Feist1998 - Psicologia Da Ciência

Review of General Psychology Copyright 1998 by the Educational Publishing Foundation
1998, Vol. 2, No. 1,3-47 1089-268W98/$3.00
The Psychology of Science: Review and Integration

of a Nascent Discipline
Gregory J. Feist Michael E. Gorman
College ofWilliam and Mary University of Virginia
Disciplines that study science are relatively well established in philosophy, history, and
sociology. Psychology of science, by comparison, is a late bloomer but has recently
shown signs of codification. The authors further this codification by integrating and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
reviewing the growing literature in the developmental, cognitive, personality, and social
This document is copyrighted by the American Psychological Association or one of its allied publishers.
psychology of science. Only by integrating the findings from each of these perspectives
can the basic questions in the study of scientific behavior be answered: Who becomes a
scientist and what role do biology, family, school, and gender play? Are productivity,
scientific reasoning, and theory acceptance influenced by age? What thought processes
and heuristics lead to successful discovery? What personality characteristics distinguish
scientists from nonscientists and eminent from less eminent scientists? Finally, how do
intergroup relations and social forces influence scientific behavior? A model that
integrates the consensual empirical findings from die psychology of science is pro-
posed.
Without the addition of a psychological dimension, I about the psychology of the scientist" (p. 349).
believe, it is impossible to appreciate fully the essence In contrast, other disciplines like philosophy,
of the scientific imagination. And without this apprecia-
tion, the origins of science, the emergence of new ideas history, and sociology have spawned clearly
about natural phenomena, must escape our grasp. identifiable subdisciplines devoted to science
Psychology is mandatory if we wish to comprehend the studies. We firmly believe that science deserves
scientific genius as the generator of science. (Simonton, more attention from psychologists, and one of
1988a, p. 200)
the goals of this review is to show how the
It is indisputable that the growth of science psychology of science has grown from being the
and the development of technology have trans- amorphous and scattered field that Fisch (1977)
formed the world, both physically and cultur- and Mahoney (1979) described to a codified,
ally. For this reason, science ought to be an albeit nascent, subdiscipline today. Psychology
object of intense psychological study. However, of science can benefit not only other psycholo-
efforts to study science and technology from a gists, but philosophers, historians, and sociolo-
psychological perspective are scattered across gists also.
the disciplines of psychology, and often there is The psychology of science applies the empiri-
too little communication among those involved. cal methods of psychological investigation to
As Mahoney (1979) wrote in the late 1970s, "In the study of scientific behavior. In other words,
terms of behavior patterns, affect, and even it is the empirical study of the cognitive,
some intellectual matters, we know more about biological, developmental, personality, and so-
alcoholics, Christians, and criminals than we do cial influences of those individuals who are
involved in the enterprise of science or who are
Gregory J. Feist, Department of Psychology, College of simulating scientific problem solving. In this
William and Mary; Michael E. Gorman, School of sense, the psychology of science is primarily
Engineering and Applied Sciences, University of Virginia. descriptive, describing actual behavior, rather
We wish to thank Eve Sou's and Stephanie Blasi for their
assistance in literature searches. We also are grateful to Erika than prescriptive (describing ideal behavior).
Rosenberg, David Klahr, Joshua Klayman, and Eric Consider the case of Michael Faraday, discov-
Freedman for their helpful feedback on earlier versions of erer of electromagnetic fields. Faraday's discov-
this article. ery can be analyzed from developmental,
Correspondence concerning this article should be ad-
dressed to Gregory J. Feist, P.O. Box 8795, College of cognitive, personality, or social psychological
William and Mary, Williamsburg, Virginia 23187-8795. perspectives. From a developmental perspec-
Electronic mail may be sent to gjfeis@facstaff.wm.edu. tive, one could study the shifts in Faraday's
FEIST AND GORMAN
beliefs, methods, and productivity rates as he article by Singer (1971), who pointed out that
grew older. Explicating the cognitive processes although a new "science of science" was a
(confirmation vs. falsification, analogy and nascent discipline as far back as the 1930s,
metaphor, ideation, and elaboration) used by "some 30 years have passed, and we do not as
Faraday in his integration of electromagnetic yet have a developed, self-conscious discipline
fields would shed important light on his of a science of science. We are now, however, in
discovery. Faraday's personality was important a better position to anticipate its arrival" (p.
in his rise from bookbinder to fellow of the 1010). Another exception from the 1970s was
Royal Society; the same patient, methodical the first major review of the field (Fisch, 1977).
devotion to a task is apparent in the way he Toward the end of the decade, Fisch echoed
bound books, conducted experiments, and kept Singer's concern and opened his review by
notebooks detailing results and speculations. pointing out the disparate and unsystematic
Furthermore, the social context from which his nature of investigations into the psychological
discoveries stemmed are critical to understand- attributes of scientists. He concluded his review
ing his scientific process. He belonged to a pessimistically: "Having now reviewed the
religious organization that referred to them- field, it is lamentably clear that basic concepts
selves as Sandemanians, and from this group are diffuse and contradictory, and rarely become
stemmed his faith that the book of nature could common to several investigations. For this and
be read by anyone who devoted careful time and other reasons, results cannot really be compared,
attention. To understand a Faraday or any other and little scholarly cumulation has resulted" (p.
scientist, a psychological perspective is as 298). In another review of the literature just 2
important and necessary as philosophical, histori- years later, Mahoney (1979) reached similarly
pessimistic conclusions about the state of the
cal, or sociological ones.
field.
A Brief History of the Psychology However, since the early to mid 1980s there
of Science has been a steady surge in works devoted to the
psychological underpinnings of science (Ghol-
No definite date can be given for the birth of son, Shadish, Neimeyer, & Houts, 1989). This
the psychology of science. Stevens (1936,1939) surge was so evident that Shadish, Fuller, and
wrote on the psychology of science in the 1930s, Gorman in 1994 could proclaim that the
but Roe's (1952a, 1952b, 1953) classic work, "psychology of science has finally arrived" (p.
along with Cattell's (R. Cattell & Drevdahl, 3). Furthermore, "substantively, psychological
1955), foreshadowed the burst of research on contributions to science studies are increasing in
psychological attributes of scientists that oc- frequency and quality. Sociologically, psycholo-
curred in the early 1960s. In general, studies in gists are beginning to identify themselves as
the 1960s placed a heavy emphasis on creativity interested in the topic" (Shadish, Houts, Ghol-
in science (Chambers, 1964; Eiduson, 1962; son, & Neimeyer, 1989, p. 1). Granted, many
Gough & Woodworth, 1960; Taylor & Barron, psychologists who study scientists and the
1963). In addition, Maslow published a book in scientific process often do so not explicitly from
1966 with the title The Psychology of Science, in a psychology of science perspective, and one
which he argued for expanding the scope of purpose of this review is to make the connection
traditional mechanistic, reductionistic views of explicit.
science to include a broader, more humanistic Before we review the literature in each
and psychological conceptualization of science. substantive area we should note a few caveats.
Finally, a precursor to the entire discipline of First, our organization is empirical rather than
cognitive psychology of science also could be theoretical. We believe that the first stage in
seen in Herbert Simon's chapter on scientific establishing a discipline is to demonstrate
discovery and the psychology of problem descriptive consensus. Second, we do not claim
solving (Simon, 1966). that the reviews are exhaustive, but we made
During the 1970s, however, there was a every effort to locate as many articles as
decline in research on the psychology of possible on the relevant topics and extract the
science, and few major works were produced on general findings. Some topics fall outside of or
the topic. One exception was a conceptual cut across traditional subdisciplinary bound-
PSYCHOLOGY OF SCIENCE
aries, in particular creativity and productivity, ics. Some children begin to display incredible
and therefore will be discussed under more than mathematical computational and reasoning skills
one category. In addition, we have chosen to as early as 2 or 3 years old, and by 10 years of
focus on the four most developed subdisci- age are already performing complex mathemati-
plines—developmental, cognitive, personality, cal calculations (Bell, 1937; Kanigel, 1991;
and social psychologies of science. We could Wiener, 1953). The list of historical examples of
have perhaps included a fifth section on the less innate and precocious mathematical genius is
developed field of the biological psychology of long and impressive: Pascal, Newton, Leibniz,
science. However, we chose to include it in the Laplace, Gauss, Boole, Wiener, Ramanujan, and
developmental section because biological and Feynman, to name but a few of the truly
genetic influences have been discussed primar- outstanding examples (Bell, 1937; Gleick, 1992;
ily in relation to gender differences in, and the Kanigel, 1991; Wiener, 1953). With the excep-
development of, mathematical ability. tion of the Bernoullis, most of these mathemati-
cally precocious geniuses came from humble
Developmental Psychology of Science and nonmathematical families (Bell, 1937).
Although some researchers have used family
One of the first and most interesting questions lineage evidence to infer genetic influence,
that can be addressed by the psychology of familial accumulation per se is irrelevant in any
science concerns how and why certain individu- nature-nurture debate (Eysenck, 1988, 1993,
als become scientists. What is the origin of the 1995). This is so for the simple reason that
necessary talents and skills required to be a genetics and environment are inherently con-
scientist? Why do some possess these talents founded within families. Theoretically, if a trait
and others not? How do these talents and were 100% genetically determined or 100%
abilities change and develop with age? These environmentally determined it would accumu-
questions are the core focus of the developmen- late in families in either case (Eysenck, 1988). If
tal psychology of science. Developmental psy- mathematical ability is to some extent geneti-
chology of science has much overlap with the cally determined, then how is it possible to have
social psychology of science because children mathematical genius spring from nonmathemati-
and adolescents are dependent on others (usu- cal families? As Bouchard and Segal (1990)
ally parents) for survival. In fact, developmental argued, such innate genius may demonstrate
questions can be placed on a continuum from Lykken's principle of emergenesis, which is
the relatively nonsocial to the very social. The defined as "the inheritance of a unique configu-
relatively nonsocial questions include biological ration of genetic factors that may explain the
and genetic influences of math ability and hereditary transmission of traits that do not
creative genius, and the moderately social topics appear to run in families" (Bouchard & Segal,
include age and productivity, age and receptivity 1990, p. 192). In other words, some genetic
to new discoveries, and gender and science. traits are so complex, and are made up of "a
Finally, the more purely social questions include configuration—rather than by a simple sum—of
mentoring and training, family influences, and polymorphic genes" (Lykken, McGue, Tellegen,
religious background. Of course, the social & Bouchard, 1992, p. 1565), that even though
continuum is meant only as a heuristic because they are genetically influenced they are not
biology and environment do have mutually likely to run in families. There is some evidence
reinforcing influences on each other. that creativity and genius are such complex
traits (Lykken et al., 1992; Waller, Bouchard,
Precocity, Giftedness, and Creativity: Lykken, Tellegen, & Blacker, 1993).
Influence of Biology and Genetics There is also direct evidence, however, that
nongenius-level mathematical ability is more
No one is born a scientist, but some are born strongly related in monozygotic than in dizy-
with talents and temperaments that form the gotic twins, and is therefore at least partially
foundation for doing science. In some children, genetically determined (Bouchard & McGue,
these talents and aptitudes are manifested very 1981; Husen, 1960; Loehlin & Nichols, 1976;
clearly and very precociously. Scarr & Saltzman, 1982; Vandenberg, 1988).
Among the more innate aptitudes is mathemat- More specifically, these studies have found that
FEIST AND GORMAN
heritability estimates [i.e., double the difference either in the late 30s or early 40s and then drops
between monozygotic and dizygotic correla- off more gradually than it rose.
tions: h2 = 2(rmz - r^)], clearly implicate a However, we must point out that age only
genetic influence in mathematical ability. accounts for a relatively small percentage of the
Whether variability in math is also attributable overall variability in productivity (Bayer &
to other physiological, neurochemical, or ana- Dutton, 1977; S. Cole, 1979; Homer et al.,
tomical differences is more debatable, even if 1986). The work of Homer et al. (1986)
the consensus is that these factors do play some illustrates this point. They sampled over 1,000
role in individual differences in math ability (see male research psychologists from four different
Benbow, 1988, and the resulting commentaries birth cohorts, and found a curvilinear relation-
in Behavioral and Brain Sciences for discussion ship between age and productivity, with the peak
occurring in the early 40s. In this sample,

of potential causes of mathematical ability).

however age accounted for 6.5% of the overall
variance in publication rate. In short, it is clear
that other individual-difference and social fac-
Age and Productivity tors (such as early levels of productivity,
rewards and honors, and institutional support)
How does productivity change with age? have at least as strong if not stronger of a
One of the oldest of the developmental psychol- relationship with productivity (S. Cole, 1979;
ogy of science questions concerns whether age Zuckerman, 1977; Zuckerman & Merton, 1972).
affects level of productivity. The question is
unique not simply because it has been asked for If the description of the relationship between
such a long time, but because its answer is now age and productivity is relatively clear and
rather consensually agreed on. There is a agreed on, its explanation is not. Little theoreti-
relationship between age and productivity in cal attention has been devoted to the topic. The
science (and other professions) and it is an few attempts at explanation can be divided into
inverted U (Bayer & Dutton, 1977; S. Cole, two general categories: extrinsic versus intrinsic
1979; Dennis, 1956; Diamond, 1986; Homer, factors (Simonton, 1988b). The primary candi-
Rushton, & Vernon, 1986; Lehman, 1953,1960, dates for extrinsic theories concern decline in
1962, 1966; Over, 1982, 1989; Simonton, 1984, physical health, increase in family and adminis-
1988a, 1988b, 1989, 1991, 1992a; Zuckerman, trative obligations, and unfavorable work condi-
1977). Furthermore, once controls are made for tions, whereas the intrinsic factors are con-
different ways of operationalizing output, the cerned with changes in motivation, experience,
curve peaks at the same age (early 40s) for intelligence, and creativity.
quality and quantity of productivity. However, it Empirical evidence does at least partially
does peak somewhat differently for various support the extrinsic theories (Hargens, McCann
disciplines (earlier in math and physics, later in & Reskin, 1978; Roe, 1972; Simonton, 1977a).
biology and geology). This is not to say that Indeed, sociologists have long argued that an
the topic of age and productivity has been extrinsic factor (reward) plays an important role
without controversy. On the contrary; it has been in maintaining high levels of productivity in
replete with controversy from its inception. In some and discouraging it in others (J. Cole & S.
particular, Lehman's seminal work has been the Cole, 1973; S. Cole, 1979; Merton, 1973;
object of frequent criticism and rebuttal (S. Zuckerman, 1977). In other words, those
Cole, 1979; Dennis, 1956, 1958; Homer et al., scientists who produce the most impactful
1986; Over, 1989; Zuckerman & Merton, 1972). works early in their careers and who are thereby
Granted some of these criticisms are valid and rewarded with tenured jobs at top departments,
justified, but once many of the controls are made financial support, and prestigious awards are the
that Lehman failed to make, the result is still an ones who are most likely to continue producing.
inverted-U relationship (Simonton, 1988b). The The main problem with this theory is that it
peak may be a little flatter and it may occur a cannot explain the single-peak curvilinear rela-
little later, but basically every study conducted tionship between age and productivity in all
on the relationship between age and productivity scientists, not only the most precociously
has shown a curvilinear relationship that peaks productive.
However, little longitudinal research has been
conducted on the intrinsic theories, namely tial equations, with the peak occurring roughly
developmental changes in motivation, intelli- 20 years into one's career and thereafter slowly
gence, and creativity across the lifespan. The declining (Simonton, 1984,1988a, 1989).
sparse empirical work conducted on change in Does producing works early predict later
intelligence across adulthood, however, points levels of productivity? Again, enough work
to a rather late and small decline (Schaie, 1984), has been conducted on this question to provide a
which suggests that age-related declines in rather consensual answer: yes, early levels of
productivity may not be a result of a drop in high productivity do regularly foreshadow
intelligence. The more likely intrinsic candidate, continued levels of high productivity across
namely motivational decline, has not received one's lifetime (S. Cole, 1979; Dennis, 1954,
much empirical attention, but one longitudinal 1966; Helson & Crutchfield, 1970; Homer et al.,
study has reported a decline in drive in scientists 1986; Lehman, 1953; Over, 1982; Reskin, 1977;
with age: "from the standpoint of satisfaction Roe, 1965; Simonton, 1988b, 1991, 1992a).
there is some diminution of the involvement in Those who are prolific early in their careers also
work—some of the gratifications are beginning tend to continue to be productive for the longest
to pall and some of the fire, drive, and curiosity periods of time. For example, Horner et al.
is gone" (Eiduson, 1974, p. 408). Theoretically, (1986) reported that the most prolific group of
however, motivational decline has received scientists outpublished medium and low publish-
attention for over 100 years (Beard, 1874). ers by more than 2 to 1 in the 25-34 age period,
Beard argued that productivity is a function of and they maintained about a paper-per-year
changes in motivation (enthusiasm) and experi- advantage over both groups during each 10 year
ence. The young are more enthusiastic and the period until the mid 60 to 70 age period. At this
old are more experienced, and both enthusiasm age all three groups dropped to approximately a
and experience are linear functions of age half a paper per year, but the precocious group
(enthusiasm negative and experience positive). still outproduced the other two groups.
Creative achievement is a result of the balance As is the case with productivity in general,
between youthful enthusiasm and the experi- sociologists tend to explain this phenomenon in
ence of old age, and hence, productivity peaks terms of the cumulative advantage or the
when these two intrinsic processes overlap (i.e., Matthew effect (S. Cole, 1979; Merton, 1973;
in the late 30s or early 40s). Zuckerman & Merton, 1972): Those who
Simonton has developed a more complex publish frequently early in their careers and are
theoretical model that attempts to predict and therefore rewarded by their peers continue to
explain the age-productivity relationship by garner more and more of their share of the
focusing on intrinsic factors, namely cognitive resources and continue to outproduce their peers
components (Simonton, 1984, 1988a, 1988b, because of the ever increasing supply of
1989, 1991). This model is based on his financial and social support. Productivity data
chance-configuration theory and consists of a are inherently positively skewed with one tenth
few key assumptions: first, each creator starts off of the scientists producing roughly one half of
with a set amount of creative potential (number all of the works (Lotka, 1926; Price, 1963). The
of contributions made over a normal, unre- rich get richer and the poor get poorer!
stricted life span). Second, the actualization of Furthermore, there is some evidence that
creative potential can be broken down into two quantity of publication matters more than
components: ideation and elaboration. Ideation quality of publication when predicting who will
is the rate at which potential ideas are expressed, receive the most peer recognition and presti-
whereas elaboration is the rate at which ideas gious honors—that is, who will become the
are put into concrete, public form. So as each most eminent (Feist, 1997).
creator produces a new work she or he "uses Compared with older scientists, do younger
up" some creative potential. The rate at which a scientists produce a disproportionate number of
creator actualizes potential and produces works high-quality works? Lehman (1953, 1960,
is a direct function of the two cognitive 1966) suggested that younger scientists (below
transformations, ideation and elaboration. To age 40) produce most of the highly cited and
graphically model this relationship, Simonton impactful works. However, Lehman's data did
has developed one of his better known differen- not take absolute number of scientists at each
FEIST AND GORMAN
age period into account, and therefore may be hand, as with productivity, age accounted for
biased towards the young simply because mere less than 10% of the variance in theory
are more young scientists. Over (1989) exam- acceptance, and can only provide weak support
ined whether older scientists were more likely to for Planck's principle. Indeed, Sulloway (1996)
produce works of lower quality than younger has recently published an exhaustive historical
scientists. He found that although it is true that a analysis of theory acceptance in science and
disproportionate number of high-quality works concluded that birth-order accounted for more
come from scientists less than 10 years post- variance than any other single variable. Further-
PhD, it is equally true that a disproportionate more, Messeri (1988) studied age differences in
number of low-quality works come from this acceptance of plate tectonics after the discovery
age group of scientists. In other words, more of sea-floor spreading by Hess and Dietz in the
high quality works are being produced by early 1960s. During the period immediately
younger scientists not because of age but following publication of this new idea, older
because of the high number of young scientists. scientists were significantly more likely to adopt
The same holds true once longitudinal rather plate tectonics than younger ones, exactly the
than cross-sectional data are examined (S. Cole, reverse of what one would expect if Planck's
1979). Longitudinal data are important because principle were true. Later, after substantial
they do not confound age and cohort effects the confirmatory data had been disseminated, age
way cross-sectional data do. no longer played a role in theory acceptance.
Are older scientists more resistant to scientific Finally, Levin, Stephen, and Walker (1995) used
revolutions than younger ones? Max Planck's logit regression on the data collected by Hull et
experience with resistance to his novel ideas al. (1978) and divided it into the same time
gave him "an opportunity to learn a new fact—a periods used by Messeri. They found no
remarkable one, in my opinion: A new scientific significant relationship between age and theory
truth does not triumph by convincing its acceptance and concluded that "No researcher
opponents and making them see the light, but to date has found substantial effects of age on
rather because its opponents eventually die, and the acceptance of new ideas" (Levin et al., 1995,
a new generation grows up that is familiar with p. 281).
it" (as quoted in Barber, 1961, p. 597). This
observation of Planck's, which has come to be Gender and Science
called Planck's principle, fits well with T.
Kuhn's notion of paradigm shift, in which the One of the more contentious and polemical
old and new paradigms are so different that they domains of the psychology of science concerns
are incommensurable. Similarly, toward the end the role that gender plays in science in general
of the On the Origin of Species, Darwin noted (cf. Keller, 1985) and in scientific and mathemati-
that cal ability and achievement in particular. The
topics of gender and science and gender
A few naturalists, endowed with much flexibility of
mind, and who have already begun to doubt on the
differences in scientific achievement could in
immutability of species, may be influenced by this and of themselves be the focus of a review
volume; but I look with confidence to the future, to article, and we leave the more exhaustive review
young and rising naturalists, who will be able to view of this literature to others. There are three
both sides of the question with impartiality, (cited in questions that we believe have accumulated
Hull, Tessner, & Diamond, 1978, p. 718)
enough literature to warrant our attention, and
Darwin's primary defender, T. H. Huxley, went each concerns gender differences: first in
further, arguing that men of science ought to be mathematical ability, second in productivity, and
strangled on their 60th birthday, lest they retard third in quality of scientific work.
scientific progress (Hull et al., 1978). Are there gender differences in mathematical
Hull et al. (1978) compared the ages of ability ? One of the more consistent and robust
scientists who accepted and rejected the idea findings in the gender-difference literature in-
that species evolved in the 10 years after volves mathematical ability, with male partici-
publication of Origin; the rejecters, on the pants scoring higher than female participants
average, were 10 years older—a statistically and throughout the distribution of scores (Astin,
practically significant difference. On the other 1975; Backman, 1972; Benbow, 1988; Benbow
& Stanley, 1980,1983; Deaux, 1985; Fischbein, explanations: attitudes toward math, perceived
1990; Fox, 1976; Holden, 1987; Keating, 1974; usefulness of math, confidence and self-efficacy,
Maccoby & Jacklin, 1974; Moore & Smith, encouragement from parents and teachers,
1987; Stanley, 1988). Both longitudinal and sex-typing, differential course taking, and career
cohort data over the last 20 years suggest that and achievement motivation. Benbow found
the gender difference is remaining constant at that some of these environmental influences do
around .50 standard deviation in favor of males. distinguish male and female participants. For
However, as Maccoby and Jacklin (1974) first example, girls do like math less, find it less
reported, there are a couple of qualifications to useful for their future goals, and have less
this generalization. First, there is little to no confidence in their ability than boys. Further-
gender difference before adolescence and sec- more, mathematics is somewhat sex-typed as a
ond, at least up through early adolescence, girls "masculine" enterprise, parents and teachers are
achieve higher grades than boys in math classes. more encouraging of male than female math-
One of the largest studies ever of mathemati- ematical achievement, differences in math
cal ability was started by Julian Stanley in 1971 courses do not explain aptitude differences, and
at Johns Hopkins University and is titled the finally male career motivation is more indepen-
Study of Mathematically Precocious Youth. In dent of parent or teacher support than female.
the tradition of Terman (1925) and Cox (1926), However, these findings do not directly
Stanley and his colleagues studied large but address the origin of the difference. As Eysenck
select samples of mathematically precocious (1988) pointed out, the situational findings
young people, who they defined as scoring at or could result from either genetic or environmen-
above 700 on the SAT-M before age 13 tal origins. To more directly address the
(Stanley, 1988; Stanley, Keating, & Fox, 1974). biological explanations, Benbow (1988) offered
That this is an extremely selective criterion is four possibilities: hemispheric laterality, aller-
beyond dispute: only 4% of male college-bound gies, hormonal influences, and myopia. For
high school seniors and fewer than 1 % of female instance, based on a high incidence of left-
college-bound high school seniors score 700 or handedness in the mathematically precocious
higher on SAT, and Stanley's sample consisted and in particular the precocious boys, and the
of preadolescents. As both Stanley and Benbow greater bilateral or diffuse cognitive functioning
have reported, one of the biggest surprises in of left-handed individuals, Benbow concluded
collecting these data, however, was the large and that bilateral or a strong right hemispheric
consistent gender difference among the extreme functioning may implicate mathematical ability.
scores—ultimately reaching as high as a 12 to 1 Furthermore, prenatal exposure to testosterone
ratio in favor of male participants. Furthermore, has been postulated to influence handedness and
Benbow's (1988) target article in Behavioral immune disorders (cf. Geschwind & Behan,
and Brain Sciences was commented on by more 1982), and therefore could be an indirect
than 40 experts, and although virtually none of influence on mathematical ability. To quote
the commentaries took issue with whether a Benbow
gender difference exists, there was little agree-
In sum, the above physiological correlates, especially
ment concerning the potential causes of this the possibility of prenatal testosterone exposure, lend
gender difference. Indeed, the gender difference credence to the view that sex differences in extremely
raises at least two very important questions: high mathematical reasoning ability may be, in part,
first, how can it be explained, and second do physiologically determined (Benbow & Stanley, 1980).
Of course, some of the above discussion on physiologi-
math scores in adolescence actually predict cal correlates is speculative. (1988, p. 182)
ultimate science and math achievement?
Possible environmental and biological causes Suffice it to say that the physiological explana-
for the gender difference in math. Unfortu- tions were the focus of most of the criticism in
nately, we can only give the briefest overview of the commentaries. However, many criticisms
this contentious area, and we refer the interested did not take issue with the fact that biological
reader to the target article by Benbow (1988) explanations may play a role, but rather that
and its commentaries for more detail. More their mechanisms are more complex than, and
specifically, the author reviewed the evidence the evidence is not as solid as, Benbow's
for seven of the more common environmental presentation.
10 FEIST AND GORMAN
Predictive validity of mathematical precocity of aptitude tests is better understood once one
and early scientific achievement. Even with a takes into account the small relationship be-
consistent gender difference, we must still tween intelligence and creativity (Barron &
address the predictive validity question, namely Harrington, 1981; Getzels, 1987; MacKinnon,
to what extent does mathematical precocity 1978; Sternberg, 1986; Wallach, 1970), or in
translate into becoming a scientist and having an Guilford's terms convergent from divergent
impactful scientific career (Farmer, 1988; Stern- thinking (Guilford, 1987). To solve quickly
berg, 1988b)? Stanley (1988) argued for the multiple-choice problems that have known
predictive value of extreme mathematical precoc- solutions involves convergent or analytical
ity: thinking skills, whereas to solve creatively
open-ended problems that have no known
these young students seem to have the potential to solutions involves divergent or intuitive think-
become the nation's superstars in pure and applied

mathematics, computer science, electrical engineering, ing skills (Guilford, 1959, 1987; Simonton,
physics, and other fields that depend heavily on great 1988a, 1989; Sternberg, 1986). Precocious
quantitative aptitude. Quite a few of the 292 [who youth are very intelligent, but not necessarily
scored £700 on the SAT-M] appear well on the way very creative.
toward excellence in such fields, (p. 206)
A general conclusion regarding gender differ-
Benbow and her colleagues have presented data ences in math is that boys and men do
showing that precocious ability predicts achieve- consistently score higher on mathematical apti-
ment in high school (Benbow & Minor, 1986; tude tests but not on achievement tests or grades.
Benbow & Stanley, 1982) and in college However, the explanation and theoretical ac-
(Lubinski & Benbow, 1994). Furthermore, counts for why these differences exist are not yet
longitudinal research has shown that those who settled, and currently we can only offer the trite
went into math or science careers scored in the and general conclusion that both biology and
90th percentile on math achievement tests in environment account for some of the variance.
high school (Wise, Steel, & MacDonald, 1979). Future empirical work must focus on testing
But as Farmer (1988) pointed out, only 42% of theoretical models and causal factors through
the male and 22% of the female extremely quasiexperimental design and multivariate and
precocious students went on to choose science latent variable analyses if more definitive
or math graduate programs (cf. Benbow, 1988; answers are to be provided.
Benbow & Lubinski, 1993). In short, only 25%
Do male and female scientists produce works
of the extremely gifted math sample continue in
at different rates ? Comparing publication rates
science and math through graduate school, and
of men and women has consistently shown that
even fewer are retained in science and math
men produce more works than women (J. Cole,
careers. The retention rate is somewhat higher
1979, 1987; J. Cole & S. Cole, 1973; Guyer &
for those who demonstrate early scientific
Fidell, 1973; Helmreich, Spence, Beane, Lucker,
achievement. Subotnik and Steiner (1992), for
& Matthews, 1980; Long, 1992; Pasewark,
instance, reported that by their mid 20s, 81% of
Fitzgerald, & Sawyer, 1975; Zuckerman &
the male and 66% of the female Westinghouse
Cole, 1975). This gender difference appears to
Science finalists were still on science-training or
hold for total number of publications and yearly
science-career tracks.
average (J. Cole, 1987). However, there is some
However, if one chooses a more real-world contradictory evidence regarding whether this
valid outcome criterion, such as actual creative gender difference increases or decreases across
achievement in math and science, very few of the course of one's career. J. Cole (1987)
even the extremely talented youth go on to have reported that the gender gap on productivity
truly influential careers (Simonton, 1988a; increases, whereas Long (1992) reported that it
Sternberg, 1988b). The evidence shows that decreases over the course of one's career.
neither grades nor aptitude tests do well at One obvious question therefore that begs to
predicting creative achievement in scientific be addressed is how to explain the gender
careers (Barron & Harrington, 1981; Gough, difference in total and yearly average publica-
1976; Guilford, 1959; Hudson, 1958; MacKin- tion rates. As with age and productivity,
non, 1960; Simonton, 1988a; Sternberg, 1988a; explanations are more contentious and less
Taylor, 1963). Such a lack of predictive validity consensual than the description of the phenom-
PSYCHOLOGY OF SCIENCE 11
enon. Differences in family obligations, prestige Development of Scientific Reasoning

of institution, rank of position, training, and
motivation each has been investigated, but with In addition to precocity, age, and gender
negative or inconsistent results. The intuitively another contribution of the developmental psy-
appealing answer that women are hindered by chology of science involves examining the
multiple roles of scientist, wife, and mother, and cognitive processes that children of different
are relegated to marginal departments seems not ages use when trying to solve scientific prob-
to have empirical support (J. Cole, 1979, 1987; lems. Jean Piaget was the great pioneer in this
S. Cole & Zuckerman, 1987). In fact, married area. He felt that, in the case of scientific
women tend to slightly outproduce single thinking, ontogeny recapitulated phylogeny: the
women, and women with one or two children child's development of scientific thought reca-
tend to outproduce women with no or more than pitulated the history of science (Piaget & Garcia,
three children (J. Cole, 1979, 1987; S. Cole & 1989). Children, in this view, begin with a kind
Zuckerman, 1987). J. Cole and S. Cole (1973) of Aristotelian view of how the world operates,
also presented evidence that gender differences and—if they reach the highest level of formal
in productivity cannot be explained by differ- operations—end up with a Newtonian or
ences of institution (college vs. university) or perhaps even an Einsteinian view, one they have
prestige of department. When both of these internalized, not merely memorized (McClos-
variables are entered first in a regression key, 1983; Wiser & Carey, 1983). Piaget also
equation and thereby held constant, the relation- made an enormous methodological contribu-
ship between gender and productivity still tion; he inspired researchers to question children
persists. closely, and pose problems for them that would
reveal not only what they knew, but how they
If prestige of institution and department do knew it, and how their knowledge could be
not moderate the relationship between gender changed by additional experimentation.
and productivity, then perhaps other variables
One fundamental question involves whether
do. The most obvious candidate for a moderat-
children's thought processes are categorically
ing variable would be rank of academic position.
similar or categorically different from those
Perhaps men outproduce women because they
used by adults (Fay, Klahr, & Dunbar, 1990;
are at higher ranks. Guyer and Fidell (1973),
Klahr, Fay, & Dunbar, 1993; D. Kuhn, 1989; D.
however, found gender differences in productiv-
Kuhn, Amstel, & O'Loughlin, 1988; Siegler &
ity even when comparing male and female
Liebert, 1975). The heuristic and decision-
professors at the associate and full ranks, but not
making movement in cognitive psychology has
at the assistant rank. Moreover, J. Cole (1987)
contributed to the widespread dissemination of
reported that promotion to higher academic rank
the intuitive scientist metaphor. In this view,
is one of the few remaining areas of sex
children and nonscientist adults construct cogni-
discrimination in science. Even holding other
tive models, evaluate evidence, and modify their
variables constant, such as quantity and quality
conceptualizations of how the world works in a
of publication, and career interruptions, women
similar but less developed manner to scientists
are still less likely to be promoted than men,
(Brewer & Samarapungavan, 1991; Karmiloff-
which could indeed be a factor in their lower
Smith, 1988). D. Kuhn (1989) and Klahr et al.
productivity.
(1993) argued that although there is some
Is there a gender difference in quality of validity to this metaphor, especially in under-
works produced? With citation counts as the standing of the world, it is quite misleading and
measure of quality and impact, some researchers inaccurate when applied to the cognitive pro-
have found that men receive more citations than cesses used by children, novice adults, and
women (J. Cole, 1987; J. Cole & S. Cole, 1973; scientists. D. Kuhn, for example, reviewed
Reskin, 1977). But this may be an artifact of much of the relevant literature and concluded
greater number of published papers by men (S. that in terms of thinking like a scientist (i.e.,
Cole & Zuckerman, 1987). In fact, once number coordinating theory and evidence) neither nov-
of publications is held constant, women produce ice adult nor child is capable of such systematic
works of greater impact than men (Long, 1992; thinking. When confronted with disconfirming
Sonnert, 1995). evidence, children often unknowingly distort,
12 FEIST AND GORMAN
selectively make use of the evidence, or tion of creative activity in science and art.
unconsciously adjust theory to fit with the Furthermore, Feist (1991) reported that 65% of
evidence. Adult scientists, however, make a the elite biological and natural scientists in his
clear distinction between theory and evidence sample reported having a significant mentor in
and therefore can systematically and con- high school, and 80% reported having one in
sciously modify and manipulate the evidence graduate school. In high school, mentors tended
one piece at a time to see what effect it has and to be either a teacher (29%) or a parent (26%),
whether the theory needs to be modified. whereas in graduate school they were over-
D. Kuhn's conclusions are in line with the whelmingly one's PhD advisor (56%) or another
work on searching in two spaces (hypothesis professor (20%). A question raised by these
and experiment) by Klahr and his colleagues, in percentages, however, is how they compare to
that they both conclude that children fail to other professions. Whether these figures are
distinguish theory from evidence. Klahr, Fay, unique to the sciences remains to be seen.
and Dunbar (1993) performed a fairly sophisti- Werts and Watley (1972) demonstrated that
cated experiment in order to test whether there the family environment can exert a strong
are developmental differences in scientific influence on choosing science as a career. They
problem solving heuristics. They tested four reported that college students who won awards
different groups who varied on age and scien- and were high achievers in science had fathers
tific-technological skill: 3rd graders, 6th grad- who were scientists. Furthermore, a consistent
ers, community college students with little and robust finding from the literature on father's
technical training, and college students with education and occupation is that scientists
technical training. Results showed rather clearly overwhelmingly come from families of profes-
that under some circumstances children can sional occupations and higher education (Berry,
perform cognitive processes that are similar to 1981; Chambers, 1964; Feist, 1991; Helson &
adults, but under other circumstances they Crutchfield, 1970; Roe, 1952a; Zuckerman,
cannot. When the actual hypotheses are plau- 1977). Either directly or indirectly, having
sible or when the experimental alternatives are well-educated parents familiar with and inter-
few in number children perform similarly to ested in science is predictive of an interest in
adults. But when the actual hypotheses are science.
implausible or the alternatives are not few in
number, adults' performance is categorically Developing an early interest in science has
superior to children's. Children were not able to importance to the extent that it translates into
consider two alternative hypotheses when the becoming and staying a scientist. Subotnik and
actual hypothesis was implausible, and they her colleagues have collected longitudinal reten-
stuck with their original plausible but incorrect tion data on a sample from finalists to the
hypothesis. Adults, on the other hand, were able prestigious Westinghouse Science Talent Search
to search for solutions in two spaces simulta- (Subotnik, Duschl, & Selmon, 1993; Subotnik
neously, namely hypothesis and experiment. & Steiner, 1992). Five years after being
Westinghouse finalists, 94 students took part in a
follow-up study that investigated why some
Mentorships and Training stayed in science and others did not. Even with
this elite and prestigious group of gifted science
What role do family members or teachers play students, almost 30% were not pursuing scien-
in promoting and retaining scientific interests? tific careers just 5 years later. Gender (being
It has long been assumed that family and female), lack of enthusiastic and motivating
school influences are critical to the development high school and college teachers and mentors,
of scientific interests. But what does the and lack of financial support were the strongest
empirical literature say about these social discriminators between the two groups. Similar
influences? Eiduson (1962) reported that roughly findings on the gender difference in retention of
half of her participants said that some older highly precocious math students have been
person was important in their developing and reported by Benbow and her colleagues (Ben-
maintaining an interest in science. John-Steiner bow, 1988; Benbow & Lubinski, 1993; Lubinski
(1985) eloquently described the importance of & Benbow, 1994).
apprenticeships and mentorships in the stimula- When students complain about lack of
enthusiastic teaching they are no doubt referring first born (e.g., J. Cattell & Brimhall, 1921;
to the fact that science often is taught in a Clark & Rice, 1982; Eiduson, 1962; Galton,
manner that emphasizes its "factual" basis and 1874; Roe, 1952a). However, an even more
ignores the process of asking questions and interesting question arises concerning whether
discovering solutions. To the extent that only the the same holds true for the most creative and
rational, factual, and objective side of science is eminent scientists. Helson and Crutchfield
taught, children develop the mistaken belief that (1970) found that creative scientists were more
science is "boring and dry." Feist (1991) likely to be first born than less creative
reported among eminent scientists a consistent scientists. The largest and most ambitious study
complaint that the "fun" part of doing science to date of birth-order and scientific eminence
was completely absent in primary education of (Sulloway, 1996) found a curvilinear relation-
science. Curiosity is a sine qua non of science. ship, with first and last born scientists being the
Consistently asking, "How does this work? Or, most eminent (cf. Feist, 1991).
what happens if I do this?" appears to be
instrumental to the development of scientific
interests, and if curiosity is squelched there is Religious Background
little hope that an otherwise intelligent child will Another way family may influence scientific
want to pursue a career in science. development is through its dominant religious
Does being trained by an eminent scientist orientation. Many researchers have reported that
predict obtained eminence ? As reported above, a disproportionate number of eminent and
having a strong mentor in high school and creative scientists come from Protestant or
college does predict staying on and pursuing a Jewish families compared with Catholic back-
scientific career (Subotnik, Duschl, & Selmon, grounds (Chambers, 1964; Datta, 1967; Feist,
1993; Subotnik & Steiner, 1992). Having an 1991; Helson & Crutchfield, 1970; Roe, 1952a;
eminent mentor also appears to be a contributing Zuckerman, 1977). For example, whereas only 2
factor in obtaining eminence (John-Steiner, to 3% of the American population comes from
1985; Simonton, 1992b; Zuckerman, 1977). Jewish backgrounds, the percentage of eminent
This finding has been most clearly demonstrated and elite scientists from Jewish backgrounds
in Harriet Zuckerman's work with Nobel ranges from 9% (Roe, 1952a) to 38% (Helson &
laureates. One of her strongest findings con- Crutchfield, 1970). In fact, among the most
cerned the cumulative advantage effect of those creative and elite groups of scientists most
young scientists who train under the scientific estimates suggest that 20 to 30% come from
elite (i.e., Nobel prize winners). They produce Jewish families (Chambers, 1964; Datta, 1967;
more at an early stage in their careers, are more Feist, 1991; Zuckerman, 1977). Religious back-
likely to produce works of high impact, and are ground, however, does not tease apart variability
more likely to win the Nobel prize themselves due to religious orientation, culture, race, or
than those who do not train under laureates even genetic influence. Moreover, we must
(Zuckerman, 1977). As Zuckerman and others make clear that these data refer to the religious
have argued, however, the causal direction of faith of one's family background and upbring-
this influence probably goes both ways: the best ing, not one's current behavior. Scientists in
young scientists are chosen by the best scien- general, and eminent scientists in particular, are
tists, which in turn feeds into the cycle of conspicuous in their rejection of organized
cumulative advantage. Simonton (1992b) also religion. The few studies that have asked
reported that American Psychological Associa- scientists about their current religious practices
tion (APA) presidents were quite likely to have have reported an almost complete absence of
been mentored by an eminent psychologist (i.e., current religious faith (Chambers, 1965; Feist,
they had an entry in a biographical dictionary). 1991; Roe, 1952a; Terman, 1954,1955).
Birth Order Conclusions From Developmental

Psychology of Science
Beginning with Galton in the 1870s, a number
of researchers have reported that compared to Among the conclusions that can be drawn
nonscientists, scientists are disproportionately about the developmental psychology of science
14 FEIST AND GORMAN
literature (see Table 1) are: (a) some people are Herbert Simon), and Wasonian (as in Peter
born with extreme mathematical talent and Wason) traditions. None of these traditions,
genius; (b) there is a curvilinear relationship however, is mutually exclusive, and in fact
between age and productivity, with the peak Simon's approach to psychology of science is
generally occurring in one's late 30s or early explicitly identified with Kuhn (Bradshaw,
40s; (c) early productivity does predict later Langley, & Simon, 1983), and Wason's with
productivity; (d) young scientists do not neces- Popper as well as Piaget (Gorman, 1992). De
sarily produce a disproportionate number of Mey (1992) also shows how the Kuhnian and
high-quality works; (e) older scientists are not Piagetian traditions can be tightly linked.
more resistant to accepting new theories com-
Therefore, we attempt to use methodological
pared to younger scientists; (f) there are gender
differences in mathematical ability, and the rather than theoretical categories to sort studies,
etiology of these differences remains unclear, focusing more on results than on the epistemolo-
but biological and social factors each appear to gies that lie behind them. In particular, we
play a role; (g) men publish more than women, divide the literature along two lines, namely the
but whether their work is of higher quality is nature of the task and the type of participant.
unclear, and differences in marital status, family The nature of the task involves essentially two
size, prestige of department and institution do types of problems: either abstract tasks that
not account for the gender differences in simulate scientific reasoning or actual scientific
quantity, but differences in academic rank may; problems. This category is not a dichotomous
(h) children's thought processes when solving variable. Some of the tasks used to simulate
scientific problems are categorically different scientific reasoning are modeled closely after sci-
from novice adults or scientists in that they are entific reasoning, and some of the actual scientific
not able to coordinate their thinking to sim- problems resemble those encountered in textbooks
ultaneously consider theory-hypothesis and evi- rather than in the laboratory. Moreover, studies have
dence-experiment; (i) enthusiastic teachers and used two categories of participants—either novices
families who value education are critical to the or experts. The novice category of participant is
development of an interest in science; (j) having mostly made up of college students of a variety of
an eminent mentor predicts scientific eminence; backgrounds, some of whom may have taken a few
(k) effects of birth order on choosing science as science courses, but none of whom are practitioners.
a career and producing high-quality works in Experts, on the other hand, are defined as practicing
science are unclear and inconsistent; and finally
scientists of varying abilities. Again, this is not a
(1) coming from a Jewish background is related
dichotomous variable. Novices can range from
to producing high-quality scientific work.
children to graduate students, and practitioners from
scientists at the beginning of their careers to eminent
veterans. This way of organizing the literature and
Cognitive Psychology of Science
its basic findings on cognitive psychology of
Instead of reviewing the entire history of science are summarized in Table 2.
attempts by psychologists to understand scien-
tific thinking (see Campbell, 1989, for a good
overview of psychological epistemologies), we Simulated Scientific Tasks: Novices
look at the current state of die field (primarily
from the late 1970s to the present), cite studies Confirmation bias. The literature using
that illustrate the major findings, and explicate simulated tasks with novices primarily has
consistent trends and general principles. To examined how (quasi-)scientific hypotheses are
organize this review, we could attempt to tested. Moreover, almost all of this literature has
categorize research based on its epistemological focused on confirmation bias or some variation
foundations, as Campbell, Tweney, and others thereof. The impetus for this line of research
have done. However, it is a virtual guarantee was Karl Popper's (1959) assertion that science
that no two psychologists will agree on these should, and the best science does, progress by
epistemological roots. For example, Tweney falsifying hypotheses, not by proving them
(1994; Tweney & Chitwood, 1995) distin- right. Wason (1960) decided to find out whether
guished Kuhnian, Piagetian, Simonian (as in novices could falsify by asking them to
determine the rule governing a sequence of the participant needs to be encouraged to seek
number triplets, given that the triplet 2-4-6 was negative evidence (Gorman, 1992).
an instance of the rule. Participants proposed Confirm early, disconfirm late. A group at
additional triples, and the experimenter told Bowling Green State University (Mynatt,
them whether each fit the rule. When they felt Doherty, & Tweney, 1977, 1978) developed an
they were ready, participants would tell the artificial universe task that bore more resem-
experimenter what they thought the rule was, blance to science than abstract problems like the
and the experimenter would tell them whether it 2—4—6. Participants spent about 10 hours on the
was the rule he had in mind. If not, they could most complex of these tasks, and none of them
continue to propose triples and make guesses. discovered the rule. The ones that made the most
Participants typically proposed triples like progress exhibited a kind of confirmation bias,
6-8-10 and 10-12-14 and guessed the rule was but with one important qualification: confirma-
something like "numbers must go up by twos." tion bias is most effectively used only early in
In fact it was "all three numbers in the triple the hypothesis testing process. Mynatt, Doherty,
must ascend in order of magnitude." Wason and Tweney concluded that confirmation was an
took this performance as evidence of a confirma- effective heuristic early in the inference process;
tion bias on the part of participants because they once a participant or scientist had discovered
found a rule sufficient to explain the first pattern and verified a pattern, then she could switch to
that generated positive instances. the search for disconfirmatory evidence. This
heuristic combination of confirmation and dis-
Novices can be trained to seek negative
confirmation also worked on abstract problems
evidence. Gorman and his colleagues (Gor-
like the 2-4-6 task, but its value became most
man, 1986; Gorman & Gorman, 1984; Gorman,
apparent on tasks that more closely simulate
Stafford, & Gorman, 1987) found that instruc-
scientific problems. Initially the best way to
tions to falsify significantly improved perfor-
follow this confirmatory heuristic would be to
mance on the 2-4-6 and similar tasks. So it
conduct positive tests. One of these positive
appeared that confirmation was a bias that could
tests might lead to disconfirmation, but the
be combated with education. However, Klay-
overall goal would be to discover a pattern and
man and Ha (1987) argued that Wason, Gorman,
verify that it was consistent enough to form the
and others had confused positive and negative
basis for a hypothesis.
test heuristics with confirmation and disconfir-
mation. On the 2-4-6 task, a positive test Computational simulation of abstract prob-
heuristic involves trying to get triples right if lems. We use the term simulation to refer to
one believes one's hypothesis is correct. If, for the programs in this section because their
example, a participant proposes 2, 6, 4, expect- primary goal is to model, understand, and
ing it to be right if her hypothesis is correct, she improve human scientific problem solving.
is following a positive test heuristic (Klayman, Naturally, this has implications for machine-
personal communication, May 14,1997). So she learning, whose goal is finding ways of making
might think the rule was "three even numbers," machines solve problems, but an effective
and if she obtains a "no" response she has problem-solving heuristic for a machine may be
obtained disconfirmation using a positive test. very different than one for a human being.
If, on the other hand, her hypothesis is "go up by Herbert Simon and a group of colleagues at
twos" and she proposes 2-6-4, this would be a Carnegie-Mellon (Langley, Simon, Bradshaw,
negative test because if the hypothesis is correct & Zykow, 1987) developed a series of computer
the triple will be incorrect and confirm rather programs designed to emulate scientific discov-
than disconfirm her hypothesis. Therefore, eries, for example Kepler's law. The simplest of
negative tests are not to be equated with these, called BACON, was given columns of
disconfirmation. Klayman and Ha argued that numbers and asked to find a relationship, using
seeking positive tests is a good all-purpose heuristics like "if the terms in two adjacent
heuristic and in most situations is even more columns increase together, compute their ratio."
likely than a negative test heuristic to produce The relationship turned out to be the numerical
falsifications (see also Oaksford & Chater, equivalent of Kepler's third law.
1994). However, when the target participant's
rule is more general than her initial hypothesis, Qin and Simon (1990) gave the same task to
14 college students. Four of them were able to
16 FEIST AND GORMAN
Table 1
Summary of Topics and Findings in the Developmental Psychology of Science Literature
Topic Results Author(s)
Mathematical Truly outstanding mathematical genius appears to be Bell, 1937
ability inborn Kanigel, 1991
Wiener, 1953
Genetics accounts for some of the variability in math Husen, 1960
ability Loehlin & Nichols, 1976
Bouchard & McGue, 1981
Scarr & Saltzman, 1982
Vandenberg, 1988
Age and produc- There is a curvilinear relationship between age and pro- Lehman, 1953,1960,1962,1966
tivity ductivity, with the peak generally occurring in one's Dennis, 1954,1956
late 30s or early 40s Bayer & Dutton, 1977

Zuckerman, 1977
S. Cole, 1979
Over, 1982, 1989
Simonton, 1984,1988a, 1988b, 1989
Diamond, 1986
Homer et al., 1986
Early productivity does predict later productivity Lehman, 1953
Dennis, 1954
Roe, 1965
Helson & Crutchfield, 1970
Over, 1982
Simonton, 1988b, 1989
Age and quality Young scientists do not necessarily produce a dispropor- S. Cole, 1979
tionate number of high-quality works Over, 1989
Age and theory Older scientists are not more resistant to accepting new Barber, 1961
acceptance theories compared to younger scientists Hull et al., 1978
Messeri, 1988
Levin et al., 1995
Gender differ- There are gender differences in mathematical ability, but Backman, 1972
ences in the etiology of these differences remains unclear; Keating, 1974
math ability biological and social factors each appear to play a Maccoby & Jacklin, 1974
role Astin, 1975
Fox, 1976
Benbow & Stanley, 1980,1983
Deaux, 1985
Holden, 1987
Moore & Smith, 1987
Benbow, 1988
Stanley, 1988
Fischbein, 1990
Gender and pro- Men publish more than women; differences in marital J. Cole & S. Cole, 1973
ductivity status, family size, prestige of department and institu- Guyer & Fidell, 1973
tion do not account for the gender differences in Pasewark et al., 1975
quantity, but differences in academic rank may Zuckerman & J. Cole, 1975
J. Cole, 1979,1987
Helmreich et al., 1980
Long, 1992
Gender and Men are more frequently cited when quantity is not taken J. Cole & S. Cole, 1973
quality into account, but women are more cited when publi- Reskin, 1977
cation total is held constant J. Cole, 1979
Helmreich et al., 1980
Long, 1992
Sonnert, 1995
Development of Children's thought processes when solving scientific prob- Siegler & Liebert, 1975
scientific lems are categorically different from novice adults or D. Kuhn et al., 1988
reasoning scientists; they are not able to coordinate their D. Kuhn, 1989
thinking to consider simultaneously theory/hypoth- Fay et al., 1990
esis and evidence/experiment Klahretal., 1993
Table 1 (continued)
Topic Results Authors)
Influence of Enthusiastic teachers and families who value education are Roe, 1952a
parents, important to the development of an interest in science Eiduson, 1962
teachers, Chambers, 1964
and mentors Helson & Crutchfield, 1970
Werts&Watley, 1972
Zuckerman, 1977
Berry, 1981
John-Steiner, 1985
Feist, 1991
Subotnik & Steiner, 1992
Subotnik et al., 1993

Having an eminent mentor predicts scientific eminence Zuckerman, 1977

John-Steiner, 1985
Birth order and Scientists are more likely to befirstboras compared with Galton, 1874
scientific nonscientists Cattell & Brimhall, 1921
eminence Roe, 1952a
Eiduson, 1962
Clark & Rice, 1982
Creative and eminent scientists are most likely to be either Helson & Crutchfield, 1970
first or last bom Feist, 1991
Sulloway, 1996
Religious back- Coming from a Jewish background is related to producing Roe, 1952a
ground high-quality scientific work Terman, 1954
Chambers, 1964
Datta, 1967
Zuckerman, 1977
Feist, 1991
emulate the computer's discovery using heuris- change made it much easier for participants to
tics similar to BACON's. Although this appears explore the limits of their hypotheses, thereby
to be a task that models the reasoning of an facilitating discovery of the target rule. Simi-
actual scientific expert, it actually bears more larly, Farris and Revlin (1989a, 1989b) argued
resemblance to the 2-4-6 task. Instead of three that many participants who appeared to be
columns of numbers, Qin and Simon's partici- following a disconfirmatory strategy were actu-
pants were given two, and the numerical rule ally searching for positive instances of a
they were seeking happened to correspond to counterfactual hypothesis. In effect, participants
Kepler's Law. Interestingly, students trying to following a counterfactual strategy were search-
solve this problem relied more on visual ing for two complementary rules (see Oaksford
representations like scatter plots than the & Chater, 1994). This counterfactual heuristic
program, which could not use diagrammatic would be best applied after the initial positive
reasoning. (More recent computational simula- test phase when the scientist or participant has
tions have this capacity; see Cheng & Simon, discerned a pattern and formulated an initial
1995; Larkin & Simon, 1987.) hypothesis. Conceptualizing the problem as a
Searching for two rules rather than one. search for two complementary hypotheses
Searching for two complementary rules rather focuses participants on exploring the boundaries
than one appears to increase successful hypoth- of both, with the result that they conduct a more
esis testing on abstract tasks. Tweney, Doherty, thorough search of the problem space. Similarly,
and Mynatt, (1981), Gorman, Stafford, and Freedman (1995) argued that limits on working
Gorman (1987), and Wharton, Cheng, and memory may explain why novices tend to focus
Wickens (1993) altered Wason's (1960) task to oh one hypothesis at a time. Instructing
make it a search for two complementary rules participants to search for multiple hypotheses
rather than a single rule. They found that this (i.e., to use strong inference strategies) did in
18 FEIST AND GORMAN
Table 2
Summary of Topics and Findings in the Cognitive Psychology of Science Literature
Task Participants Topic Results Authors)
Simulated Novices Hypothesis testing Hypothesis testing is biased toward Wason, 1960
confirming evidence
Confirm early/disconfirm late heu- Mynatt, Doherty & Tweney, 1977,
ristic is most effective strategy 1978
Success in hypothesis testing is Tweney et al., 1981
increased by searching for two Gorman, Stafford, & Gorman, 1987
complementary rules rather than Wharton, Cheng & Wickens, 1993
College students form commonsense MeCloskey, 1983

representations of scientific
phenomena
Instructions to falsify improves per- Gorman & Gorman, 1984
formance on hypothesis testing
Confirmation bias is not synonymous Klayman & Ha, 1987
with positive-test heuristic
Successful hypothesis testing is Klahr&Dunbar, 1988
facilitated by thinking simulta- Klahr, Dunbar, & Fay, 1990
nously in two problem spaces
(possible alternatives)
Scientific knowledge structures Gholson & Houts, 1989
change by the development of
more differentiated and elabo-
rate conceptual nodes
Disconfirmatory strategies may be Farris & Revlin, 1989a
positive instances of counterfac-
tual hypothesis testing
Error encourages hypothesis perse- Gorman, 1989,1992
veration
Adults tentatively make separation Brewer & Chinn, 1992
between dual spaces
Adults have metacognitive skills Klahr, Fay, & Dunbar, 1993
absent in children that allow for
better coordination of searches
in two spaces
Novices' inability to search in dual Freedman, 1995
space is related to working-
memory capacity
Computer simulation Novices use heuristics similar to a Langley et al., 1987
computer simulation Qin & Simon, 1990
Experts Hypothesis testing Scientists are more prone to confir- Mahoney, 1977
mation bias than ministers
In scientists, experiments & hypoth- Mahoney, 1977
esis spaces are fully separated, D. Kuhn, 1989
whereas in children experiment
& hypothesis spaces are
merged; in novice adults, they
are partially separated
Actual Novices Textbook problems Novices gradually develop metacog- Anzai, 1991
nitive skills as they learn to
solve new problems
Scientists work on problems in a Larkin, McDermitt, Simon, &
forward, abstract manner rather Simon, 1980
than backward and concretely Larkin, 1983
as novices do
Scientists use informal, qualitative Clement, 1991
logic (analogous, simpler cases)
solve problems
Scientists form abstract representa- Chi, Feltovich, & Glaser, 1981
tions (1st principles)
Table 2 (continued)
Task Participants Topic Results Authors)
Experts Cognitive bias Scientists are just as prone to cogni- Hanson, 1962
tive bias as nonscientists Kruglanski, 1994
Mahoney, 1977,1979
Faust, 1984
Mahoney & DeMonbreun, 1977
Analogy Creating and using analogies facili- John-Steiner, 1985
tates creative insight and scien- DeMey, 1989
tific discovery Gentner & Jeziorski, 1989
Dunbar, 1995
Laboratory/field Senior scientists are least likely to Dunbar, 1995
problems show confirmation bias, and in

fact exhibit a falsification bias;

experts are more likely to
modify or discard hypotheses
compared with novices; the
most successful labs make use
of local, domain-specific analo-
gies and heuristics; finally, ser-
endipity plays an important role
in discovery
Cognitive complexity Scientists are more complex thinkers Suedfeld, 1985
than nonscientists Feist, 1994
Historical case studies Darwin developed a network of cog- Gruber, 1981,1989
nitive enterprises Tweney, 1985,1989
Faraday confirmed early and discon-
tinued later
Faraday used a large repertoire of Gooding, 1985,1990
hands-on procedures, which
influenced his mental models
In physics, imagery and metaphor Miller, 1989
have changed from being per-
ceptually based to being propo-
sitionally based
Use of metaphor and analogy is De Mey, 1989
critical to creation of scientific
knowledge; conceptual schemes
can greatly inhibit the creation
of new knowledge
Computer simulation Krebs used both general and domain- Kulkarni & Simon, 1988
of historical case specific heuristics to discover
studies the ornithine cycle
Computer programs demonstrate the Cheng & Simon, 1995
importance of visual diagrams
in scientific discovery
Developed a computer simulation Gooding & Addis, 1993
program that allows researchers
to reconstruct steps of discovery
Inventors search in Wright brothers made use of dual Bradshaw, 1992
dual space space
Bell made use of mental models and Gorman, 1995
heuristics in the invention of the
telephone
fact enhance their ability to test hypotheses scientific problems is that they include no
successfully. possibility of error in the results, even though
Replication and error. One of the limita- working scientists struggle constantly to sepa-
tions of most abstract tasks and simulated rate patterns from noise (see Gorman, 1992, for
20 FEIST AND GORMAN
an extended discussion). Gorman (1989) told novices form a kind of commonsense representa-
participants that anywhere from 0 to 20% of tion of scientific phenomena (Anzai, 1991).
their results on an abstract task similar to the
2-4-6 might be erroneous (i.e., a trial that was Simulated Scientific Tasks: Experts
classified as inconsistent with the rule might be
consistent and vice versa). Errors would occur at Metacognitive coordination of search in two
random, as determined by a random-number spaces. Another perhaps related heuristic in-
generator on a calculator. Initially, the error rate volved in successful hypothesis testing is the
was set at 0; participants encountered no actual ability to think simultaneously of two or more
errors. Gorman found that participants used problem spaces, where this term denotes a set of
replication plus extension to eliminate the possible alternatives. Klahr and Dunbar (1988)
possibility of error: they proposed experiments and Klahr, Dunbar, and Fay (1990) asked
that were similar to, but not exactly the same as, participants to learn how a device called a Big
previous experiments in an effort to replicate the Trak functions by conducting experiments. They
current pattern. This strategy resembles the found that it was most useful to think metacogni-
positive test heuristic recommended by Klay- tively in terms of separate problem spaces,
man and Ha. where one space contained ideas for possible
But replication and replication-plus-exten- experiments and another contained space for
sion are costly heuristics: they require a possible hypotheses. The most successful partici-
significant investment of time and resources of a pants reacted to falsificatory evidence in the
laboratory while other competing laboratories experimental space by developing new hypoth-
may be pursuing novel research (see Gorman, eses that represented a shift in the way they
1992, for a discussion). The cost and complexity represented the function of the device, which in
of replication can be increased on experimental turn suggested new areas of the problem space
tasks that incorporate the possibility of error. For to search for evidence.
example, when participants have to replicate an As discussed in the developmental psychol-
entire sequence of experiments rather than just a ogy of science section above, D. Kuhn (1989)
single result, the possibility of error encourages argued that the dual-space model of Klahr and
hypothesis perseveration, or a reluctance to his colleagues helps to account for major
discard a hypothesis in the face of occasional differences between child, novice adult, and
disconfirmation. When the possibility of 20% scientist. In the child, experiment and hypoth-
error is converted to actual error, participants esis spaces are merged into a single mental
had even more difficulty using replication and model; theory and evidence are adjusted to
replication plus extension to combat hypothesis maintain this representation. In the scientist,
perseveration (Gorman, 1989). Even on very theory and evidence are clearly separated. The
simple artifical tasks, replication alone is not novice adult falls somewhere between. Brewer
sufficient to isolate and eliminate errors. Obvi- and Chinn (1992) have explored the scientific
ously, scientists rely on other kinds of checks in beliefs of novice adults by giving them brief
addition to replication, for instance refinement readings on quantum theory or special relativity
of procedures. Future experiments on error that made predictions that conflicted with
should use scientific problems and tasks that commonsense beliefs about space and time and
simulate them, and also compare the perfor- cause and effect. Some participants simply
mance of scientists to novices. rejected the new information, resembling those
Form commonsense representations. Re- scientists who cling to the old paradigm. Other
searchers like McCloskey (1983), Clement participants interpreted the answer in terms of
(1982), and Carey (1992; Wiser & Carey, 1983) existing beliefs, for example, by treating relativ-
have established parallels between the mental istic phenomena as optical illusions. Similar to
models of modern novices and historical figures children, these participants adjusted evidence to
in the evolution of science. For example, fit beliefs. A final group of participants showed
McCloskey found that college students held at least partial assimilation of the new material:
beliefs about momentum that resembled those of they were able to give an answer that corre-
Philoponus (6th century) and Buridan (14th sponded to what they had read, but they "sure
century). These historical analogies suggest that didn't believe it" (1992, p. 70). These partici-
pants are beginning to make the separation the yo-yo will move to the left (Anzai, 1991).
between hypothesis and evidence, though they Once an expert has classified a problem in this
do not trust their conclusions. The key, accord- way, she can work forward rapidly to a solution
ing to D. Kuhn (1989) and Klahr, Fay, and (Green &Gilhooly, 1992).
Dunbar (1993) is the development of metacogni- Expert use of analogical reasoning. There
tive skills that permit delineation of theory and is a long literature on the importance of
evidence, and a coordinated search in two metaphor and analogy in scientific problem
spaces. solving (Clement, 1989; DeMey, 1992; Holyoak
Confirmation bias. Mahoney (1977) com- & Thagard, 1995) and even in psychology
pared a small sample of scientists working on (Leary, 1990). Clement (1991) compared the
the 2-4-6 task to a sample of Protestant way technical experts and novices solved more
ministers and, surprisingly, found that the unusual problems like determining what hap-
former were more prone to confirmation bias pens when the width of the coils on a spring is
than the latter. Kruglanski (1994) has argued doubled and the suspended weight is held
that scientists are subject to some of the same constant. Experts used informal, qualitative
cognitive biases as nonscientists, including reasoning processes; for example, they often
confirmation (see also Faust, 1984; Hanson, constructed an analogous simpler case, for
1962; Mahoney, 1977, 1979; Mahoney & instance, imagining what happens if the coils
DeMonbreun, 1977). were replaced by a U-shaped spring of the same
length. Then they related the analogy to the case
(see John-Steiner, 1985). Analogies are one very
Actual Scientific Problems
important means for arriving at appropriate
Experts work forward. Larkin, McDermitt, problem representations. Gentner and Gentner
Simon, and Simon (1980) used kinematics (1983) demonstrated that novices who used a
problems from an elementary physics textbook; flowing-waters analogy to understand electric
experts, of course, solved the problems much circuits formed a mental model that was
more rapidly and with fewer errors. But the appropriate for battery problems but not for ones
critical finding was that more efficient problem involving resistors. Green and Gilhooly (1992)
solving uses qualitative problem-solving strate- argued that
gies. Similarly, according to Larkin (1983),
the standard expert-novice contrastive paradigm by
experts worked forward from the information requiring use of problems accessible to novices has led
given, reasoning qualitatively until they arrived to a relative neglect of how experts tackle difficult
at a representation that suggested what set of problems and how experts detect and recover from
equations to use. Novices, in contrast, worked errors in the face of task difficulty, (p. 67)
backwards from the possible solution, applying
One solution to this difficulty is to look at how
equations early in the hopes of finding the
experts solve difficult, novel problems. Gentner
values of specific variables.
and Jeziorski (1989) compared the way Robert
Experts form abstract representations. In Boyle and Sadi Carnot used analogies to the
contrast to novices who tend to form common- way alchemists used them and concluded that
sense representations, expert scientists form following certain criteria when using analogies
abstract representations of scientific phenom- (e.g., avoiding mixed analogies, understanding
ena. Chi, Feltovich, and Glaser (1981) found that analogy is not causation) was the key to
that experts tended to categorize problems into types distinguishing scientific reasoning from the
that are defined by the major physics principles that pseudoscientific. But Carnot and Boyle had
will be used in solution, whereas novices tend to different styles of analogical reasoning: the
categorize them into types as defined by the entities former relied on a single analogy, deriving
contained in the problem statement, (p. 150)
principles from it, whereas the latter preferred to
For example, when asked to predict whether a work with a whole family of analogies. Nerses-
yo-yo on a table will roll to the left or right when sian (1992) observed that James Clerk Maxwell
one pulls on a string, novices say right based on used analogies iteratively, that is, he constantly
their commonsense experience with yo-yos, modified them to fit his growing understanding
whereas experts classify the problem in terms of of the constraints of the target domain. Alex-
momentum and force equilibrium and conclude ander Graham Bell deliberately followed the
22 FEIST AND GORMAN
analogy of nature and used the human ear as a positive or negative tests. When confronted with
mental model for his telephone; like Maxwell, a disconfirmatory result, however, the scientists
he was able to modify this analogy as he learned typically did one of three things: they either
more about his target domain (Gorman, 1995). changed a corollary assumption of the current
Dunbar (1995), in a cognitive study of hypothesis, attributed an anomalous result to
molecular biology laboratories, noticed that the error, or displayed a falsification bias, discard-
least successful of his four laboratories used ing results mat appeared to confirm a hypoth-
virtually no analogies, whereas the other three esis. Dunbar speculated that this falsification
used local analogies to change representations bias was a protection against airing hypotheses
and procedures. A local analogy involves that might later be proved wrong, a frequent
drawing on a similar experiment to solve a experience for the senior scientists.
problem with the current one. The backgrounds Dunbar also explicitly compared experts and
of the members of the laboratory that used no novices. Experts were more willing to modify or
analogies were too similar; they all drew on the discard hypotheses than novices. Part of this
same knowledge base. Dunbar also noted that willingness came from the fact that group
expert scientists made more analogies than interaction helped scientists articulate alternate
relative novices because the deep, structural hypotheses. In scientific practice, much of the
features of a domain were obvious to them and coordination between hypothesis and evidence
they could therefore map them readily onto goes on in groups. Perhaps that explains the
other domains. apparent difference between Tweney's and
Experts confirm early and disconfirm late. Dunbar's results: Tweney studied a detailed
Tweney (1985, 1989, 1991) used his experi- record of Faraday's experiments, and Dunbar
mental work on confirmation and disconfirma- focused on laboratory meetings. Here the
tion to frame and enrich a cognitive account of cognitive psychology of science begins to merge
how Michael Faraday used these strategies. with the social psychology of science (see
Tweney constructed detailed problem-behavior below)—conceptual change occurs in a group
graphs of Faraday's problem-solving processes. setting.
Faraday wrote about the dangers of inertia of the Experts and metacognitive dual-space search.
mind, by which he meant premature attachment Like Klahr and Dunbar's (1988) most success-
to one's own ideas (see Chamberlain, 1890/ ful participants, Alexander Graham Bell decided
1965), but he also argued that it is important to early on that he would focus more on the
ignore disconfirmatory evidence when one is hypothesis space than on the experimental one.
dealing with a new hypothesis (Tweney, 1989). He conducted a small number of experiments
In general, Faraday followed the confirm-early- and reflected constantly on the relationship
disconfirm-late heuristic: confirm until you have between evidence and hypothesis. Similarly,
a well-corroborated hypothesis, then try to Bradshaw (1992) has argued that the success of
disconfirm it. For example, his initial attempts to the Wright Brothers is due to the fact that they
use magnets to induce an electric current did a coordinated search in function and design
produced apparent discontinuations, but he spaces.
ignored them—a single confirmation was more Experts form a complex network of enter-
powerful than half-a-dozen disconfirmations, prises. Gruber (1981, 1989) originally ex-
especially given the high possibility of error in pected to rely on the work of historians in his
his initial experiments. When he obtained a cognitive analysis of Darwin's development of
more powerful magnet, he was able to reduce evolutionary theory, but he found that his
the level of noise and obtain consistent confirma- Piagetian background enabled him to see
tions. Again, Tweney's work illustrates how patterns in Darwin's activities that had eluded
experimental studies can provide a framework historians. He noticed that Darwin's apparently
for analyzing naturalistic situations, which in disparate activities fit into a network of enter-
turn, force alterations in the framework. prises including what Klahr and colleagues
Dunbar (1995) focused on laboratory meet- would call observational and hypothesis spaces
ings rather than the actual conduct of experi- (Klahr & Dunbar, 1988; Klahr, Dunbar, & Fay,
ments; therefore, it is impossible to assess 1990). Darwin's observational space included
whether scientists in his study tried to generate detailed studies of barnacles, worms, and coral
and influenced his work in the evolutionary One important aspect of embodiment is
hypothesis space in ways that are worth tracing visualization, and here computational simula-
in detail. Indeed, such a complex cognitive tions are making progress. Cheng and Simon
network appears to be common among profes- showed that it might have been easier for
sional scientists, especially the most creative Huygens and Wren to have discovered the law
ones (Gruber, 1981,1989). of conservation of momentum using diagrams
Computational simulation of historical scien- rather than deriving it from theory or by
tific problem solving. Kulkarni and Simon data-driven processes similar to those used by
(1988) used Holmes' (1980) detailed study of BACON (Cheng & Simon, 1992). Cheng then
the discovery of the ornithine cycle to create a created HUYGENS, a more general computa-
computer simulation that followed Krebs' discov- tional simulation of discovery by one-dimen-
ery process as closely as possible. KEKADA, as sional diagrams. HUYGENS uses a kind of
the program was called, relied on a dual-space dual-space search: From given numerical data,
search and a hierarchy of heuristics to accom- HUYGENS switches to a space of diagrams in
plish this goal. The hierarchy included general its search for regularities by looking for patterns
heuristics that could have been used across a in the diagrams. When patterns have been
wide range of scientific problems and specific found, the regularities are simply transformed
back into equations. The change to diagram-
ones limited to the domain of organic chemistry.
matic representation permits different operators,
One of the conclusions from KEKADA is that
regularity spotters, and heuristics to be used that
experts possess both general and domain-
are more effective than those used in the direct
specific heuristics, whereas novices are more
search of a space of algebraic terms (Cheng &
likely to possess only the more general ones. Simon, 1995). Cheng admits that we cannot be
There were even some heuristics possessed only sure the real Huygens (the discoverer of
by Krebs and a few others, including a conservation of momentum) used this method,
tissue-slicing technique that greatly facilitated but it is historically plausible, and HUYGENS
the discovery. Obviously, KEKADA could not demonstrates that it would have been more
simulate the kinds of hands-on skills that play efficient than alternatives. Instead of claiming he
such an important role in discovery. developed a program that discovers, Cheng
Shrager and Langley (1990), in an excellent argued instead that he provided computational
volume on computational simulations of scien- evidence for the importance of using diagrams
tific discovery, describe in scientific discovery, evidence that could be
combined with material from other sources
two important aspects of intellectual activity—
embedding and embodiment—that have significant (e.g., fine-grained case studies of the way
bearing on science but that have not been addressed by diagrams are used in actual discoveries). Cheng's
existing computational models. Briefly, science takes goal appears to be to provide both a normative
place in a world that is occupied by the scientist, by the account—how diagrams should be used to
physical system under study, and by other agents, and
this world has indefinite richness of physical structure
discover—and a historically plausible one—
and constraint. Thus the scientist is an embodied agent how diagrams probably were used by Huygens.
embedded in a physical and social world, (p. 15)
Gooding and Addis (1993) took another step
This criticism applies to BACON, KEKADA, in the direction of computational simulations of
and a variety of other computational ap- embodiment by developing a programming
proaches, including Paul Thagard's ECHO, a environment called CLARITY that allows them
connectionist simulation that embodies Th- to simulate Faraday's problem-solving pro-
agard's theory that the scientific hypothesis with cesses. Although CLARITY cannot conduct
the most explanatory coherence wins in dis- experiments itself, it allows researchers to
putes. ECHO has been applied to the oxygen- incorporate fine-grained details of experimental
phlogiston debate and the controversy surround- procedures in an effort to reconstruct an inventor
ing the extinction of the dinosaurs (Thagard, or scientist's path. It also allows the researcher
1988; Thagard & Nowak, 1990). This simula- to model the way in which information from
tion is directed more toward testing philosophi- articles and interaction with others influenced
cal norms for settling controversies than emulat- Faraday's thinking. Instead of making discover-
ing the psychological processes of participants. ies like BACON and KEKADA, or settling
24 FEIST AND GORMAN
controversies like ECHO, "CLARITY diagrams effective and successful heuristic in hypothesis
make hypotheses about inference and learning testing; (b) searching for solutions in two spaces
processes accessible; they can be discussed and simultaneously leads to more successful hypoth-
criticized more readily than computer code and esis testing than searching in one space; (c)
are therefore open to revision and experimenta- creating analogies and using metacognitive
tion in ways that most code-based modeling is skills facilitates successful problem solving; (d)
not" (Gooding & Addis, 1993, p. 8). In other introducing error increases hypothesis persevera-
words, CLARITY serves as an expert assistant, tion; (e) computer simulations can shed insight
helping scholars understand discovery. Such on the heuristics involved in scientific discov-
simulations will need psychologists of science ery; (f) complexity of scientific thought is
to supply detailed data on human processes in associated with scientific eminence.
similar domains. It is exactly this sort of rich and

detailed data that make simulations like

KEKADA, BACON, HUYGENS, and CLAR- Personality Psychology of Science
ITY so powerful.
Experts are cognitively complex. The study As a way of organizing the literature on the
of dispositional cognitive styles, such as integra- personality psychology of science, we catego-
tive complexity, provides a link between cogni- rize it around four fundamental topics, namely
tive and personality psychology of science (see consistent personality differences between scien-
next section). Integrative complexity is a tists and nonscientists, consistent personality
measure of complexity of thinking and is differences between eminent and less eminent
divided into two components: differentiation scientists, whether scientists of different theoreti-
and integration (Schroder, Driver, & Streufert, cal persuasions differ in terms of personality,
1967; Tetlock & Suedfeld, 1988). The simple and finally the directional influence of personal-
thinker makes relatively few qualifications and ity on scientific behavior (see Table 3).
sees things in black and white terms. In contrast,
the complex thinker not only makes distinctions Comparing Personality Characteristics
and qualifications, but integrates into a synthetic
whole the opposing points of view. Only two of Scientists to Nonscientists
studies have been conducted on integrative In 1874 Francis Galton published the first
complexity in scientists. Suedfeld (1985) re- scientific investigation of the psychological
ported that the APA presidents not only had the characteristics of scientists. Galton collected
highest complexity means compared to all qualitative self-report data from 180 English
nonscientist samples, but that the most eminent men of science and found that they were
psychologists gave the most complex presiden- energetic, were physically healthy, were perse-
tial addresses. Feist (1994) interviewed a group vering, had good memories, and were very
of eminent scientists and, among other things, independent. The study of genius was furthered
had them respond to a set of semistructured by J. McKeen Cattell (1910), Terman (1925),
questions, which were transcribed and coded on and Cox (1926). Under the guidance of Terman
integrative complexity. The mean levels of and using J. Cattell's eminent sample as
complexity in these physicists, chemists, and participants, Cox carried out the most ambitious,
biologists were even higher than those in the systematic, and most quantitative of the early
Suedfeld study. These eminent scientists were investigations into genius. Although she did not
complex thinkers about their research but not focus exclusively on scientists, she did report
about other issues (such as science education). findings broken down by group. Cox found the
traits that most clearly distinguished scientists
Conclusions From the Cognitive from nonscientific eminent men were the desire
Psychology of Science to excel, originality, reason, tendency not to be
changeable, determination, and neatness and
To summarize the consensual findings from accuracy of work. Since the 1950s, however,
the cognitive psychology of science (see Table more systematic work has focused on personal-
2): (a) confirm early, disconfirm late is an ity and scientists, with scientists being defined
as any sample that consisted of either students of female scientists being compared to female
science, engineers, inventors, social scientists, artists (Bachtold, 1976; Barton & H. Cattell,
biological scientists, or natural scientists. This 1972). The other negative finding on dominance
body of literature can be summarized with the was on a student sample (Scott & Sedlacek,
conclusions discussed in the following sections. 1975).
Scientists are more conscientious. The em- Scientists are more independent, introverted,
pirical research has revealed a consistent pattern and less sociable. Scientists, relative to nonsci-
of greater conscientiousness among scientists entists, do prefer to be alone and are somewhat
compared to nonscientists (Albert & Runco, less social (Arvey & Dewhirst, 1976; Bachtold,
1995; Bachtold, 1976; Barton & H. Cattell, 1976; Bachtold & Werner, 1976; Butcher, 1969;
1972; Feist & Barron, 1995; Gough, 1987; Ham R. Cattell & Drevdahl, 1955; Eiduson, 1962;
& Shaughnessy, 1992; Kline & Lapham, 1992; Feist, 1987; Feist & Barron, 1995; Pearce, 1968;
Schaefer, 1969; Udell, Baker, & Albaum, 1976; Roe, 1952a; Scott & Sedlacek, 1975; Wilson &
Wilson & Jackson, 1994). For example, Kline Jackson, 1994). The recent paper by Wilson and
and Lapham (1992) used the Eysenck Personal- Jackson (1994) is representative. They adminis-
ity Questionnaire (EPQ) to measure the person- tered the EPQ to 109 male and 133 female
ality characteristics of 326 science majors and physicists and compared their scores to the
compared them to 357 art majors. They reported population norms. Among the many differences,
a difference on conscientiousness between the male physicists were almost one standard
two groups that translates into an effect-size d deviation {d=— .94) and female physicists
(i.e., the difference between the two means were approximately three fourths of a standard
divided by the average standard deviation; see deviation (d = —.73) lower on the Sociable-
Cohen, 1988) of 1.59. Given the highly
Unsociable dimension. The only contrary find-
structured and organized nature of scientific
ing was reported by Mohan and Kaur (1993),
investigation, it is not surprising that scientists
who reported a mean on Extroversion for a
would have dispositions towards orderliness and
sample of scientists that was higher than the
conscientiousness.
normative mean. However, the Mohan and Kaur
Scientists are more dominant, achievement sample was from India and therefore cultural
oriented, and driven. Scientists are also more differences may be responsible for the negative
ambitious, driven, and dominant in personality relationship.
than nonscientists (Albert & Runco, 1995;
Arvey & Dewhirst, 1976; Bachtold & Werner, Scientists are emotionally stable and impulse
1972; R. Cattell & Drevdahl, 1955; Feist & controlled. Compared to nonscientists, scien-
Barron, 1995; Gough, 1987; Ham & Shaugh- tists tend to be relatively emotionally stable, low
nessy, 1992; Pearce, 1968; Schaefer, 1969; on neuroticism, and more likely to control their
Udell, Baker, & Albaum, 1976). Dominance and impulses (Albert & Runco, 1995; Bachtold,
drive appear to be distinguishing characteristics 1976; Bamber, Bill, Boyd, & Corbett, 1983;
of both female and male scientists. For example, Barton & H. Cattell, 1972; Butcher, 1969; R.
Bachtold and Werner (1972) collected personal- Cattell & Drevdahl, 1955; Eiduson, 1962; Feist
ity data on 116 female biologists and chemists & Barron, 1995; Gough, 1987; Ham & Shaugh-
listed in Who's Who in America and Who's Who nessy, 1992; Mossholder, Dewhirst, & Arvey,
of American Women and compared them to 1981; Scott & Sedlacek, 1975; Terman, 1954;
female norms. Using Cattell's Sixteen Personal- Wilson & Jackson, 1994). For example, Cattell
ity Factor (16PF) as the measure of personality, and Drevdahl (1955) gave the 16PF to scientists
they found that the women scientists were more who were classified as primarily researchers,
dominant, confident, intelligent, radical, and teachers, or administrators. Taking the means on
adventurous than women in general. Further- only the researcher subsample (n = 144) and
more, the personality profile of female scientists comparing them to norms, researchers were
was quite consistent with that of male scientists more than half a standard deviation higher on
(Bachtold & Werner, 1972). Two of the three impulse control (d = .54) and more than two
studies that found scientists to be less dominant thirds of a standard deviation higher on ego
than a comparison group was on samples of strength (d = .70).
26 FEIST AND GORMAN
Table 3
Summary of Topics and Findings in the Personality Psychology of Science Literature
Scientists vs. nonscientists Scientists are more conscientious or Schaefer, 1969
orderly Barton & Cattell, 1972
Bachtold, 1976
Gough, 1987
Ham & Shaughnessy, 1992
Kline & Lapham, 1992
Wilson & Jackson, 1994
Albert & Runco, 1995
Feist & Barron, 1995
Scientists are more dominant, driven, or R. Cattell & Drevdahl, 1955

achievement oriented Pearce, 1968

Schaefer, 1969
Bachtold & Werner, 1972
Scott &Sedlacek, 1975
Arvey & Dewhirst, 1976
Udell, Baker, & Albaum, 1976
Gough, 1987
Scientists are more independent and less Roe, 1952a
sociable Cattell & Drevdahl, 1955
Terman, 1955
Eiduson, 1962
Pearce, 1968
Butcher, 1969
Bachtold & Wemer, 1972
Scott & Sedlacek, 1975
Arvey & Dewhirst, 1976
Bachtold, 1976
Feist, 1987
Scientists are more emotionally stable or Roe, 1952a
impulse controlled Cattell & Drevdahl, 1955
Eiduson, 1962
Butcher, 1969
Barton & Cattell, 1972
Scott & Sedlacek, 1975
Bachtold, 1976
Mossholder, Dewhirst, & Arvey, 1981
Bamber, Bill, Boyd, & Corbett, 1983
Gough, 1987
Eminent/creative vs. less eminent/cre- Creative scientists are more dominant, VanZelst&Kerr, 1954
ative arrogant, self-confident, or hostile Gough, 1961
Wispe, 1963
Chambers, 1964
Garwood, 1964
Parloff&Datta, 1965
McDermid, 1965
Davids, 1968
Parloff et al., 1968
Shapiro, 1968
Schaefer, 1969
Erickson et al., 1970
Table 3 (continued)
Gantz et al., 1972
Lacey & Erickson, 1974
Rushton, Murray, & Paunonen, 1987
Helmreich, Spence, & Pred, 1988
Feist, 1993
Creative scientists are more autonomous, Van Zelst&Kerr, 1954
independent, or introverted Holland, 1961
Chambers, 1964
Garwood, 1964
Parloff&Datta,1965
Davids, 1968
Schaefer, 1969
Helson & Crutehfield, 1970
Smithers & Batcock, 1970
Helson, 1971
Busse & Mansfield, 1984
Feist, 1993
Roco, 1993
Feist & Barren, 1995
Creative scientists are more driven, Van Zelst&Kerr, 1954
ambitious, or achievement oriented Gough, 1961
Holland, 1961
Wispe, 1963
Chambers, 1964
Davids, 1968
Shapiro, 1968
Schaefer, 1969
Gantz et al., 1972
Busse & Mansfield, 1984
Dcapaahindi, 1987
Helmreich, Spence, & Pred, 1988
Creative scientists are more open and Van Zelst&Kerr, 1954
flexible in thought or behavior Gough, 1961
Wispe, 1963
Garwood, 1964
Parloffetal., 1968
Shapiro, 1968
Schaefer, 1969
Helson & Crutehfield, 1970
Helson, 1971
Rushton et al., 1987
Roco, 1993
Theoretical predilection Personality influences theory creation, Atwood & Tomkins, 1976
acceptance, and orientation Johnson et al., 1988
Hart, 1982
Royalty & Magcon, 1985
Directional influence Directional influence between person- Eiduson, 1974
ality and scientific behavior is Feist, 1993
uncertain Feist & Barron, 1995
28 FEIST AND GORMAN
Comparing Personality Characteristics Spence, & Pred, 1988; Holland, 1961; Dcpaa-
of Eminent and Creative to Less Eminent hindi, 1987; Lacey & Erickson, 1974; Rushton
and Creative Scientists et al., 1987; Schaefer, 1969; Shapiro, 1968; Van
Zelst & Kerr, 1954; Wispe, 1963). Busse and
In addition to the general research on the Mansfield (1984), for instance, studied the
distinctive personality characteristics of scien- personality characteristics of 196 biologists, 201
tists, another body of work has focused more chemists, and 171 physicists. Holding age and
specifically on the unique personality character- professional age constant, commitment to work
istics of the most successful, eminent, and (i.e., "need to concentrate intensively over long
creative scientists. periods of time on one's work") was the
Eminent-creative scientists are more domi- strongest predictor of productivity (i.e., publica-
nant, arrogant, hostile, and self-confident. In tion quantity). Of course, drive and ambition are
the highly competitive world of science, espe- predictive of success in other fields also, but
cially "big" science, where the most productive their effect in science are nevertheless important
and influential continue to be rewarded with to demonstrate.
more and more of the resources, success is more Eminent-creative scientists are more autono-
likely for those who thrive in competitive mous, introverted, and independent. If scien-
environments: the dominant, arrogant, hostile, tists in general are more aloof, asocial, and
and self-confident (Chambers, 1964; Davids, introverted than nonscientists, then these charac-
1968; Erickson, Gantz, & Stephenson, 1970; teristics appear to be even more salient for the
Feist, 1993; Gantz, Erickson, & Stephenson, scientific elite (Busse & Mansfield, 1984;
1972; Garwood, 1964; Gough, 1961; Helm- Chambers, 1964; Davids, 1968; Erickson et al.,
reich, Spence, and Pred, 1988; Helson & 1970; Garwood, 1964; Helson, 1971; Helson &
Crutchfield, 1970; Lacey & Erickson, 1974; Crutchfield, 1970; Holland, 1961; Lacey &
McDermid, 1965; Parloff & Datta, 1965; Erickson, 1974; Parloff & Datta, 1965; Roco,
Parloff, Datta, Kleman, & Handlon, 1968; 1993; Rushton, Murray, & Paunonen, 1987;
Rushton, Murray, & Paunonen, 1987; Schaefer, Schaefer, 1969; Smithers & Batcock, 1970; Van
1969; Shapiro, 1968; Van Zelst & Kerr, 1954; Zelst & Kerr, 1954). Chambers (1964) used the
Wispe, 1963). For example, Van Zelst and Kerr 16PF to obtain self-reported personality data
(1954) collected personality self-descriptions from 225 chemists and 213 psychologists.
from 514 technical and scientific personnel from Based on awards and honors, some of these
a research foundation and a university. Holding scientists were classified in the creative group,
age constant, they reported significant partial whereas the rest were classified in the less
correlations between productivity and the self- creative group. The creative scientists were .39
description of argumentative {rv = .23), assertive of a standard deviation higher than the less
(rp = .22), and self-confident (rp = .35). Simi- creative scientists on self-sufficiency. Further-
larly, Feist (1993) presented a structural- more, Helson (1971) compared creative female
equations model of scientific eminence in which mathematicians matched on IQ with less cre-
the path between observer-related hostility and ative female mathematicians. Observers blindly
eminence was direct and the path between rated the former as having more "unconven-
arrogant working style and eminence was tional thought processes," as being more
indirect (mediated by productivity) but signifi- "rebellious and nonconforming," and as being
cant. less likely to judge "self and others in
Eminent-creative scientists are more driven, conventional terms" than the latter.
ambitious, and achievement oriented. Related Eminent-creative scientists are more open to
to their hostility, arrogance, dominance, and experience or flexible in thought and behavior.
self-confidence, the most eminent and creative A final consistent effect of personality on
scientists also tend to be more driven, ambitious, creativity in science is the finding that creative
and achievement oriented than their less emi- and eminent scientists tend to be more open to
nent peers (Busse & Mansfield, 1984; Cham- experience and more flexible in thought than
bers, 1964; Davids, 1968; Erickson et al., 1970; less creative and eminent scientists (Feist &
Gantz et al., 1972; Gough, 1961; Helmreich, Barron, 1995; Garwood, 1964; Gough, 1961;
Helson, 1971; Helson & Crutchfield, 1970; whether these traits are causes or effects of
Parloff & Datta, 1965; Parloff et al., 1968; Roco, scientific behavior. To put it most simply, do
1993; Schaefer, 1969; Shapiro, 1968; Van Zelst smart, conscientious, introverted, driven, and
& Kerr, 1954; Wispe, 1963). Many of these controlled people become scientists or does
findings stem from data on the Flexibility scale science create smart, conscientious, introverted,
(Fe) of the California Psychological Inventory driven, and controlled people? Out of logical
(Feist & Barron, 1995; Garwood, 1964; Gough, necessity, it would seem very unlikely that any
1961; Helson, 1971; Helson & Crutchfield, of these characteristics would be nonexistent
1970; Parloff & Datta, 1965). The Fe scale taps until one became a scientist, and therefore
into flexibility and adaptability of thought and unlikely that being a scientist actually caused
behavior as well as the preference for change these traits of personality. However, some of
and novelty (Gough, 1987). The few studies that

them may in fact become more pronounced after

have reported either no effect or a negative
being trained as a scientist and after practicing
effect of flexibility in scientific creativity have
been with student samples (Davids, 1968; science. As is often the case, however, the model
Smithers & Batcock, 1970). that may best fit the relationship between
personality and scientific behavior is probably
bidirectional, going from personality to scien-
Personality and Theoretical Predilection tific behavior and from scientific behavior to
personality. Because one cannot perform experi-
Earlier we reviewed the literature on age and mental designs on either occupational interest or
theory acceptance and showed that there is little personality, the best methodology one can use to
evidence for Planck's principle (new theories address issues of causality is longitudinal
are accepted only once old scientists die and design. However, although longitudinal data are
younger ones take control). The work on able to address the first two criteria of causality,
personality can also shed light on theory namely covariation and temporal precedence,
acceptance and even theory creation. Or stated they still do not easily address the third and most
as a question, do certain personality styles difficult criterion: ruling out extraneous variable
predispose a scientist to create, accept, and/or
explanations (Rosenthal & Rosnow, 1991).
reject certain theories? The first work on this
question was done by Atwood and Tomkins Of the dozen or so studies that have
(1976), who showed through case studies how examined scientific behavior longitudinally
the personality of the theorist influences his or (Arvey, Dewhirst, & Brown, 1976; S. Cole,
her theory of personality. More systematic 1979; Diamond, 1986; Eiduson, 1974; Feist &
empirical investigations have expanded this Barron, 1995; Hinrichs, 1972; Horner et al.,
work and have demonstrated that personality 1986; Roe, 1965; Root-Bernstein, Bernstein, &
influences not only theories of personality, but Gamier, 1995; Simonton, 1991, 1992a; Subot-
also the theoretical orientation of behavioral nik et al., 1993; Terman, 1954), most have
scientists and how quantitatively or qualitatively focused on questions of age and productivity
oriented they are (Conway, 1988; Hart, 1982; and only two have looked at personality across
Johnson, Germer, Efran, & Overton, 1988; time (Eiduson, 1974; Feist & Barron, 1995).
Royalty & Magoon, 1985; Zachar & Leong, Initial results examining the directionality ques-
1992). However, all of these studies have been tion from the Feist and Barron study show that
with psychologists, so answering the question of
certain personality traits, such as dominance,
whether these results generalize to the biological
may become more pronounced during and after
and natural sciences remains a task for future
psychologists of science. a career in science (Feist & Barron, 1995),
suggesting a directional influence from career to
personality. However, before one can confirm
Directional Influence Between Personality such an inference, one must rule out alternative
and Scientific Behavior variable explanations. For instance, perhaps age
and maturation, not scientific careers, leads to
The most pressing question that begs to be this difference in dominance. Lack of research
addressed from the personality findings is on longitudinal personality change and stability
30 FEIST AND GORMAN
is one of the real shortcomings of the personality thought, feeling and behavior of individuals are
psychology of science literature. influenced by the actual, imagined or implied
presence of others" (Allport, 1985, p. 3). As
Conclusions From the Personality Shadish et al. (1994) noted, substituting individu-
Psychology of Science als with scientists in Allport's quotation creates
a good working definition of the social psychol-
In sum, the empirical literature over the last ogy of science. Social psychology of science
40 years has revealed rather consistent portraits may not be as well-developed as the developmen-
of the scientific personality, both in comparison tal, cognitive, or personality psychologies of
with nonscientists and to less creative scientists science, but the recent book The Social Psychol-
(see Table 3). Furthermore, personality charac- ogy of Science (Shadish & Fuller, 1994)
teristics appear to be related to which domain of suggests the field is on the verge of blossoming.
science one is attracted to (i.e., physical vs. Here, some of the main figures in social
biological vs. social science). Results have psychology have begun to produce work that is
converged on a description of scientists as more directly relevant to the social psychology of
conscientious, driven, introverted, stable, and science, and are starting to make the connection
controlled compared with nonscientists. More- quite explicitly. In what follows we review some
over, the empirical literature also suggests that of the main contributions to the social psychol-
creative scientists are more dominant, arrogant, ogy of science literature. Because of this
hostile, driven, introverted, and open and dichotomy between actual and potential, our
flexible than less creative scientists. In addition, review of social psychology of science is
personality dispositions appear to influence the divided into extant and potential-proposed
kinds of theories behavioral scientists are likely topics of investigation.
to create or accept. Finally, suggestive work on
the extent to which personality is a cause or an
effect of scientific behavior needs to be supple- Extant Social Psychology of Science
mented with more systematic research. Literature
Rosenthal's (1976, 1994) work on experi-
Social Psychology of Science menter, observer, interpreter, and expectancy
effects is without a doubt one of the more
Science is unquestionably a cognitive activ- persuasive and powerful bodies of literature
ity, and the social-cognitive and attributional relevant to the psychology of science in general
perspectives, with their emphasis on cognitive and a social psychology of science in particular.
heuristics, biases, and causal explanations, can For example, research on experimenter effects
complement the work we cited earlier on has demonstrated that participants' responses
cognitive psychology of science. Science also is can be influenced by the personality, attractive-
unquestionably a highly social activity, with ness, attire, or gender of the experimenter
much of the work being done cooperatively or (Barnes & Rosenthal, 1985). In addition,
competitively with other research teams. Ad- observer effects occur when systematic error
dressing the social factors involved in science, exists in observations of raw data, whereas
the field of social psychology of science finds interpreter effects exist when there is systematic
itself in an unusual situation. It is potentially one error in the interpretation of data. Observer and
of the richest and most stimulating areas in the interpreter effects may be quite perceptual and
psychology of science, but as yet remains more cognitive in nature, but they also have very clear
latent than actual. One can easily apply all of the social ramifications. Finally, expectancy effects
major social psychological phenomena—social are concerned with "how the investigator's
cognition, attribution theory, attitude and atti- expectation can come to serve as a self-fulfilling
tude change, conformity and social influence prophecy" (Rosenthal & Rosnow, 1991, p. 129).
and persuasion, and intergroup relations—to the That a researcher's prior expectations can affect
study of science and scientists. However, as yet, the observations, final results, and interpreta-
much of this work has not been conducted. The tions of research has been demonstrated not only
province of social psychology can be defined as when the participants are humans (Rosenthal &
"an attempt to understand and explain how the Fode, 1963a; Rosenthal & Jacobson, 1992;
Stanton & Baker, 1942), but also when they are essential tension between traditional knowledge
animals (Cordaro & Ison, 1963; Rosenthal & and revolutionary, not yet accepted knowledge
Fode, 1963b). The history of the physical, (see T. Kuhn, 1970). Empirically, Simonton has
biological, and behavioral sciences is replete shown through analysis of historical and archi-
with examples of observational and interpreta- val data how mentors and role models (see also
tional disagreements about data. More than Subotnik et al., 1993; Subotnik & Steiner, 1992)
once, junior researchers have lost jobs or were war, and political upheaval or stability influence
delegated to obscure jobs for disagreements creative output in science. Using cross-lagged
with senior colleagues over observations. In panel designs, Simonton (1975, 1976a, 1976b,
fact, often these differences are cast off as 1980) has examined the causal influence of war
simply being a result of error on the junior on scientific productivity. For instance, in
researcher's part. Granted, this may be the case examining the influence of war on productivity
in some instances, but there are many known in seven European countries from 1500 to 1900,
instances for which history has shown these Simonton reported that war had a significant
simply to be honest differences in observation influence on productivity rather than the other
(Rosenthal, 1976). way around, but the influences were complex
One contribution the social psychology of and inconsistent across country (1976a). Finally,
science has begun to make is to shed light on Simonton (1988a) has reported that often the
these experimenter and observer effects, which most creative contributions come from those
will increase the understanding of the social- who know two different cultures, suggesting
cognitive processes involved in the creation and that exposure to multiple cultural frames of
development of scientific knowledge. In Krug- reference is important for creative productivity
lanski's (1994) words, in science.
cognitive and motivational biases that influence scien-
Shadish (1989) has written about the impor-
tific conclusions are fundamentally inevitable and are tance of a psychological perspective in the
an integral part of how all knowledge is acquired. evaluation of quality in science; quality evalua-
Rather than regarding them as impediments to truth, it tions are at the heart of the scientific enterprise.
may be more practical to take them into account to Such evaluations and their criteria and measure-
improve the quality and persuasiveness of one's
research, (p. 211) ment are what determines who gets which job,
who gets tenure, who gets which grants, and
It would not be an exaggeration to say that the who gets which awards and honors (Feist,
whole field of experimenter effects could be 1997). Science is a competitive enterprise and
categorized as a subdiscipline under social and resources (read, reward and recognition) are
cognitive psychology of science. Indeed, this scarce. Of course, the question of quality in
body of work provides a prototypic example of science immediately raises a few other critical—
how social psychology has much to offer and social psychological—questions: Whose
science studies and implicitly has been doing so perceptions should be used to evaluate quality?
for years. What criteria are used? How is it decided how to
Another key figure in the social psychology weigh the various criteria? Are these criteria and
of science is Dean Simonton, whose work has evaluations fair or biased against particular
more explicitly explored how social structures individuals or groups of individuals? Until
influence the creation and maintenance of recently, philosophy, history, and sociology may
science. Theoretically, Simonton's chance- have been the disciplines most likely to address
configuration model (1988a, 1989) provides an these questions, but as Shadish wrote
explanation for how an individual scientist's Why should we think that psychology offers an
conceptual configurations and insights develop, important perspective on our understanding of science
are articulated, are communicated, are accepted quality? The reason is this: The perception of quality in
or rejected, become influential, and potentially science probably exercises an inordinate amount of
influence in scientific reward systems, and perception is
develop into a school of similar-thinking indi- largely a psychological variable, (p. 407)
viduals; how those who produce the most ideas
are most likely to wield wide-ranging influence Social negotiations and self-presentation tactics
by their high-quality work; and how individual involved with promoting one's own career
differences and social factors contribute to the clearly play a role in influencing the perceptions
32 FEIST AND GORMAN
of the "powers that be." Few would deny this. reasoning task performed no better than the best
The real question then becomes how much of a individual in a coacting group (i.e., those whose
role does self-presentation play in career suc- members work separately but were informed of
cess? The cynic may say a major role, whereas other members' hypotheses), and disconfirma-
the more naive may say no role. Rather than tory instructions were superior to confirmatory
leave the question to one's predilection towards except when there is possibility of error in the
cynicism or gullibility, we argue that the data. These findings were replicated with
question is fundamentally an empirical one and individual participants (Gorman, 1989; Gorman
therefore should be examined empirically. & Gorman, 1984), suggesting that groups
One of the so-called objective measures that perform about as well as the best of an equal
Shadish argues is important in quality evalua- number of individuals on these scientific reason-
ing tasks. Steiner (1972) called such tasks

tion is citation analysis; the importance of a

particular scientist's opus easily and fairly eureka problems, and proposed that groups
reliably can be measured by counting the would only perform better than the best of an
number of times her or his works are cited by equal number of individuals on divisible prob-
peers. Most frequently used by sociologists, lems. Modern research teams succeed in part
citation analysis seldom has examined the because they divide labor effectively among
cognitive and psychological reasons authors participants with different skills and resources.
have for citing any particular paper. Shadish, For instance, Gholson and Houts (1989) reported
Tolliver, Gray, and Gupta (1995), however, have that coacting groups were more prone to confirma-
used surveys to address the citation-analysis tion bias than interacting groups. More exper-
issue for a large sample of psychologists. They imental work needs to be done comparing group
found that oft-cited works were considered and individual performance on divisible tasks
exemplars, seen as higher in quality, published that simulate aspects of scientific reasoning.
longer ago, and often used as sources of
methods or designs. Interestingly and unexpect-
edly, frequently cited articles were also per- Potential Social Psychology of Science
ceived to be less creative. It is not clear, Literature
however, why psychologists believe highly
cited papers are high in quality but low in Possible historical case studies. Because
creativity, especially because Sternberg and relatively little actual empirical work has been
Gordeeva (1996) reported that papers were carried out on the social psychology of science,
highly cited because they were novel. The latter a few researchers have outlined how various
also found that importance of theoretical contri- methods could be applied in investigating social
bution and whether papers generated research elements in science. Shadish et al. (1994), for
were rated as the most important reasons example, outlined a simulated experimental
psychologists gave for citing a paper. Clearly, paradigm that would allow one to investigate
additional research is needed in understanding issues raised by the case study of the Devonian
the explicit and implicit reasons that scientists controversy in geology and controversies over
cite works. the existence of canals on Mars (Gorman, 1992).
There is a long and distinguished literature on In this case, the discovery of the Devonian
group processes in social psychology, and yet period in geological history was not the product
only recently has any of it focused on variables of a single individual; rather, it emerged out of a
mix of cooperative and competitive interactions
and tasks involved in science. For example,
among a group of geologists. Attribution theory
work on small-group processes in science has
(Kelley, 1967) could easily be applied to help
made use of experimental methods and provided
explain how and why one particular "discov-
some insight into differences between individu-
erer" managed to get most of the credit. In this
als and groups working on scientific problems particular case, the external attribution comes
(Gorman, 1986; Gorman, Gorman, Latta, & from the consensual perspective of the commu-
Cunningham, 1984). Gorman and his colleagues nity of geologists awarding Murchison the label
found that interacting groups (i.e., those whose discoverer, the end result of a process of social
members interacted directly) on a scientific-
negotiations. Social psychology of science can Conclusions From a Social Psychology

help unpack these negotiations. of Science
In addition, the Murchison case study can
provide insight into the role of minority Social psychology has recently begun to
influence on majority opinion in science. The actualize its vast potential to shed light on
literature on social influence suggests that a critical issues confronting science studies. Minor-
unanimous majority can cause a minority of one ity influence, attitudes and attitude change,
to conform to an erroneous position on an persuasion, and small-group processes have
unambiguous perceptual task (Asch, 1956). begun to build up at least a small, albeit it
However, a consistent, determined minority also somewhat disconnected literature. Attributional
can influence the judgments of a majority in an processes, decision making, and conformity
ambiguous perceptual task (Moscovici & Nem- each have nearly nonexistent literatures in
eth, 1974). Moscovici takes the view that science studies, and therefore have the most
minority influence forces the majority to look potential. Not only does social psychology of
more closely at the stimuli that are the focus of science have important methodological contribu-
argument. For example, in the beginning, tions to make, but also may make fundamental
Murchison's was a novel, minority view, but theoretical and substantive contributions. Only
there was no consistent majority to oppose it. social psychology of science combines an
But Murchison was a persuasive scientist and emphasis on the individual with a social context,
his consistent, determined arguments fueled a and this places social psychology of science in a
close study of those aspects of the data that he very strong position to shed light on how
thought were particularly important. Gradually, individual scientists influence and are influ-
enced by the complex social network in which
Murchison's position became the majority view.
they work.
Possible experimental paradigms. How can
these minority influence processes be studied
experimentally? One could study the circum- General Conclusions and Discussion
stances under which a minority can force a
majority to look more carefully at the data on a The psychology of science is still in its
scientific-simulation task such as the artificial infancy. There is still much that is not known
universe used by Mynatt et al. (1978), in which about the development, thought processes,
participants shot particles at shapes on a screen personalities, motives, and social factors in-
in order to discover laws governing then- volved in scientific behavior. However, there are
motion. Such experiments could be conducted clear signs that the field is solidifying and
by (a) manipulating task ambiguity by introduc- developing an autonomous sense of identity.
ing different levels of error; (b) using a Much of this growth in fact has been stimulated
confederate to play the role of minority member by confrontations with critics from philosophy,
and varying the style of argument that she or he sociology, history, and even within psychology.
uses (Rosenwein, 1994); (c) manipulating the
credibility of the minority-group members, Integrating What Is Known: A Theoretical
perhaps by presenting them as having had Structural-Equation Model
previous success with a similar task; and (d)
looking at minority influence across generations The overall findings from each of the four
(cf. Jacobs & Campbell, 1961), in which psychology of science subdisciplines (summa-
members of an original group are replaced ries of which are found at the end of each
one-by-one and each new member can consider section) can be integrated into a theoretical
the minority's arguments anew. In addition, structural model (see Figure 1). This model is a
Gorman and Rosenwein (1995) have proposed a combination and generalization of the path
possible quasiexperiment in which groups of analyses of Helmreich et al. (1980) and
individual participants try to solve problems that Mansfield and Busse (1981) and the structural
mimic scientific reasoning in a multifaceted models proposed by Feist (1993), Reynolds and
environment that simulates the social negotia- Walberg (1992), and Simonton (1977b). The
tions found in scientific communities. content of each of the latent and measured
34
FEIST AND GORMAN
variables1 and direction of their structural paths tackle unfamiliar problems, especially those that
are based on actual empirical findings, with the arise on the border of disciplines. In more
exception of the social influences latent vari- general terms, very little is known about how
able. The logic of the model is based on scientists solve actual laboratory problems (see
temporal precedence and purported causal Dunbar, 1995).
influence. The first structures that develop are Convergence in describing the scientific
the biological and developmental characteris- personality is clearly an important step forward,
tics, which, we propose, have direct influences but a critical and difficult issue remains: how to
on the personality, cognitive, and social pro- develop a cogent theoretical explanation for the
cesses. Biological and developmental effects are dynamic relationship between the personalities
also hypothesized to have significant indirect of scientists and their science. Understanding
influences on scientific behavior. Personality

that differences between scientists and nonscien-

and cognitive and social processes, in turn, are tists and between creative and less creative
hypothesized to be the most direct influences on scientists exist reveals very little about why
scientific behavior (i.e., the development of an those differences exist or how they developed.
interest in science, school achievement and The next phase of personality research on
retention, obtaining a job in science, producing scientists must focus on developmental and
works of varying quality, and receiving honors directional issues. Such questions can only be
and awards for one's achievements). One may
addressed if longitudinal research is conducted
argue that the model is so general that most any
that begins in childhood and continues up
behavior could be placed in the outcome.
through retirement.
Indeed, that is precisely to point: scientific
behavior is like any other set of complex, Another relatively ignored domain of psychol-
integrated behaviors and can be examined from ogy of science is the motivation to do science.
each of the psychological subdisciplines. We Why do people become scientists and why do
call this a theoretical structural model because it they continue to do science? The small area of
has yet to be tested directly, even though it is research that has addressed motivation in
based on empirical findings from several scientists was mainly conducted in the 1950s
different studies. We consider this model to be and 1960s (Chambers, 1964; Eiduson, 1962;
tentative and open to modification, and indeed, Maddi, 1965; McClelland, 1962; Roe, 1952a).
we hope that it will stimulate and inspire McClelland (1962), for example, reported that
researchers in the field to test it directly. creative physical scientists are unusually hard
working, even obsessed; they avoid, and are
disturbed by complex human emotions, espe-
What Is Not Known and Prescriptions cially aggression; and they develop early in life
for the Future a strong interest in the analysis of the structure
of things. Although only few have done research
The recent past has made clear what is known on motivation, one of the common and consis-
about the psychology of science, but an equally tent findings is that the most creative and
critical question, and even more important for eminent scientists are the most persistent, hard
the future, is what is not known (Houts, 1989). working, and driven (Amabile, 1996; Chambers,
Much more work is needed on the biological 1964; Eiduson, 1962; Maddi, 1965; Mansfield &
and physiological predictors of scientific apti- Busse, 1981; McClelland, 1962; Roe, 1952a;
tude and talent, as well as on the motivational Simon, 1974; Zuckerman, 1977). In addition,
changes with age. Gender differences in math some have explored intrinsic and extrinsic
performance need further exploration and empiri- motivation and found that in general the most
cal scrutiny. The development of scientific creative scientists tend to be driven more by
thinking needs to be examined longitudinally internal than external reward (Amabile, 1996;
rather than cross-sectionally, and the role that Feist, 1993).
parents, teachers, and religious orientation play
in the development and maintenance of an
interest in science must receive more attention. 1
For ease of presentation, we present the measured
Furthermore, psychologists can provide more variables inside the latent variables, rather than as rectangles
knowledge about what happens when scientists outside the latent variables.
36 FEIST AND GORMAN
Importance of Psychology of Science Future Prospects for a Psychology

for Other Metasciences of Science
Because of the recent increase in activity in So how can psychology of science acquire a
the psychology of science, there are now signs louder voice, both within and outside of
that the more mature metascientific disciplines psychology? It must follow the route of the
are taking notice of the psychology of science other metascientific disciplines, which have
and even wanting to cooperate on problems formed their own organizations and journals. If
rather than hold fast to their disciplinocentrism any discipline is to establish itself as a
(i.e., the belief in the superiority of one's own legitimate, viable, and healthy field, it must
discipline and the uselessness of others; Feist, ultimately reach the last of three stages of
progress (Mullins, 1973): Stage 1, individual

1995). For example, the epistemological ques-

scientists working on similar problems in
tion of what scientific knowledge is stands to
isolation; Stage 2, explicit identification with a
gain much from the recent developments in
field that is attracting colleagues to it; and Stage
cognitive science and artificial intelligence 3, the establishment of conferences, journals,
(Gholson & Houts, 1989; Gorman, 1992; Miller, and departments.
1989; Simon, Langley, & Bradshaw, 1981;
Tweney, 1989; Tweney et al., 1981). Indeed, an Without doubt, psychology of science has
reached Stage 1, whereby individual scientists
increasing number of contemporary philoso-
work on similar problems. This began in earnest
phers are developing the discipline of natural
in the 1950s. The field is now moving into Stage
epistemology (Fuller, 1988, 1993; Giere, 1988; 2, with more and more psychologists identifying
Gooding, 1990; Heyes, 1989; Thagard, 1988), themselves as psychologists of science. The
and they directly acknowledge the value of field is clearly not at Stage 3, although
psychology in addressing metascientific ques- conferences may indeed appear in the near
tions. future. Currently there are occasional panels at
But psychology of science is still a relatively various conferences, but these cannot substitute
unrecognized specialty, unlike history, philoso- for a regular conference that brings psycholo-
phy, and sociology of science. It is often ignored gists of science together. Further, a journal
in debates about the nature of science. For devoted to the psychology of science is still not
example, a recent book written by a scientist and on the horizon. Finally, psychology of science
needs to become a legitimate area of inquiry
a mathematician (Gross & Levitt, 1994) ap-
within university psychology departments. De-
plauded those metascience disciplines—primar-
partments at Carnegie-Mellon, Bowling Green,
ily history and philosophy of science—that and University of Memphis are pioneers in this
reinforce the view that science is a rational direction, but without a journal and a regular
enterprise. They attacked sociology of science, conference, it is likely that psychology of
feminist science studies, and other metascien- science will remain an avocation even at these
tific movements that question or undermine this institutions.
view. Psychology of science was not mentioned
at all, nor has it played any role in the ensuing
controversy, though psychology of science Importance of Psychology of Science
could contribute much to this debate. Psychol-
ogy of science could transcend and illuminate We believe the empirical findings and concep-
tual contributions of a psychology of science are
these debates by contributing empirical research
important for at least five reasons:
that seeks neither to undermine nor sanctify
1. Psychology simply cannot afford to ignore
science. Studies on how science relates to other one of the most important human activities, one
human activities (i.e., how the thinking of expert that has transformed the very world. Granted,
scientists differs from novices, and what sorts of the consequences of such knowledge are not
personalities and which kinds of group interac- uniformly constructive and positive. Combined
tion are most likely to lead to success in science), with the ability to understand and create comes
are most likely to advance theory. the ability to annihilate and destroy. This is all
the more reason to understand the psychology including science (Goleman, 1995; Sternberg,
behind science and scientific knowledge, and in 1986, 1988a; Sternberg, Wagner, Williams, &
particular how it is created, communicated, and Horvath, 1995). To the extent that gatekeepers
applied to new technologies. Having a citizenry continue to use selection criteria that measure
that is ignorant of these processes and therefore but one part of what leads to doing good science
unable to evaluate the end product of research (i.e., entrance exams and IQ tests), identifying
can only lead to misguided and misunderstood and retaining scientific talent will remain poor
attempts to control and regulate science. To the (Gardner, 1993; Subotnik et al., 1993).
extent that politicians follow the will of the 5. Studying the scientist will force psychologi-
people and control the amount of money cal theories into an important new domain,
devoted to basic and applied research, a public is leading to changes in psychological concepts. In
needed that has more than a superficial under- fact, Simonton (1990, 1995) argued that the
standing of science and why it is important. psychology of science offers an ideal and rich
2. The other metasciences—philosophy, his- field for testing general psychological theory.
tory, sociology (and more recently, anthropol- We agree with Singer (1971):
ogy)—are trying to supply their own answers to
Psychology can make important practical contributions
psychological questions concerning conceptual to the progress of science. The philosophy of science
change, theory choice, motives, and personal has generated many fundamental questions about
styles of scientists. There is no doubt that scientific behavior which can be translated into
psychology has the conceptual and theoretical research problems for the psychologist. A psychology
artillery to attack precisely these questions. of science, which seeks to understand the psychological
nature of science and of knowing in general, is a rich
3. A better psychological understanding of and fascinating enterprise, (p. 1014)
science is already leading to improvements in
pedagogy, both for those who will become Psychology of science will also encourage
scientists and for those who need to understand collaboration among psychologists from various
science in order to be informed citizens. If the subareas, helping the field achieve coherence
goals of having an informed adult populace rather than continued fragmentation. Staats
regarding science are to be met, then psycholo- (1991) has long argued for the necessity of a
gists of science need to know how to teach it in unified discipline of psychology if the field is to
ways that unleash children's natural curiosity mature. The complete and fully developed
about how the world works. Furthermore, the psychology of science must by definition
psychology of science can contribute greatly to include developmental, biological, clinical, cog-
the understanding of how children conceptual- nitive, personality, and social perspectives.
ize the physical world, and this is precisely what Collaboration and cooperation among the subdis-
the Piagetian and neo-Piagetian literatures have ciplines is required. It would be no small feat
done (Inhelder & Piaget, 1958; Klahr & and of no small import if the psychology of
Robinson, 1981; Klahr et al., 1993; D. Kuhn, science could become a model for the parent
1989). Having accurate and sophisticated cogni- discipline on how to combine resources and
tive models of children's understanding of the study science from a unified perspective.
natural world is absolutely necessary in under-
standing the development from child to adult
scientific knowledge (cf. Klahr & Robinson,
1981).
References
4. Along similar applied lines, by understand- Albert, R., & Runco, M. (1987). The possible
ing the actual psychological processes behind different personality dispositions of scientists and
science, and in particular the best science, non-scientists. In D. N. Jackson & J. P. Rushton
(Eds.), Scientific Excellence (pp. 67-97). Beverly
perhaps a psychology of science can have a loud Hills, CA: Sage.
and clear voice about selection criteria for Albeit, R., & Runco, M. (1995). [Personality
potential graduate students and faculty. It is dispositions of scientists and non-scientists]. Unpub-
becoming increasingly evident that purely cogni- lished raw data.
tive and intellectual skills are but one of many Allport, G. W. (1985). The historical background of
important predictors of successful careers, social psychology. In G. Lindzey & E. Aronson
38 FEIST AND GORMAN
(Eds.), Handbook of social psychology (3rd ed., Bell, E. T. (1937). Men of mathematics. New York:
Vol. 1, pp. 1-46) Hillsdale, NJ: Erlbaum. Simon & Sinister.
Amabile, T. (1996). Creativity in context. New York: Benbow, C. P. (1988). Sex differences in mathemati-
Westview. cal reasoning ability in intellectually talented
Anzai, Y. (1991). Learning and use of representations preadolescents: Their nature, effects, and possible
for physics expertise. In K. A. Ericsson & J. Smith causes. Behavioral and Brain Sciences, 11, 169-
(Eds.), Toward a general theory of expertise. 183.
Cambridge, England: Cambridge University Press. Benbow, C. P., & Lubinski, D. (1993). Psychological
Arvey, R. D., & Dewhirst, H. D. (1976). Goal-setting profiles of the mathematically talented. Some sex
attributes, personality variables, and job satisfac- differences and evidence supporting their biologi-
tion. Journal of Vocational Behavior, 9, 179-189. cal basis. In G. R. Rock & K. Ackrill (Eds.), The
Arvey, R. D., Dewhirst, H., & Brown, E. M. (1976). A origins and development of high ability (pp.
longitudinal study of the impact of changes in goal 44-66). Chichester, England: Wiley & Sons.
setting on employee satisfaction. Personnel Psy- Benbow, C. P., & Minor, L. L. (1986). Mathemati-
chology, 31, 596-608. cally talented students and achievement in the high
Asch, S. E. (1956). Studies of independence and school sciences. American Educational Research
conformity. A minority of one against a unanimous Journal, 23, 425^36.
majority. Psychological Monographs, 70(9, Whole Benbow, C. P., & Stanley, J. C. (1980). Sex
No. 16). differences in mathematical ability: Fact or arti-
Astin, H. S. (1975). Sex differences in mathematical fact? Science, 210, 1262-1264.
and scientific precocity. Journal of Special Educa- Benbow, C. P., & Stanley, J. C. (1982). Consequences
tion, 9, 79-91. in high school and college of sex differences in
Atwood, G. E., & Tomkins, S. S. (1976). On mathematical reasoning ability: A longitudinal
subjectivity of personality theory. Journal of the perspective. American Educational Research Jour-
History of the Behavioral Sciences, 12, 166-177. nal, 19, 598-622.
Bachtold, L. M. (1976). Personality characteristics of Benbow, C. P., & Stanley, J. C. (1983). Sex
women of distinction. Psychology of Women differences in mathematical ability: More facts.
Quarterly, 1, 70-78. Science, 222, 1029-1031.
Bachtold, L. M., & Werner, E. E. (1972). Personality Berry, C. (1981). The Nobel scientists and the origins
characteristics of women scientists. Psychological of scientific achievement. British Journal of
Reports, 31, 391-396. Sociology, 32, 381-391.
Backman, M. E. (1972). Patterns of mental abilities: Bouchard, T. J., Jr., & McGue, M. (1981). Familial
Ethnic, socioeconomic, and sex differences. Ameri- studies of intelligence: A review. Science, 212,
can Educational Research Journal, 9, 1-12. 1055-1059.
Bamber, J. H., BUI, J. M., Boyd, F. E., & Corbett, W. Bouchard, T. J., Jr., & Segal, N. L. (1990). Advanced
D. (1983). In two minds—arts and science mathematical reasoning ability: A behavioral per-
differences at sixth-form level. British Journal of spective. Behavioral and Brain Sciences, 13,
Educational Psychological, 53, 222-233. 191-192.
Barber, B. (1961). Resistance by scientists to Bradshaw, G. F. (1992). The airplane and the logic of
scientific discovery. Science, 134, 596-602. invention. In R. N. Giere (Ed.), Cognitive models
Barnes, M. L., & Rosenthal, R. (1985). Interpersonal of science (pp. 239-250). Minneapolis: University
effects of experimenter attractiveness, attire, and of Minnesota Press.
gender. Journal of Personality and Social Psychol- Bradshaw, G. R, Langley, P. W, & Simon, H. A.
ogy, 48, 435^46. (1983). Studying scientific discovery by computer
Barron, R, & Harrington, D. (1981). Creativity, simulation. Science, 222, 971-975.
intelligence, and personality. Annual Review of Brewer, W. R, & Chinn, C. A. (1992). Entrenched
Psychology, 32, 439-476. beliefs, inconsistent information, and knowledge
Barton, K., & Cattell, H. (1972). Personality charac- change. In L. Birnbaum (Ed.), Proceedings of 1991
teristics of female psychology, science and art international conferences on the learning sciences
majors. Psychological Reports, 31, 807-813. (pp. 67-73). Charlottesville, VA: Association for
Bayer, A. E., & Dutton, J. E. (1977). Career age and the Advancement of Computing in Education.
research—professional activities of academic scien- Brewer, W. R, & Samarapungavan, A. (1991). Child
tists: Tests of alternative non-linear models and theories vs. scientific theories: Differences in
some implications for higher education faculty reasoning or differences in knowledge? In R. R.
policies. Journal of Higher Education, 48, 259- Hoffman & D. S. Palermo (Eds.), Cognition and
282. the symbolic processes: Applied and ecological
Beard, G. M. (1874). Legal responsibilities in old perspectives (pp. 209-232). Hillsdale, NJ: Erlbaum.
age. New York: Russell Sage. Busse, T. V, & Mansfield, R. S. (1984). Selected
personality traits and achievement in male scien- creativity in science. In J. A. Glover, R. R.

tists. The Journal of Psychology, 116,117-131. Ronning, & C. R. Reynolds (Eds.), Handbook of
Butcher, H. J. (1969). The structure of abilities, creativity: Perspectives on individual differences
interests and personality in 1,000 Scottish school (pp. 341-381). New York: Plenum.
children. British Journal of Educational Psychol- Clement, J. (1991). Experts and science students: The
ogy, 39, 154-165. use of analogies, extreme cases, and physical
Campbell, D. T. (1989). Fragments of the fragile intuition. In J. F. Voss, D. N. Perkins, & J. W. Segal
history of psychological epistemology and theory (Eds.), Informal reasoning and education (pp.
of science. In B. Gholson, W. R. Shadish, Jr., R. A. 345-362). Hillsdale, NJ: Erlbaum.
Neimeyer, & A. C. Houts (Eds.), Psychology of Cohen, J. (1988). Statistical power analysis for
science (pp. 21—46). Cambridge, England: Cam- behavioral sciences (2nd ed.). Hillsdale, NJ:
bridge University Press. Erlbaum.
Carey, S. (1992). The origin and evolution of

Cole, J. R. (1979). Fair science: Women in the

everyday concepts. In R. N. Giere (Ed.), Cognitive scientific community. New York: Free Press.
models of science (pp. 89—128). Minneapolis: Cole, J. R. (1987). Women in science. In D. Jackson
University of Minnesota Press. & P. Rushton (Eds.), Scientific Excellence (pp.
Cattell, J. M. (1910). A further statistical study of 359-375). Beverly Hills, CA: Sage.
American men of science. Science, 32, 633-648. Cole, J. R., & Cole, S. (1973). Social stratification in
Cattell, J. M., & Brimhall, D. R. (1921). American science. Chicago: The University of Chicago Press.
Men of Science (3rd ed.). Garrison, NY: Science Cole, S. (1979). Age and scientific performance.
Press. American Journal of Sociology, 84, 958-977.
Cattell, R. B., & Drevdahl, J. E. (1955). A comparison Cole, S., & Zuckerman, H. (1987). Marriage,
of the personality profile (16PF) of eminent motherhood, and research performance in science.
researchers with that of eminent teachers and Scientific American, 256, 119-125.
administrators, and the general population. British
Conway, J. B. (1988). Differences among clinical
Journal of Psychology, 46, 248-261.
psychologists: Scientists, practitioners, and science-
Chamberlain, T. C. (1965). The method of multiple
practitioners. Professional Psychology: Research
working hypotheses. Science, 148,754-759. (Origi-
and Practice, 19, 642-655.
nal work published 1890)
Cordaro, L., & Ison, J. R. (1963). Observer bias in
Chambers, J. A. (1964). Relating personality and
classical conditioning of the planarian. Psychologi-
biographical factors to scientific creativity. Psycho-
cal Reports, 13, 787-789.
logical Monographs: General and Applied, 78,
1-20. Cox, C. (1926). Genetic studies of genius: Vol. II. The
Chambers, J. A. (1965). Comments. Science, 147, 67. early mental traits of three hundred geniuses.
Cheng, P. C.-H., & Simon, H. (1992). The right Stanford, CA: Stanford University Press.
representation for discovery: Finding the conserva- Datta, L. E. (1967). Family religious background and
tion of momentum. In D. Sleeman & P. Edwards early scientific creativity. American Sociological
(Eds.), Machine learning: Proceedings of the Ninth Review, 32, 626-635.
International Conference (pp. 62-71). San Mateo, Davids, A. (1968), Psychological characteristics of
CA: Morgan Kaufmann. high school male and female potential scientists in
Cheng, P. C.-H., & Simon, H. A. (1995). Scientific comparison with academic underachievers. Psychol-
discovery and creative reasoning with diagrams. In ogy in the Schools, 3, 79-87.
S. M. Smith, T. B. Ward, & R. A. Finke (Eds.), The Deaux, K. (1985). Sex and gender. Annual Review of
creative cognition approach (pp. 205-228). Cam- Psychology, 36, 49-81.
bridge, MA: MIT Press. De Mey, M, (1989). Cognitive paradigms and the
Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981). psychology of science. In B. Gholson, W. R.
Categorization and representation of physics prob- Shadish, R. A. Neimeyer, & A. C. Houts (Eds.),
lems by experts and novices. Cognitive Science, 5, Psychology of science: Contributions to meta-
121-152. science (pp. 275-295). Cambridge, England: Cam-
Clark, R. D., & Rice, G. A. (1982). Family bridge University Press.
constellations and eminence: The birth-orders of De Mey, M. (1992). The cognitive paradigm.
Nobel prize winners. Journal of Psychology, 110, Chicago: University of Chicago Press.
281-287. Dennis, W. (1954). Predicting scientific productivity
Clement, J. (1982). Students' preconceptions in in later decades from records of earlier decades.
introductory mechanics. American Journal of Journal of Gerontology, 9, 465-467.
Physics, 50, 66-71. Dennis, W. (1956). Age and productivity among
Clement, J. (1989). Learning via model construction scientists. Science, 123, 724-725.
and criticism: Protocol evidence on sources of Dennis, W. (1958). The age decrement in outstanding
40 FEIST AND GORMAN
scientific contributions: Fact or artifact? American Feist, G. J. (1993). A structural model of scientific
Psychologist, 13, 457-460. eminence. Psychological Science, 4, 366-371.
Dennis, W. (1966). Creative productivity between the Feist, G. J. (1994). Personality and working style
ages of 20 and 80 years. Journal of Gerontology, predictors of integrative complexity: A study of
21, 1-8. scientists' thinking about research and teaching.
Diamond, A. M. (1986). The life-cycle research Journal of Personality and Social Psychology, 67,
productivity of mathematicians and scientists. 474-484.
Journal of Gerontology, 41, 520-525. Feist, G. J. (1995). Psychology of science and history
Dunbar, K. (1995). How scientists really reason: of psychology: Putting behavioral generalizations
Scientific reasoning in real-world laboratories. In to the test. Psychological Inquiry, 6, 119-123.
R. J. Sternberg & J. Davidson (Eds.), The nature of Feist, G. J. (1997). Quantity, impact, and depth of
insight (pp. 365-396). Cambridge, MA: MIT research as influences on scientific eminence: Is
Press. quantity most important? Creativity Research

Eiduson, B. T. (1962). Scientists: Their psychological Journal, 10, 325-335.

world. New York: Basic Books. Feist, G. J., & Barren, F. (1995, October). Do hostile
Eiduson, B. T. (1974). 10 year longitudinal Rorsch- and arrogant scientists become eminent or are
achs on research scientists. Journal of Personality eminent scientists likely to become hostile and
Assessment, 38, 405^10. arrogant? Paper presented at the annual conference
Erickson, C , Gantz, B. S., & Stephenson, R. W. of Society for Social Studies of Science, Charlottes-
ville, VA.
(1970). Logical and construct validation of a
short-form biographical inventory predictor of Fisch, R. (1977). Psychology of science. In I.
scientific creativity. In Proceedings of the Annual Spiegel-Rosing & D. de Solla Price (Eds.), Science,
Convention of the American Psychological Associa- technology, and society: A cross disciplinary
tion, 5(Pt. 1), 151-152. perspective (277-318). London: Sage.
Fischbein, S. (1990). Biosocial influences on sex
Eysenck, H. J. (1988). O tempora, o mores!
differences for ability and achievement test results
Behavioral and Brain Sciences, 11, 189-190.
as well as marks at school. Intelligence, 14,
Eysenck, H. J. (1993). Creativity and personality: 127-139.
Suggestions for a theory. Psychological Inquiry, 4,
Fox, L. H. (1976). Sex differences in mathematical
147-178.
precocity: Bridging the gap. In D. P. Keating (Ed.),
Eysenck, H. J. (1995). Genius: The natural history of Intellectual talent: Research and development (pp.
creativity. Cambridge, England: Cambridge Univer- 32-54). Baltimore: Johns Hopkins University
sity Press. Press.
Farmer, H. S. (1988). Predicting who our future Freedman, E. (1995, October). Working memory and
scientists and mathematicians will be. Behavioral testing multiple hypotheses. Paper presented at
and Brain Sciences, 11, 190-191. Annual Convention of the Society for Social
Farris, H., & Revlin, R. (1989a). The discovery Studies of Science, Charlottesville, VA.
process: A counterfactual strategy. Social Studies of Fuller, S. (1988). Social epistemology. Bloomington,
Science, 19, 497-513. IN: Indiana University Press.
Farris, H., & Revlin, R. (1989b). Sensible reasoning Fuller, S. (1993). Philosophy of science and its
in two tasks: Rule discovery and hypothesis discontents (2nd ed.). New York: Guilford.
evaluation. Memory & Cognition, 17(2), 221-232. Galton, F. (1874). English men of science. London:
Faust, D. (1984). The limits of scientific reasoning. Macmillan.
Minneapolis: University of Minnesota Press. Gantz, B. S., Erickson, C. O., & Stephenson, R. W.
Fay, A. L., Klahr, D., & Dunbar, K. (1990). Are there (1972). Some determinants of promotion in a
developmental milestones in scientific reasoning? research and development population. In Proceed-
In Proceedings of the Twelfth Annual Conference ings of Annual Convention of the American
of Cognitive Science Society (pp. 333-339). Psychological Association, 7(Pt. 1), 451-452.
Hillsdale, NJ: Erlbaum. Gardner, H. (1993). Multiple intelligences: The
Feist, G. J. (1987). The emotional and cognitive theory in practice. New York: Basic Books.
aspects and personality dimensions of artistic and Garwood, D. S. (1964). Personality factors related to
scientific creativity. Unpublished manuscript. Uni- creativity in young scientists. Journal of Abnormal
versity of California, Berkeley. and Social Psychology, 68, 413-419.
Feist, G. J. (1991). The psychology of science: Gentner, D., & Gentner, G. R. (1983). Flowing waters
Personality, cognitive, motivational and working or teeming crowds: Mental models of electricity. In
styles of eminent and less eminent scientists. D. Gentner & A. L. Stevens (Eds.), Mental models
Unpublished doctoral dissertation, University of (pp. 99-129). Hillsdale, NJ: Erlbaum.
California, Berkeley. Gentner, D., & Jeziorski, M. (1989). Historical shifts
in the use of analogy in science. In B. Gholson, W. Gorman, M. E., & Gorman, M. E. (1984). A Com-
R. Shadish, R. A. Neimeyer, & A. C. Houts (Eds.), parison of disconfirmatory, confirmatory and a
Psychology of science: Contributions to metas- control strategy on Wason's 2-4-6 task. Quarterly
cience (pp. 296-325). Cambridge, England: Cam- Journal of Experimental Psychology, 36, 629-648.
bridge University Press. Gorman, M. E., & Rosenwein, R. (1995). Simulating
Geschwind, N., & Behan, P. (1982). Left-handedness: social epistemology. Social Epistemology, 9(1),
Association with immune disease, migraine, and 71-79.
developmental learning disorder. Proceedings of Gorman, M. E., Stafford, A., & Gorman, M. E.
the National Academy of Sciences, 79, 5097-5100. (1987). Disconfirmation and dual hypotheses on a
Getzels, J. W. (1987). Creativity, intelligence, and more difficult version of Wason's 2-4-6 task.
problem finding: Retrospect and prospect. In S. G. Quarterly Journal of Experimental Psychology, 39,
Isaksen (Ed.), Frontiers of creativity research (pp. 1-28.
88-102). Buffalo, NY: Bearly. Gorman, M. E., Gorman, M. E., Latta, R. M , &
Gholson, B., & Houts, A. C. (1989). Toward a Cunningham, G. (1984). How disconfirmatory,
cognitive psychology of science. Social Epistemol- confirmatory and combined strategies affect group
ogy, 3, 107-127. problem-solving. British Journal of Psychology,
Gholson, B., Shadish, W. R., Neimeyer, R. A., & 75, 65-79.
Houts, A. C. (Eds.). (1989). The psychology of Gough, H. G. (1961, February). A personality sketch
science: Contributions to metascience. Cambridge, of the creative research scientist. Paper presented
England: Cambridge University Press. at 5th Annual Conference on Personnel and
Giere, R. N. (1988). Explaining science: A cognitive Industrial Relations Research, Los Angeles, CA.
approach. Chicago: University of Chicago Press. Gough, H. G. (1976). What happens to creative
Gleick, J. (1992). Genius: Richard Feynman and medical students. Journal of Medical Education,
modern physics. New York: Pantheon. 51, 461-467.
Goleman, D. (1995). Emotional intelligence. New Gough, H. G. (1987). California Psychological
York: Bantam Books. Inventory Manual. Palo Alto, CA: Consulting
Gooding, D. C. (1985). In nature's school: Faraday as Psychologists Press.
an experimentalist In D. Gooding & F. James Gough, H. G., & Woodworth, D. G. (1960). Stylistic
(Eds.), Faraday rediscovered: Essays on the life variations among professional research scientists.
and work of Michael Faraday: 1791-1867. New Journal of Psychology, 49, 87-98.
York: Stockton Press. Green, A. J. K., & Gilhooly, K. J. (1992). Empirical
Gooding, D. C. (1990). Experiment on the making of advances in expertise research. In M. T. Keane &
meaning: Human agency in scientific observation K. J. Gilhooly (Eds.), Advances in the psychology
and experiment. Dordrecht, Netherlands: Kluwer of thinking (pp. 45-70). London: Harvester and
Academic Publishers. Wheatchief.
Gooding, D. C , & Addis, T. R. (1993). Modelling Gross, P. R., & Levitt, N. (1994). Higher superstition:
Faraday's experiments with visual functional The academic left and its quarrels with science.
programming 1: Models, methods and examples Baltimore: Johns Hopkins Press.
(Working Paper No. Joint Research Councils' Gruber, H. E. (1981). Darwin on man: A psychologi-
Initiative on Cognitive Science & Human Com- cal study of scientific creativity. Chicago: Univer-
puter Interaction Special Project Grant #9107137). sity of Chicago Press.
Bath, England: University of Bath. Gruber, H. E. (1989). Networks of enterprise in
Gorman, M. E. (1986). How the possibility of error creative scientific thought. In B. Gholson, W. R.
affects falsification on a task that models scientific Shadish, R. A. Neimeyer, & A. C. Houts (Eds.),
problem-solving. British Journal of Psychology, Psychology of science: Contributions to metas-
77, 85-96. cience (pp. 246-266). Cambridge, England: Cam-
Gorman, M. E. (1989). Error, falsification and bridge University Press.
scientific inference: An experimental investigation. Guilford, J. P. (1959). Traits of creativity. In H. H.
Quarterly Journal of Experimental Psychology, 41, Anderson (Ed.), Creativity and its cultivation (pp.
385-412. 142-161). New York: Harper.
Gorman, M. E. (1992). Simulating science: Heuris- Guilford, J. P. (1987). A review of a quarter century of
tics, mental models and technoscientific thinking. progress. In S. G. Isaksen (Ed.), Frontiers of
Bloomington: Indiana University Press. creativity research (pp. 45-61). Buffalo, NY:
Gorman, M. E. (1995). Confirmation, disconfirmation Bearly.
and invention: The case of Alexander Graham Bell Guyer, L., & Fidell, L. (1973). Publications of men
and the telephone. Thinking and Reasoning, 7(7), and women psychologists: Do women publish
31-53. less? American Psychologist, 28, 157-160.
42 FEIST AND GORMAN
Ham, S., & Shaughnessy, M. F. (1992). Personality Houts (Eds.), Psychology of science (pp. 47-88).
and scientific promise. Psychological Reports, 70, Cambridge, England: University of Cambridge
971-975. Press.
Hanson, N. R. (1962). Scientists and logicians: A Hudson, L. (1958). Undergraduate academic record
confrontation. Science, 138, 1311-1313. of Fellows of the Royal Society. Nature, 182, 1326.
Hargens, L. L., McCann, J. C , & Reskin, B. F. Hull, D. L., Tessner, P. D., & Diamond, A. M. (1978).
(1978). Productivity and reproductivity: Fertility Planck's principle: Do younger scientists accept
and professional achievement among research new scientific ideas with greater alacrity than older
scientists. Social Forces, 57, 154—163. scientists? Social Studies of Science, 202, 717-723.
Hart, J. J. (1982). Psychology of the scientists: XLVI: Hus6n, T. (1960). Abilities of twins. Scandinavian
Correlation between theoretical orientation in Journal of Psychology, 1, 25-35.
psychology and personality type. Psychological Ikpaahindi, L. (1987). The relationship between the
Reports, 50, 795-801. needs for achievement, affiliation, power, and

Helmreich, R. L., Spence, J. T., Beane, W. E., Lucker, scientific productivity among Nigerian veterinary
G. W., & Matthews, K. A. (1980). Making it in surgeons. The Journal of Social Psychology, 127,
academic psychology: Demographic and personal- 535-537.
ity correlates of attainment. Journal of Personality Inhelder, B., & Piaget, J. (1958). The growth of
and Social Psychology, 39, 896-908. logical thinking from childhood to adolescence.
Helmreich, R. L., Spence, J. T., & Pred, R. S. (1988). New York: Basic Books.
Making it without losing it: Type A, achievement Jacobs, R. C , & Campbell, D. T. (1961). The
motivation and scientific attainmentrevisited.Person- perpetuation of an arbitrary tradition through
ality and Social Psychology Bulletin, 14, 495-504. several generations of a laboratory microculture.
Helson, R. (1971). Women mathematicians and the Journal of Abnormal and Social Psychology, 83,
creative personality. Journal of Consulting and 649-658.
Clinical Psychology, 36, 210-220. Johnson, J. A., Germer, C. K., Efran, J. S., & Overton,
Helson, R., & Crutchfield, R. S. (1970). Mathemati- W. F. (1988). Personality as the basis for theoretical
cians: The creative researcher and the average predilections. Journal of Personality and Social
PhD. Journal of Consulting and Clinical Psychol- Psychology, 55, 824-835.
ogy, 34, 250-257. John-Steiner, V. (1985). Notebooks of the mind.
Heyes, C. M. (1989). Uneasy chapters in the Albuquerque: University of New Mexico Press.
relationship between psychology and epistemol- Kanigel, R. (1991). The man who knew infinity: A life
ogy. In B. Gholson, W. R. Shadish, R. A. Neimeyer, of the genius Ramanujan. New York: Scribner's &
& A. C. Houts (Eds.), Psychology of science: Sons.
Contributions to metascience (pp. 115-137). Cam- Karlsson, J. (1991). Genetics of human mentality.
bridge, England: Cambridge University Press. New York: Praeger.
Hinrichs, J. R. (1972). Value adaptation of new PhDs Karmiloff-Smith, A. (1988). A child is a theoretician,
to academic and industrial environments: A com- not an inductivist. Mind and Language, 3, 183-
parative longitudinal study. Personnel Psychology, 195.
25, 545-565. Keating, D. P. (1974). The study of mathematically
Holden, C. (1987). Female math anxiety on the wane. precocious youth. In J. C. Stanley, D. P. Keating, &
Science, 236, 600-601. L. H. Fox (Eds.), Mathematical talent: Discovery,
Holland, J. (1961, June). The assessment and description, and development (pp. 23-46). Balti-
prediction of creative performance of high aptitude more: Johns Hopkins University Press.
youth. Paper presented at the 4th University of Keller, E. F. (1985). Reflections on gender and
Utah Conference on the Identification of Creative science. New Haven, CT: Yale University Press.
Scientific Talent, Salt Lake City, UT. Kelley, H. H. (1967). Attribution theory in social
Holmes, F. L. (1980). Hans Krebs and the discovery psychology. In D. L. Vine (Ed.), Nebraska
of the ornithine cycle. Federation Proceedings, 39, symposium on motivation (pp. 192-240). Lincoln:
216-225. University of Nebraska Press.
Holyoak, K. J., & Thagard, P. (1995). Mental leaps. Klahr, D., & Dunbar, K. (1988). Dual space search
Cambridge, MA: MIT Press. during scientific reasoning. Cognitive Science, 12,
Homer, K. L., Rushton, J. P., & Vernon, P. A. (1986).
Relation between aging and research productivity Klahr, D., Dunbar, K., & Fay, A. L. (1990). Designing
of academic psychologists. Psychology and Aging, good experiments to test bad hypotheses. In J.
4, 319-324. Shrager & P. Langley (Eds.), Computational
Houts, A. (1989). Contributions of the psychology of models of discovery and theory formation (pp.
science to metascience: A call for explorers. In B. 355-402). San Mateo, CA: Morgan Kaufmann
Gholson, W. R. Shadish, R. A. Neimeyer, & A. C. Publishers, Inc.
Klahr, D., Fay, A. L., & Dunbar, K. (1993). Heuristics of present versus past generations of scientists.
for scientific experimentation: A developmental Journal of Gerontology, 17, 409-^17.
study. Cognitive Psychology, 25, 111-146. Lehman, H. C. (1966). The psychologist's most
Klahr, D., & Robinson, M. (1981). Formal assess- creative years. American Psychologist, 21, 363-
ment of problem-solving and planning processes in 369.
preschool children. Cognitive Psychology, 13, Levin, S. G., Stephen, P. E., & Walker, M. B. (1995).
113-148. Planck's principle revisited: A special note. Social
Klayman, J., & Ha, Y.-W. (1987). Confirmation, Studies of Science, 25, 275-283.
disconfirmation and information in hypothesis Loehlin, J. C , & Nichols, R. C. (1976). Heredity,
testing. Psychological Review, 94, 211-228. environment, and personality: A study of 850 sets
Kline, P., & Lapham, S. L. (1992). Personality and of twins. Austin, TX: University of Texas Press.
faculty in British universities. Personality and Long, J. S. (1992). Measures of sex differences in
Individual Differences, 13, 855-857. scientific productivity. Social Forces, 71, 159-178.
Kraglanski, A. W. (1994). The social-cognitive bases Lotka, A. J. (1926). The frequency distribution of
of scientific knowledge. In W. R. Shadish & S. scientific productivity. Journal of the Washington
Fuller (Eds.), The social psychology of science (pp. Academy of Sciences, 16, 317-323.
197-213). New York: Guilford. Lubinski, D., & Benbow, C. P. (1994). The study of
Kuhn, D. (1989). Children and adults as intuitive mathematically precocious youth: The first three
scientists. Psychological Review, 96(4), 674-689. decades of a planned 50-year study of intellectual
Kuhn, D., Amstel, E., & O'Loughlin, M. (1988). The talent. In R. F. Subotnik & K. D. Arnold (Eds.),
development of scientific thinking skills. New York: Beyond Terman: Longitudinal studies ofgiftedness
Academic Press. and talent (pp. 255-281). Norwood, NJ: Ablex.
Lykken, D. T., McGue, M., Tellegen, A., & Bouchard,
Kuhn, T. S. (1970). The structure of scientific
revolutions (2nd ed.). Chicago: University of T. J., Jr. (1992). Emergenesis: Genetic traits that
Chicago Press. may not run in families. American Psychologist,
47,1565-1577.
Kulkami, D., & Simon, H. A. (1988). The processes
Maccoby, E. E., & Jacklin, C. N. (1974). The
of scientific discovery: The strategies of experimen-
psychology of sex differences. Stanford, CA:
tation. Cognitive Science, 12, 139-175.
Stanford University Press.
Lacey, L. A., & Erickson, C. E. (1974). Psychology of MacKinnon, D. W. (1960). The highly effective
the scientist: XXXI: Discriminability of a creativ-
individual. Teachers College Record, 61, 367-378.
ity scale for the Adjective Check List among MacKinnon, D. W. (1978). In search of human
scientists and engineers. Psychological Reports,
effectiveness. Buffalo, NY: Bearly.
34, 755-758. Maddi, S. (1965). Motivational aspects of creativity.
Langley, P., Simon, H. A., Bradshaw, G. L., & Zykow, Journal of Personality, 33, 338-347.
J. M. (1987). Scientific discovery: Computational Mahoney, M. J. (1977). Publication prejudices: An
explorations of the creative processes. Cambridge, experimental study of confirmatory bias in the peer
MA: MIT Press. review system. Cognitive Therapy and Research,
Larkin, J. (1983). The role of problem representation 1, 161-175.
in physics. In D. Gentner & Stevens, A. L. (Ed.), Mahoney, M. J. (1979). Psychology of the scientist:
Mental models (pp. 75-98). Hillsdale, NJ: Lawrence An evaluative review. Social Studies of Science, 9,
Erlbaum Associates. 349-375.
Larkin, J. H., McDermitt, J., Simon, D. P., & Simon, Mahoney, M. J., & DeMonbreun, B. J. (1977).
H. A. (1980). Expert and novice performance in Psychology of the scientist: An analysis of problem
solving physics problems. Science, 208, 1335- solving bias. Cognitive Therapy and Research, 1,
1342. 229-238.
Larkin, J. H., & Simon, H. A. (1987). Why a diagram Mansfield, R. S., & Busse, T. V. (1981). The
is (sometimes) worth ten thousand words. Cogni- psychology of creativity and discovery: Scientists
tive Science, 11, 65-99. and their work. Chicago: Nelson-Hall.
Leary, D. E. (Ed.). (1990). Metaphors in the history of Maslow, A. (1966). The psychology of science. New
psychology. New York: Cambridge University York: Harper & Row.
Press. McClelland, D. (1962). On the psychodynamics of
Lehman, H. C. (1953). Age and achievement. creative physical scientists. In H. Gruber, G.
Princeton, NJ: Princeton University Press. Terrell, & M. Wertheimer (Eds.), Contemporary
Lehman, H. C. (1960). The age decrement in approaches to creative thinking (pp. 141-174).
outstanding scientific creativity. American Psy- New York: Atherton.
chologist, 15, 128-134. McCloskey, M. (1983). Naive theories of motion. In
Lehman, H. C. (1962). The creative production rates D. Gentner & A. L. Stevens (Eds.), Mental models
44 FEIST AND GORMAN
(pp. 299-324). Hillsdale, NJ: Lawrence Erlbaum productivity? Australian Psychologist, 17, 129—
Associates. 139.
McDermid, C. D. (1965). Some correlates of Over, R. (1989). Age and scholarly impact. Psychol-
creativity in engineering personnel. Journal of ogy and Aging, 4, 222-225.
Applied Psychology, 49, 14-19. Parloff, M. B., & Datta, L. (1965). Personality
Merton, R. K. (1973). The sociology of science: characteristics of the potentially creative scientist.
Theoretical and empirical investigations. Chicago: Science and Psychoanalysis, 8, 91-105.
University of Chicago Press. Parloff, M. B., Datta, L., Kleman, M., & Handlon, J.
Messeri, P. (1988). Age differences in the reception of H. (1968). Personality characteristics which differ-
new scientific theory: The case of plate tectonics entiate creative male adolescents and adults.
theory. Social Studies of Science, 18, 91-112. Journal of Personality, 36, 528-552.
Miller, A. (1989). Imagery, metaphor, and physical Pasewark, R. A., Fitzgerald, B. J., & Sawyer, R. N.
reality. In B. Gholson, W. R. Shadish, R. A. (1975). Psychology of the scientist: XXXII. God at

Neimeyer, & A. C. Houts (Eds.), Psychology of the synapse: Research activities of clinical, experi-
science (pp. 326-341). Cambridge, England: Cam- mental, and physiological psychologists. Psycho-
bridge University Press. logical Reports, 36, 671-674.
Mohan, J., & Kaur, M. (1993). Adjustment, personal Pearce, C. (1968). Creativity in young science
intelligence, values, and SES of male-female students. Exceptional Children, 35, 121-126.
university research scholars. Social Science Inter- Piaget, J., & Garcia, R. (1989). Psychogenesis and the
national, 9, 39-50. history of science. New York: Columbia University
Press.
Moore, E. G. J., & Smith, A. W. (1987). Sex and
Popper, K. (1959). The logic of scientific discovery.
ethnic group differences in mathematics achieve-
New York: Science Editions.
ment: Results from the National Longitudinal
Price, D. (1963). Little science, big science. New
Study. Journal for Research in Mathematics
York: Columbia University Press.
Education, 18, 25-36.
Qin, Y, & Simon, H. A. (1990). Laboratory
Moscovici, S., & Nemeth, C. (1974). Social influence replication of scientific discovery processes. Cogni-
II: Minority influence. In C. Nemeth (Ed.), Social tive Science, 14, 281-312.
psychology: Classic and contemporary integra- Reskin, B. F. (1977). Scientific productivity and the
tions (pp. 217-249). Chicago: Rand McNally. reward structure of science. American Sociological
Mossholder, K. W., Dewhirst, H. D., & Arvey, R. D. Review, 42, 491-504.
(1981). Vocational interest and personality differ- Reynolds, A. J., & Walberg, H. J. (1992). A structural
ences between development and research person- model of science achievement and attitude: An
nel: A field study. Journal of Vocational Behavior, extension to high school. Journal of Educational
19, 233-243. Psychology, 84, 371-382.
Mullins, N. (1973). Theories and theory groups in Roco, M. (1993). Creative personalities about cre-
contemporary American sociology. New York: ative personality in science. Revue Roumaine de
Harper and Row. Psychologie [Romanian Review of Psychology],
Mynatt, C. R., Doherty, M. E., & Tweney, R. D. 37, 27-36.
(1977). Confirmation bias in simulated research Roe, A. (1952a). The making of a scientist. New York:
environment: An experimental study of scientific Dodd, Mead.
inference. Quarterly Journal of Experimental Roe, A. (1952b). Apsychologist examines 64 eminent
Psychology, 29, 85-95. scientists. Scientific American, 187, 21-25.
Mynatt, C. R., Doherty, M. E., & Tweney, R. D. Roe, A. (1953). A psychological study of eminent
(1978). Consequences of confirmation and discon- psychologists and anthropologists, and a compari-
firmation in a simulated research environment. son with biological and physical scientists. Psycho-
Quarterly Journal of Experimental Psychology, 30, logical Monographs: General and Applied, 67,
395-406. 1-55.
Nersessian, N. J. (1992). How do scientists think? Roe, A. (1965). Changes in scientific activities with
Capturing the dynamics of conceptual change in age. Science, 150, 313-318.
science. In R. N. Giere (Ed.), Cognitive models of Roe, A. (1972). Maintenance of creative output
science (pp. 3-44). Minneapolis: University of through the years. In C. W. Taylor (Ed.), Climate
Minnesota Press. for creativity (pp. 167-191). New York: Pergamon.
Oaksford, M., & Chater, N. (1994). Another look at Root-Bernstein, R. S., Bernstein, M., & Gamier, H.
eliminative and enumerative behaviour in a concep- (1995). Correlations between avocations, scientific
tual task. The European Journal of Cognitive style, work habits, and professional impact of
Psychology, 6, 149-169. scientists. Creativity Research Journal, 8, 115-
Over, R. (1982). Is age a good predictor of research 137.
Rosenthal, R. (1976). Experimenter effects in behav- Shadish, W. R., Fuller, S., & Gorman, M. E. (1994).
ioral research (enlarged edition). New York: Social psychology of science: A conceptual and
Irvington. empirical research program. In W. R. Shadish & S.
Rosenthal, R. (1994). On being one's own case study: Fuller (Eds.), Social psychology of science (pp.
Experimenter effects in behavioral research—30 3-123). New York: Guilford Press.
years later. In W. Shadish & S. Fuller (Eds.), The Shadish, W. R., Houts, A. C , Gholson, B., &
social psychology of science. New York: Guilford. Neimeyer, R. A. (1989). The psychology of
Rosenthal, R., & Fode, K. L. (1963a). Psychology of science: An introduction. In B. Gholson, W. R.
the scientist: V: Three experiments in experimenter Shadish, R. A. Neimeyer, & A. C. Houts (Eds.),
bias. Psychological Reports, 12, 491-511. Psychology of science: Contributions to meta-
Rosenthal, R., & Fode, K. L. (1963b). The effect of science (pp. 1-16). Cambridge, England: Cam-
experimenter bias on the performance of the albino bridge University Press.
tat. Behavioral Science, 8, 183-189. Shadish, W. R., Tolliver, D., Gray, M., & Gupta, S. K.
Rosenthal, R., & Jacobson, L. (1992). Pygmalion in (1995). Author judgements about works they cite:
the classroom (enlarged edition). New York: Three studies from psychology journals. Social
Irvington. Studies of Science, 25(3), 477-498.
Rosenthal, R., & Rosnow, R. L. (1991). Essentials of Shapiro, R. J. (1968). Creative research scientists.
behavioral research: Methods and data analysis PsychologiaAfricana Monograph Supplement, No.
(2nd ed.). New York: McGraw-Hill. 4 (180).
Rosenwein, R. (1994). Social influence in science: Shrager, J., & Langley, P. (1990). Computational
Agreement and dissent in achieving scientific models of scientific discovery and theory forma-
consensus. In W. R. Shadish & S. Fuller (Eds.), The tion. San Mateo, CA: Morgan Kaufmann Publish-
social psychology of science (pp. 262-285). New ers, Inc.
York: Guilford. Siegler, R. S., & Liebert, R. M. (1975). Acquisition of
Royalty, G. M , & Magoon, T. M. (1985). Correlates formal scientific reasoning by 10- and 13-year olds:
of scholarly productivity among counseling psy- Designing a factorial experiment. Developmental
chologists. Journal of Counseling Psychology, 32, Psychology, 10, 401-402.
458-461. Simon, H. A. (1966). Scientific discovery and the
Rushton, J. P., Murray, H. G., & Paunonen, S. V. psychology of problem solving. In R. Colodny
(1987). Personality characteristics associated with (Ed.), Mind and cosmos (22-40). Pittsburgh, PA:
high research productivity. In D. Jackson & J. P. University of Pittsburgh Press.
Rushton (Eds.), Scientific excellence (pp. 129- Simon, H. A. (1974). The work habits of eminent
148). Beverly Hills, CA: Sage. scientists. Sociology of Work and Occupations, 1,
Scarr, S., & Saltzman, L. C. (1982). Genetics and 327-335.
intelligence. In R. J. Steinberg (Ed.), Handbook of Simon, H. L., Langley, P. W., & Bradshaw, G. (1981).
human intelligence (pp. 792-816). Cambridge, Scientific discovery as problem solving. Syntheses,
England: Cambridge University Press. 47, 1-27.
Schaefer, C. E. (1969). The self-concept of creative Simonton, D. K. (1975). Invention and discovery
adolescents. The Journal of Psychology, 72, among the sciences: A p-technique factor analysis.
233-242. Journal of Vocational Behavior, 7, 275-281.
Schaie, K. W. (1984). Midlife influences upon Simonton, D. K. (1976a). The causal relation between
intellectual functioning in old age. International war and scientific discovery. Journal of Cross-
Journal of Behavior Development, 7, 463-478. Cultural Psychology, 7, 133-144.
Schroder, H. M., Driver, M. J., & Streufert, S. (1967). Simonton, D. K. (1976b). Interdisciplinary and
Human information processing. New York: Holt, military determinants of scientific productivity: A
Reinhart, & Winston. cross-lagged correlation analysis. Journal of Voca-
Scott, N. A., & Sedlacek, W. E. (1975). Personality tional Behavior, 9, 53-62.
differentiation and prediction of persistence in Simonton, D. K. (1977a). Creative productivity, age,
physical science and engineering. Journal of and stress: A biographical time-series analysis of
Vocational Behavior, 6, 205-216. 10 classical composers. Journal of Personality and
Shadish, W. R. (1989). The perception and evaluation Social Psychology, 35, 791-804.
of quality in science. In B. Gholson, W. R. Shadish, Simonton, D. K. (1977b). Eminence, creativity, and
R. A. Neimeyer, & A. C. Houts (Eds.), Psychology geographic marginality: A recursive structural
of science: Contributions to metascience (pp. equation model. Journal of Personality and Social
383-426). Cambridge, England: Cambridge Univer- Psychology, 35, 805-816.
sity Press. Simonton, D. K. (1980). Techno-scientific activity
Shadish, W. R., & Fuller, S. (Eds.). (1994). Social and war: A yearly time-series analysis, 1500-1903
psychology of science. New York: Guilford. A.D. Scientometrics, 2, 251-255.
46 FEIST AND GORMAN
Simonton, D. K. (1984). Creative productivity and theory of human intelligence. New York: Viking
age: A mathematical model based on a two-step Penguin Inc.
cognitive process. Developmental Review, 4, 77- Sternberg, R. J. (1988a). A three-facet model of
111. creativity. In R. J. Sternberg (Ed.), The nature of
Simonton, D. K. (1988a). Scientific genius: A creativity (pp. 125-147). Cambridge, England:
psychology of science. Cambridge, England: Cam- Cambridge University Press.
bridge University Press. Sternberg, R. J. (1988b). The male/female difference
Simonton, D. K. (1988b). Age and outstanding is there: Should we care? Behavioral and Brain
achievement: What do we know after a century of Sciences, 11, 210-211.
research? Psychological Bulletin, 104, 251-267. Sternberg, R. J., & Gordeeva, T. (1996). The anatomy
Simonton, D, K. (1989). Chance-configuration theory of impact: What makes an article influential.
of scientific creativity. In B. Gholson, W. R. Psychological Science, 7, 69-75.
Shadish, R. A. Neimeyer, & A. C. Houts (Eds.), Sternberg, R. J., Wagner, R. K., Williams, W. M., &
Psychology of science (pp. 170-213). Cambridge, Horvath, J. A. (1995). Testing common sense.
England: Cambridge University Press. American Psychologist, 50, 912-927.
Simonton, D. K. (1990). Psychology, science, and Stevens, S. S. (1936). Psychology: The propaedeutic
history: A treatise on their convergence. New science. Philosophy of Science, 3, 90-104.
Haven, CT: Yale University Press. Stevens, S. S. (1939). Psychology and the science of
Simonton, D. K. (1991). Career landmarks in science: science. Psychological Bulletin, 36, 221-263.
Individual differences and interdisciplinary con- Subotnik, R. F., Duschl, R. A., & Selmon, E. H.
trasts. Developmental Psychology, 27, 119-130. (1993). Retention and attrition of science talent: A
Simonton, D. K. (1992a). The social context of career longitudinal study of Westinghouse science talent
success and course for 2,026 scientists and search winners. International Journal of Science
inventors. Personality and Social Psychology Education, 15, 61-72.
Bulletin, 18, 452-^*63. Subotnik, R. F , & Steiner, C. L. (1992). Adult
Simonton, D. K. (1992b). Leaders in American manifestations of adolescent talent in science.
psychology, 1879-1967: Career development, cre- Roeper Review, 15, 164-169.
ative output and professional achievement. Journal Suedfeld, P. (1985). APA presidential addresses: The
of Personality and Social Psychology, 62, 5-17. relation of integrative complexity to historical,
Simonton, D. K. (1995). Behavioral laws in histories professional and personal factors. Journal of
of psychology: Psychological science, metas- Personality and Social Psychology, 49, 1634—
cience, and the psychology of science. Psychologi- 1651.
cal Inquiry, 6, 89-114. Sulloway, F. (1996). Born to rebel: Birth order, family
Singer, B. F. (1971). Toward a psychology of science. dynamics, and creative lives. New York: Pantheon.
American Psychologist, 26, 1010-1015. Taylor, D. W. (1963). Variables related to creativity
Smithers, A. G., & Batcock, A. (1970). Success and and productivity among men in two research
failure among social scientists and health scientists laboratories. In C. W. Taylor & F. Barren (Eds.),
at a technological university. British Journal of Scientific creativity: Its recognition and develop-
Educational Psychology, 40, 144-153. ment (pp. 228-250). New York: Wiley & Sons.
Sonnert, G. (1995). What makes a good scientist?: Taylor, C. W, & Barren, F. (1963). Scientific
Determinants of peer evaluation among biologists. creativity: Its recognition and development. New
Social Studies of Science, 25, 35-55. York: John Wiley & Sons.
Staats, A. (1991). Unified positivism and unification Terman, L. M. (1925). Genetic studies of genius: Vol.
psychology. American Psychologist, 46, 899-912. 1. Mental and physical traits of a thousand gifted
Stanley, J. C. (1988). Some characteristics of SMPY's children. Palo Alto, CA: Stanford University Press.
"700-800 on SAT-M before age 13 group": Youths Terman, L. M. (1954). Scientists and nonscientists in
who reason extremely well mathematically. Gifted a group of 800 men. Psychological Monographs,
Child Quarterly, 32, 205-209. 68, (7, Whole No. 378).
Stanley, J. C , Keating, D. P., & Fox, L. H. (Eds.). Terman, L. M. (1955). Are scientists different?
(1974). Mathematical talent: Discovery, descrip- Scientific American, 192, 25-29.
tion, and development. Baltimore: Johns Hopkins Tetlock, P. E., & Suedfeld, P. (1988). Integrative
University Press. complexity coding of verbal behavior. In C. Antaki
Stanton, F., & Baker, K. H. (1942). Interviewer bias (Ed.), Lay explanation (pp. 43-59). Beverly Hills,
and the recall of incompletely learned materials. CA: Sage.
Sociometry, 5, 123-134. Thagard, P. (1988). Computational philosophy of
Steiner, I. D. (1972). Group process and productivity. science. Cambridge, MA: MIT Press.
New York: Academic Press. Thagard, P., & Nowak, G. (1990). The conceptual
Sternberg, R. J. (1986). The triarchic mind: A new structure of the geological revolution. In J. Shrager
& P. Langley (Ed.), Computational models of heritability, familiality: Which word does not
scientific discovery and theory formation (pp. belong? Psychological Inquiry, 4, 235-237.
27-72). San Mateo, CA: Morgan Kaufmann Wason, P. C. (1960). On the failure to eliminate
Publishers, Inc. hypotheses in a conceptual task. Quarterly Journal
Tweney, R. D. (1985). Faraday's discovery of of Experimental Psychology, 12, 129-140.
induction: A cognitive approach. In D. Gooding & Werts, C. E., & Watley, D. J. (1972). Paternal
F. James (Eds.), Faraday rediscovered: Essays on influence on talent development. Journal of
the life and work of Michael Faraday: 1791-1867. Counseling Psychology, 19, 367-373.
New York: Stockton Press. Wharton, C. M., Cheng, P. W., & Wickens, T. D.
Tweney, R. D. (1989). A framework for the cognitive (1993). Hypothesis-testing strategies: Why two
psychology of science. In B. Gholson, W. R. goals are better than one. Quarterly Journal of
Shadish, Jr., R. A. Neimeyer, & A. C. Houts (Eds.), Experimental Psychology, 46(4), 743-758.
Psychology of Science (pp. 342-366). Cambridge, Wiener, N. (1953). Ex-prodigy: My childhood and
MA: Cambridge University Press. youth. Cambridge, MA: MIT Press.

Tweney, R. D. (1991). Faraday's notebooks: The Wilson, G. D., & Jackson, C. (1994). The personality
active organization of creative science. Physics of physicists. Personality and Individual Differ-
Education, 26, 301-306. ences, 16, 187-189.
Tweney, R. D. (1994). Towards a cognitive psychol- Wise, L. L., Steel, L., & McDonald, C. (1979).
ogy of science. Mexican Journal of Behavioral Origins and career consequences of sex differences
Analysis, 20, 35-46. in high school mathematics achievement. Washing-
Tweney, R. D., & Chitwood, S. T. (1995). Scientific ton, DC: American Institute for Research.
reasoning. In S. E. Newstead & S. St. B. T. Evans Wiser, M., & Carey, S. (1983). When heat and
(Eds.), Perspectives on thinking and reasoning: temperature were one. In D. Gentner & A. L.
Essays in honour of Peter Wason (pp. 241-260). Stevens (Eds.), Mental models (pp. 267-297).
Hove, England: Erlbaum. Hillsdale, NJ: Lawrence Erlbaum Associates.
Tweney, R. D., Doherty, M. E., & Mynatt, C. R. Wispe, L. G. (1963). Traits of eminent American
(Eds.). (1981). On scientific thinking. New York: psychologists. Science, 141, 1256-1261.
Columbia University Press. Zachar, P., & Leong, F. T. L. (1992). A problem of
Udell, G. G., Baker, K. G., & Albaum, G. S. (1976). personality: Scientist and practitioner differences
Creativity: Necessary, but not sufficient. Journal of in psychology. Journal of Personality, 60, 667-
Creative Behavior, 10, 92-103. 677.
Vandenberg, S. G. (1988). Could these sex differences Zuckerman, H. (1977). Scientific elite. New York:
be due to genes? Behavioral and Brain Sciences, Free Press.
11, 212-214. Zuckerman, H., & Cole, J. R. (1975). Women in
Van Zelst, R. H., & Kerr, W. A. (1954). Personality American science, Minerva, 13, 82-102.
self-assessment of scientific and technical person- Zuckerman, H., & Merton, R. K. (1972). Age, aging,
nel. Journal ofApplied Psychology, 38, 145-147. and age structure. In M. W. Riley, M. Johnson, &
Vernon, P. E. (1989). The nature-nurture problem in A. Foner (Eds.), Aging and Society, Vol. 3.
creativity. In A. Glover, R. R. Ronning, & C. R. (292-356). New York: Russell Sage Foundation.
Reynolds (Eds.), Handbook of creativity: Perspec- Zuckerman, M. (1994). Behavioral expressions and
tives on individual differences (pp. 93-110). New biosocial bases of sensation seeking. New York:
York: Plenum Press. Cambridge University Press.
Wallach, M. A. (1970). Creativity. In P. H. Mussen
(Ed.), Manual of child psychology (pp. 1211-
1272). New York: Wiley & Sons. Received May 20,1997
Waller, N. G., Bouchard, T. J., Lykken, D. T., Revision received September 17,1997
Tellegen, A., & Blacker, D. M. (1993). Creativity, Accepted September 17,1997 •

Feist1998 - Psicologia Da Ciência

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Feist1998 - Psicologia Da Ciência

Transféré par

Droits d'auteur :

Formats disponibles

Review of General Psychology Copyright 1998 by the Educational Publishing Foundation

1998, Vol. 2, No. 1,3-47 1089-268W98/$3.00

The Psychology of Science: Review and Integration

occurring in the early 40s. In this sample,

of potential causes of mathematical ability).

become the nation's superstars in pure and applied

enon. Differences in family obligations, prestige Development of Scientific Reasoning

Birth Order Conclusions From Developmental

late 30s or early 40s Bayer & Dutton, 1977

Subotnik et al., 1993

Having an eminent mentor predicts scientific eminence Zuckerman, 1977

College students form commonsense MeCloskey, 1983

problems show confirmation bias, and in

fact exhibit a falsification bias;

similar domains. It is exactly this sort of rich and

detailed data that make simulations like

Scientists are more dominant, driven, or R. Cattell & Drevdahl, 1955

achievement oriented Pearce, 1968

and novelty (Gough, 1987). The few studies that

them may in fact become more pronounced after

ing tasks. Steiner (1972) called such tasks

tion is citation analysis; the importance of a

negotiations. Social psychology of science can Conclusions From a Social Psychology

influences on scientific behavior. Personality

that differences between scientists and nonscien-

Importance of Psychology of Science Future Prospects for a Psychology

progress (Mullins, 1973): Stage 1, individual

1995). For example, the epistemological ques-

personality traits and achievement in male scien- creativity in science. In J. A. Glover, R. R.

Carey, S. (1992). The origin and evolution of

Cole, J. R. (1979). Fair science: Women in the

Press. quantity most important? Creativity Research

Eiduson, B. T. (1962). Scientists: Their psychological Journal, 10, 325-335.

Reports, 50, 795-801. needs for achievement, affiliation, power, and

reality. In B. Gholson, W. R. Shadish, R. A. (1975). Psychology of the scientist: XXXII. God at

MA: Cambridge University Press. youth. Cambridge, MA: MIT Press.

Vous aimerez peut-être aussi