
1993 The NIH Implementation Plan for Health and Behavior Re
search. Bethesda, MD: NationalInstitutes of Health.
National Science Board (NSB)
1991 Science and Engineering Indicators - 1991. Washington,
D.C.: National Science Foundation.
O'Keefe, J. and L. Nadel
1978 The Hippocampus as a Cognitive Map. London: Oxford Uni
versity Press.
Pion, G.
1993 Trends in behavioral science human resources: their implica
tions for research training. Paper prepared for the Committee 1987
on National Needs for Biomedical and Behavioral Research
Personnel.
Ries, P. and D.H. Thurgood
1993 Summary Report 1992: Doctorate Recipients from United
States Universities. Washington, D.C.: National Academy
Press.
54
Salovey, P. and D. Birnbaum
1989 Influence of mood on health-relevant cognitions. Journal of
Personality and Social Psychology. 57:539-51.
Scheier, M. F. and C.S. Carver
1992 Effects of optimism on psychological and physical well-be-
ing: theoretical overview and empirical update. Cognitive
Therapy and Research. 16:201-228.
Siegrist, J., R. Peter, A. Junge, P. Cremer, and D. Seidel
1990 Low status control, high effort at work and ischemic heart
disease: prospective evidence from blue-collar men. Social
Science and Medicine. 31:1127-1134.
Memory and the hippocampus: a synthesis from findings with
rats, monkeys, and humans. Psychological Review. 99:195
231.
ROC analysis applied to the evaluation of medical imaging
techniques. Investigative Radiology. 14:109-121.
1988 Measuring the accuracy of diagnostic systems. Science.
240:1285-1293.
Swets, J.A., D.J. Getty, R.M. Pickett, C.J. D'Orsi, S.E. Seltzer, and B.J.
McNeil
1991 Enhancing and evaluating diagnostic accuracy. Medical De-
cisionMaking. 11:9-18.
Swets, J.A. and R.M. Pickett
1982 Evaluation of Diagnostic Systems: Methods from Signal De-
tection Theory. New York: Academic Press.
Thompson, R.F.
1986 The neurobiology of learning and memory. Science. 223:941-
947.
Zuckerman, H.
Persistence and Change in the Careers of Men and Women
Scientists and Engineers. In Women: Their
Underrepresentation and Career Differentials in Science and
Engineering, ed. L.S. Dix, pp. 127-156. Washington, D.C.:
National Academy Press.
Representative terms from entire chapter:
behavioral science
