ArticlePDF AvailableLiterature Review

Scientific and Social Significance of Assessing Individual Differences: “Sinking Shafts at a Few Critical Points”



This chapter reviews empirical findings on the importance of assessing individual differences in human behavior. Traditional dimensions of human abilities, personality, and vocational interests play critical roles in structuring a variety of important behaviors and outcomes (e.g. achieved socioeconomic status, educational choices, work performance, delinquency, health risk behaviors, and income). In the review of their importance, the construct of general intelligence is featured, but attributes that routinely add incremental validity to cognitive assessments are also discussed. Recent experimental and methodological advances for better understanding how these dimensions may contribute to other psychological frameworks are reviewed, as are ways for determining their scientific significance within domains where they are not routinely assessed. Finally, some noteworthy models are outlined that highlight the importance of assessing relatively distinct classes of individual-differences attributes simultaneously. For understanding fully complex human phenomena such as crime, eminence, and educational-vocational development, such a multifaceted approach is likely to be the most productive.
Annu. Rev. Psychol. 2000. 51:405–444
Copyright q 2000 by Annual Reviews. All rights reserved
0084–6570/00/0201–0405$12.00 405
‘Sinking Shafts at a Few Critical Points’
David Lubinski
Department of Psychology and Human Development, Vanderbilt University, Nashville,
Tennessee 37203; e-mail:
Key Words differential psychology, general intelligence, total evidence rule,
Abstract This chapter reviews empirical findings on the importance of assessing
individual differences in human behavior. Traditional dimensions of human abilities, per-
sonality, and vocational interests play critical roles in structuring a variety of important
behaviors and outcomes (e.g. achieved socioeconomic status, educational choices, work
performance, delinquency, health risk behaviors, and income). In the review of their
importance, the construct of general intelligence is featured, but attributes that routinely
add incremental validity to cognitive assessments are also discussed. Recent experimental
and methodological advances for better understanding how these dimensions may con-
tribute to other psychological frameworks are reviewed, as are ways for determining their
scientific significance within domains where they are not routinely assessed. Finally, some
noteworthy models are outlined that highlight the importance of assessing relatively dis-
tinct classes of individual-differences attributes simultaneously. For understanding fully
complex human phenomena such as crime, eminence, and educational-vocational devel-
opment, such a multifaceted approach is likely to be the most productive.
Introduction .......................................................................................
Literature Reviewed ............................................................................407
Dispositional Attributes: Abilities, Interests, and Personality..........................407
Cognitive Abilities ..............................................................................407
Intellectual Development.......................................................................424
Vocational Adjustment..........................................................................426
Work Performance..............................................................................427
Creativity and Eminence.......................................................................428
Crime ............................................................................................430
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Health Risk Behavior...........................................................................431
Life Span Development.........................................................................431
Methodological Issues..........................................................................432
Causal Modeling................................................................................432
Causality and Confounds......................................................................432
Total Evidence Rule ............................................................................433
Throughout most of this century, a broad introduction to the psychology of individual
differences or differential psychology was standard background for graduate training
in applied psychology. Its importance was underscored by Scott: ‘‘Possibly the great-
est single achievement of the American Psychological Association is the establish-
ment of the psychology of individual differences’’ (1920:85).Differentialpsychology
comprises the psychometric assessment of abilities, personality, and vocational inter-
ests, with special emphasis devoted to their real-world significance and their devel-
opmental antecedents. Topics of interest included educational, interpersonal, and
vocational behaviors, especially those relevant to facilitating optimal adjustment to
life and work and tailoring opportunities for positive growth. Anastasi (1937), Tyler
(1965), and Willerman (1979) all wrote classic texts covering these topics, and pro-
vided the conceptualunderpinnings for psychologists working in educational,clinical,
industrial, and military settings.
Emerging out of these early conceptual foundations, accumulating empirical evi-
dence has made it clear that differential psychology can contribute to better under-
standing of academic achievement (Benbow & Stanley 1996, Snow 1991), the
particulars of intellectual development (Ackerman 1996), creativity (Eysenck 1995,
Jensen 1996), crime and delinquency (Gordon 1997, Lykken 1995), educational and
vocational choice (Dawis 1992, Snow et al 1996), health-risk behavior (Caspi et al
1997, Lubinski & Humphreys 1997), income and poverty (Hunt 1995, Murray1998),
occupational performance (Hunter & Schmidt 1996, Hough 1997), social stratifica-
tion (Gottfredson 1997), clinical prediction (Dawes 1994, Grove & Meehl 1996),
and life-span development (Harris 1995, Holahan & Sears 1995, Rowe 1994, Schaie
1996). As a matter of fact, causal models of these phenomena that do not incorporate
individual differences variables are likely to be underdetermined. In addition, as
differential psychologists devote particular attention to socially relevant phenomena,
their findingsare germane to the work of medical and socialscientists studyingpeople
at risk for negative outcomes or showing promise for positive outcomes.
As developments in differential psychology unfolded, however, and specialization
progressed, the study of individual differences became less likely to be viewed (and
reviewed) as a cohesive body of knowledge. Willerman’s (1979) comprehensive text
was the last of its kind. Basic researchers (and textbook writers) have tended since
to restrict their activities to specific classes of attributes: e.g. either human abilities,
interests, personality, or their biological and environmental antecedents. Indeed, few
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
research programs have examined these attributes simultaneously or systematically
for their collective role in explaining and predicting human psychological phenom-
ena. Yet, a much richer picture of humanity and psychological diversity is brought
into focus when constellations of individual-differences variables are assembled for
research and practice. By teaming relatively independent individual-differences vari-
ables to model human behavior, it becomes easy to illustrate how they operate in
many important contexts (whether they are measured or not). A new millennium
marks a good time to examine the study of individual differences more holistically.
Literature Reviewed
Following an examination of Cattell’s (1890) classic, wherein the term ‘‘mental test’
was first introduced,Galton (1890: 380) appendedtwo pagesof profoundly influential
remarks underscoring the importance of assessing psychological phenomena of sub-
stantive significance: ‘One of the most important objects of measurement is hardly
if at all alluded to here and should be emphasized. It is to obtain a generalknowledge
of . . . capacities ...bysinking shafts, as it were, at a few critical points. In order
to ascertain the best points for the purpose...Wethus may learn which of the
measures are the most instructive.’
The reviewed literature reveals a number of ‘deep shafts’ that would likely
impress Galton himself. First, three classes of dispositional attributes will be
reviewed: abilities, interests,and personality. To keepthis review downtomanageable
dimensions, abilities will be restricted to cognitive abilities, interests will focus on
educational and vocational interests, and omitted from consideration in personality
dimensions are the familiar psychopathological traits (e.g. schizophrenia, manic
depressive disorders, etc). Without this curtailment, a wide-angle view of differential
psychology would be prohibitive. Some research combining ability, interest, and
personality variables will be reviewed, followed by a discussion of methodological
issues pertaining to mis-specified causal modeling. This chapter concludes by expli-
cating some ideas behind the concept of ‘niche building’ (i.e. how individuals seek
out, build, and create environments that correspond to their personal attributes). This
analysis may resolve conflicts between various groups, e.g. the tensions observed
between Snow’s (1967) ‘two cultures’ (the humanists and the scientists) or, closer
to home, psychologists who work with people (see clients) versus psychologists who
do not. As the psychology of individual differences illuminates issues surrounding
human diversity, it may furnish tools for facilitating cross-cultural empathy (Dawis
Cognitive Abilities
The last two decades have witnessed many ambitious examinations of cognitive
ability measures and the constructs they assess. Discussion has focused on the con-
struct of general intelligence (g). However, discourse has also extended into cognitive
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
abilities beyond g and grappled with the full dimensionality and psychometric orga-
nization of the resulting array of intellective components. At broader levels of anal-
ysis, group differences (e.g. sex, race) have been explored, along with attendant
questions about whether test bias might place certain groups at a disadvantage in the
assessment process. A further topic of inquiry has been an observed tendency for
scores on intelligence tests to rise cross-culturally over the century. Finally,biological
correlates of g have been explored, leading ultimately to speculation on the evolu-
tionary derivation of general cognitive ability. Findings from each of these areas of
investigation are reviewed below.
General Intelligence (‘‘g’’) Large-scale studies have addressed the psychological
nature of g, biological interconnections, and the validity of well-known tools pur-
porting to index g for predicting socially valued criteria. During the past decade,
treatments have intensified exponentially (Carroll 1993, Jensen 1998, Neisser et al
1996), across both familiar (core) as well as less familiar (peripheral) criterion
domains. Researchhas sharpened validity generalizationsforecastingeducationalout-
comes (Benbow 1992, Benbow & Stanley 1996, Snow 1996), occupational training,
and work performance (Hunter & Schmidt 1996, Schmidt & Hunter 1998). More is
also now known about periphery phenomena surrounding gs nomological network:
aggression, delinquency, and crime (Caspi & Moffitt 1993, Gordon 1997, Wiegman
et al 1992); health risks (Lubinski & Humphreys 1997, Macklin et al 1998); and
income and poverty (Hunt 1995, Murray 1998).
For some benchmarks, general cognitive ability covaries 0.70–0.80 withacademic
achievement measures, 0.40–0.70 with military training assignments, 0.20–0.60 with
work performance (higher values reflect higher job complexity families), 0.30–0.40
with income, and 0.20 with law abidingness (Brody 1992, 1996; Gordon 1997).
Willis & Schaie (1986) haveshed considerable light on the role ofgeneral intelligence
for practical intelligence in later life, and O’Toole (1990) has done the same for motor
vehicle accident proneness. A nice compilation of positive and negative correlates of
g is Brand’s (1987) Table 2, which documents a variety of modest correlations
between general intelligence and altruism, sense of humor, practical knowledge,
response to psychotherapy, social skills, supermarket shopping ability (positive cor-
relates), and impulsivity, accident proneness, delinquency, smoking, racial prejudice,
and obesity (negative correlates), among others. These outer-layer peripheral corre-
lates are especially thought provoking because they reveal how individual differences
in g ‘pull’ with them cascades of primary (direct) and secondary (indirect) effects
(Gottfredson 1997).
Contemporary psychologistsat opposite poles ofthe appliededucational-industrial
spectrum, such as Snow (1989) and Campbell (1990), respectively, have showcased
g in law-like empirical generalizations.
Given new evidence and reconsideration of old evidence, [g] can indeed be
interpreted as ‘ability to learn’ as long as it is clear that these terms refer to
complex processes and skills and that a somewhat different mix of these con-
stituents may be required in different learning tasks and settings. The old view
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
that mental tests and learning tasks measure distinctly different abilities should
be discarded. (Snow 1989:22)
General mental ability is a substantively significant determinant of individ-
ual differences in job performance for any job that includes information-pro-
cessing tasks. If the measure of performance reflects the information
processing components of the job and any of several well-developed standard-
ized measures used to assess general mental ability, then the relationship will
be found unless the sample restricts the variances in performance or mental
ability to near zero. The exact size of the relationship will be a function of the
range of talent in the sample and the degree to which the job requires infor-
mation processing and verbal cognitive skills. (Campbell 1990:56)
These views are widely accepted among psychometricians (Barrett & Depinet
1991, Carroll 1997, Gottfredson 1997). They will be welcomed by researchers who
have searched in vain for genuine moderator variables and felt compelled therefore
to accept, however reluctantly, Ghiselli’s (1972:270) influential but dyspeptic
appraisal: ‘It is possible that moderators are as fragile and elusive as that other will-
o-the-wisp, the suppressor variable.’ The following empirical generalization is now
one of the most robust in all of psychology: The positive correlation between work
performance (Y) and general intelligence (X) is moderated by job complexity (Z).
Substituting general academic learning for Y and accelerated abstract-curriculum for
Z, another robust empirical generalization of a moderated relationship is revealed for
curriculum and instruction (Benbow & Stanley 1996).
Yet, contentious debate has been common for research pertaining to g (Campbell
1996). Indeed, psychologists can be found on all sides of the complex set of issues
engendered by assessing general intelligence (Snyderman & Rothman 1987). This is
not new, however. Heated debate has followed this important construct since shortly
after Spearman’s (1904) initial article (cf. Chapman 1988). Nevertheless, recently,
many scientists have been determined to understand g and the means of assessing it
better. Even prior to 1994, the date marking publication of Herrnstein & Murray’s
(1994) controversial book, a number of highly visible publications appeared that
attempted (among other things) to explicate the social significance of g. For by the
1980s it was becoming clear that g played a prominent role in learning and work
(Ackerman 1988, Thorndike 1985). This development bore out Cronbach’s
(1970:197) earlier evaluation: ‘The general mental test stands today as the most
important technical contribution psychology has made to the practical guidance of
human affairs.’ Thorndike (1994:150) summarized years of research findings on
cognitive abilities: ‘‘[T]he great preponderance of the prediction that is possible from
any set of cognitive tests is attributable to the general ability that they share. What I
have called ‘empirical g’ is not merely an interesting psychometric phenomenon, but
lies at the heart of the prediction of real-life performances ....Meehl (1990:124)
remarked: ‘Almost all human performance (work competence) dispositions, if care-
fully studied, are saturated to some extent with the generalintelligence factorg, which
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
for psychodynamic and ideological reasons has been somewhat neglected in recent
years but is due for a comeback.’
By 1995, largely in response to exchanges stimulated by the Bell Curve (Herrn-
stein & Murray 1994) (both within scholarly outlets and the popular press), the APA
formed a special task force (Neisser et al 1996). Contemporaneously with the work
of this task force, several major psychological outlets published special issues (Ceci
1996, Sternberg 1997, Gottfredson 1997).
The final chapter to this story is far from complete. However, one thing is clear:
The intensity of research on intellectual abilities continues unabated. Jensen (1998)
has just unveiled his most recent book which, like Carroll’s (1993), is destined to
become a classic (Bouchard 1999, Neisser 1999). In Meehl’s (1998) words: ‘‘Verbal
definitions of the intelligence concept have never been adequate or commanded con-
sensus. Carroll’s [1993] and Jensen’s [1998] books, Human Cognitive Abilities and
The g Factor (which will be the definitive treatises on the subject for many years to
come), essentially solve the problem.’ In both works, general intelligence has been
conceptualized through a (perhaps, the) fundamental predicate of science—
covariation. General intelligence is defined by the covariation cutting across various
problem solving mediums (numerical, pictorial, verbal), assessment modalities
(group, individual), and populations (cross culturally); it reflects the general factor—
or communality—shared by these multiple operations.
To the extent that this general factor reaches out and connects with external
phenomena—covariation—a basis is formed for evaluating its scientific significance.
Jointly, these two systems of covariation (internal operations of assessment tools and
external links to extra-assessment phenomena) form the nexus of the general intel-
ligence construct. g is viewed as the central node of this nexus, with its meaning
successively clarified as conceptual and empirical interrelationships develop through
research and establish the causal directionality of the network’s strands. Spearman
(1927:89) referred to the essence of g as ‘mental energy,’ which manifested itself
in individual differences in ‘the eduction of relations and correlates.’ This was a
respectable pioneering beginning but, as indicated below, there are other ways to
construe this attribute.
While Meehl (1998) is correct that verbal definitions of intelligence have never
been ‘adequate or commanded consensus’ because writers tend to focus on the
unique features of their formulation rather than the communality that they share (cf.
Sternberg & Detterman 1986), literary definitions do have their place. For example,
they frequently point to critical core criteria and relevant peripheral criteria that con-
stitute differential degrees of importance for establishing construct validity of mea-
sures purporting to assess the attribute in question. Such distinctions can bring the
fruitfulness of a particular line of research into focus. Early psychophysical measures
of intelligence were rejected, for example, because they failed to covary with edu-
cational outcomes, rate of learning academic material, and teacher ratings—criteria
thought to be central to the meaning of intelligence; for measures not to display an
appreciable relationship with these criteria would violate the essence of what intel-
ligence was intended to embody. It was natural, therefore, that when Binet and Spear-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
man produced tests predictive of these core criteria, investigators shifted their focus
and began using the new tools in their empirical research (Thorndike & Lohman
Today, for example, there is a fair amount of agreement among measurement
experts that measures of g assess individual differences pertaining to ‘‘abstract think-
ing or reasoning,’ ‘the capacity to acquire knowledge,’’ and ‘‘problem-solving abil-
ity’ (see Snyderman & Rothman’s 1987 survey of 641 experts and Gottfredson
1997). Naturally, individual differences in these attributes influence aspects of life
outside of academic and vocational arenas because abstract reasoning, problem-
solving, and rate of learning touch so many facets of life, especially now in our
information intense society. These quoted characteristics fit with correlates at both
the core and the periphery of gs nexus. They are compatible with empirical facts.
Investigators who conceptualize intelligence differently are probably talking about
something other than psychometric g, and something less central to learning and
work performance.
Dimensionality and Organization Over the past 20 years, an understandingof how
cognitive abilities are organized (hierarchically) has emerged, through hierarchical
factor analysis (Carroll 1993, Humphreys 1994), radex scaling (Snow & Lohman
1989), and structural equation modeling (Gustafsson & Undheim 1996). To psycho-
logical researchers working outside the field of cognitive abilities, variations across
these methods mirror Allport’s distinction between his and Henry Murray’s view of
personality: ‘narcissisms of subtle difference.’ Most impressive is Carroll’s (1993)
treatment of cognitive abilities, which confirmed what a number of investigatorshave
maintained all along. Cognitive abilities are organized hierarchically and, when
administered to a wide range of talent, approximately 50% of the common variance
in heterogeneous collections of cognitive tests comprise a general factor. There is
clearly a conspicuous red thread running through variegated conglomerations ofcog-
nitive tests (and the items that form them). It reflects the largest vein of construct-
valid variance uncovered by differential psychology in terms of its external
connections. Yet, to be sure, there is psychological significance beyond the general
factor. Quantitative, spatial, and verbal reasoning abilities all possess psychological
import beyond g. This is especially true for predicting educational and career tracks
that people self-select (Achter et al 1999, Austin & Hanisch 1990, Humphreys et al
1993), but also for individual differences in criterion performance (Carroll 1993,
Jensen 1998). However, as Carroll (1993:689) has noted, the scientific significance
of various abilities comes in degrees: ‘[A]bilities are analogous to elements in the
periodic table: Some, like fluid intelligence [‘‘g’’], are obviously as important as
carbon or oxygen; others are more like rare earth elements ...’
Although Carroll’s (1993) nomenclature is presented below, other approaches
would paint a similar picture. All of the aforementioned treatments are centered by
a general factor at the apex of a hierarchy (stratum III) that is defined by the com-
munality running through a secondary tier of more content specific abilities (stratum
II): mathematical, spatial/mechanical, and verbal reasoning abilities. The stratum III
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
general factor is a global marker of intellectual complexity or sophistication,whereas
the stratum II abilities are content specific strengths and relative weaknesses. There
are others, but the above abilities command the most scientific significance. Finally,
under these dimensions are more circumscribed abilities closely associated with spe-
cific tests (stratum I), such as arithmetic reasoning, block design, vocabulary, etc.
Carroll’s (1993) three-stratum theory is, in many respects, not new. Embryonic out-
lines are seen in earlier psychometric work (Burt, Cattell, Guttman, Humphreys, and
Vernon, among others). But the empirical bases for Carroll’s (1993) conclusions are
unparalleled; readers should consult this source for a systematic detailing of more
molecular abilities.
In view of these developments, some have concluded that a fairly comprehensive
picture of the structure and forecasting capabilities of cognitive abilities has been
drawn. Consequently, little is likely to come of further examining phenotypic aspects
of intellectual behavior. For example, Jensen (1998) has argued that basic research
needs to uncover more fundamental (biological) vertical paths and develop more
ultimate (evolutionary) explanations, for genuine advances to occur. There are, how-
ever, at least two issues worthy of additional examination. The first involves the
scientific significance of lower-order dimensions of human abilities (those beyond g)
and how best to appraise their scientific worth. The second has to do with population
changes and differences.
Cognitive Abilities Beyond g Specific abilities beyond g contribute to real-world
forecasts. This becomes especially true at higher g levels [e.g. continuous gradations
extending from bright, to gifted, to profoundly gifted populations (cf. Achter et al
1996, Benbow 1992)], where the major markers of g successively pull apart (dis-
sociate). In complex educational (graduate school) and vocational (doctoral-level
occupational) environments, range truncation on g is intense because an appreciable
amount of g is necessary to operate with competence in these ever-changing, sym-
bolically dense environments (Hunt 1995, 1996). Hence, the predictive power of
other factors increases relative to general intelligence, but again, only for populations
highly selected on g. This is akin to Tanners (1965) intriguing discriminant function
analysis. The physical (body build) profiles of Olympic athletes enabled Tanner to
identify their domains of excellence (events they were competing in) with great
accuracy. Yet, within a given event, the individual differences dimensions utilized to
classify these gifted athletes were not impressive performance predictors. [That the
American Psychologist (1998) recently devoted nine letters and 12 pages to pointing
out how range truncation can attenuate correlations is commentary on the poor cumu-
lative nature of some psychological research. Reading McNemars (1964) article
would have forestalled the need for this exchange.]
Probably the simplest model of human cognitive abilities (beyond g) is Eysenck’s
(1995) two-dimensional model: the general factor and a bipolar spatial-verbal factor.
Vernon (1961) used verbal-educational-numerical (v:ed) and mechanical-practical-
spatial (k:m) as major group factors subservient to g, while Cattell (1971) has pro-
posed a fluid/crystallized distinction. Snow and his colleagues (Snow 1991, Snow &
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Lohman 1989) have discussed verbal/linguistic, quantitative/numerical, and spatial/
mechanical abilities, in addition to the general factor defined by what is common to
these symbolic, problem-solving systems. Over a variety of educational/vocational
contexts, these three regions represent important sectors of concentration; they have
also demonstrated incremental validity relative to g. However, traditional factor-
analytic treatments have not proceeded with incremental validity in mind. That is,
factor analytic models of cognitive abilities have (for the most part) focused on the
internal structure of assessment tools. Models have been based on within-instrument
covariance structure.
Many factor analysts seem to hold as their implicit (if not explicit)goal accounting
for all the common variance in a correlation matrix. However, this goal fails to
consider the psychologically significant dimensionality that might result. For exam-
ple, Carroll (1994:196) writes: ‘I have pointed out (Carroll 1993) that the general
factor on the average contributes only a little more than half the common factor
variance ofa given test; thus, lower order factors can have almostas much importance
as the general factor.’’ But it must be asked, Is this view plausible? Given the breadth
and depth of the g nexus, is it conceivable that, even collectively, lower-order cog-
nitive factors, all independent of g, could evince external relationships as important
as g by itself? Mathematically, of course, it is conceivable; but is it psychologically
conceivable based on what we know about various ability dimensions that are inde-
pendent of g when in competition with the general factor for predicting important
external criteria? Based on existing evidence, it does not seem likely.
Furthermore, Carroll (1994) seems to imply that all the dimensions resulting from
common variance among cognitive abilities have the potential of being psychologi-
cally important. Again, although this is technically possible, it is unlikely; in fact,
there is reason to suspect otherwise. This is especially true when all of the variables
in a factor analysis are assessed by the same (monomethod) modality (Carroll’s ‘a
given test’’). Understanding this idea is important, because it generalizes to issues
involving the number of dimensions needed to model both personality andvocational
interests discussed in subsequent sections. For example, in the context of a discussion
on the number of dimensions needed to characterize personality, Block (1995a:189)
noted: ‘[T]he amount of variance ‘explained’ internally by a factor need not testify
to the external psychological importance of the factor.’
Within a domain of individual-differences measures, only aportion of the common
variance should be expected to have psychological import. This can be illustrated
through basic concepts from Campbell & Fiske’s (1959) multi-trait multi-method
matrix. When examining construct validity through multiple sources, monomethod
correlations are essentially always larger than their heteromethod counterparts.
Indeed, this comparison is most germane to calibrating the magnitude of methods
variance operating. It indicates that some portion of common variance running
through cognitiveability tests is methods varianceand, assuch, isconstruct irrelevant.
Dimensions emerging primarily from this aspect of common variance are bestviewed
as undesirable contaminants for the ultimate psychological solution (but not neces-
sarily for a mathematical solution aiming to account for all of the common variance
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
regardless of construct relevance). The basic idea, carried to its logical conclusion,
challenges the assumed desirability of accounting for all of the common variance in
a correlation matrix through factor-analytic techniques (when attempting to under-
stand the psychological structure underlying a representative collection of individual
differences measures). It suggests that only a fraction of the common variance is
construct relevant.
A factor solution accounting for 85% of the common variance among 50 variables
with a three-factor solution, and reinforced by a sharp ‘elbowed’ scree between
eigenvalues three and four would constitute, by many, a clear-cut, if not elegant,
triadic solution. But what if factors two and three provided little incremental validity
over factor one in the prediction of relevant (group membership or performance)
criteria, and none that held up under cross-validation (Lubinski & Dawis 1992,
Schmidt et al 1998)? Should we consider these dimensions psychologically important
too? Or, might these factors constitute nuisance variables—namely, reliable variance
akin to what Cook & Campbell (1979) have referred to as construct irrelevancies, or
‘systematic bias’ (Humphreys 1990), ‘constant error (Loevinger 1954), ‘‘system-
atic ambient noise’ (Lykken 1991), ‘crud’ (Meehl 1990), or, ‘methods variance’
(Campbell & Fiske 1959)? There is no a priori reason to assume that all of the
common variance in a correlation matrix is psychologically significant; but to deter-
mine whether it is (and to what extent) is an empirical question (cf. Thurstone
If the amount of common variance accounted for in a factor analysis need not
translate into the importance of a factor, what does? A proposalstemming from earlier
recommendations by Humphreys (1962) and McNemar (1964) has been reinstated
(Lubinski & Dawis 1992). Humphreys and McNemar stress the importance of incre-
mental validity. That is, when attempting to ascertain the number of dimensions
necessary to characterize cognitive abilities (orany domain of individual differences),
consider the amount of incremental validity gleaned over and beyond what is already
available. Given that the general factor accounts for about 50% of the common
variance among cognitive tests (coupled with the breadth and depth of its external
linkages), parsimony suggests that investigators begin here. By adding variables to
multiple regression equations (following the general factor), investigators can work
their way down the hierarchy of cognitive abilities and, as long as lower-tier dimen-
sions add incremental validity to what prior dimensions provided and these incre-
ments hold up on cross-validation (Lubinski & Dawis 1992), more molecular
dimensions thus achieve the status of psychologically significant parameters of indi-
viduality. Messick (1992:379) has communicated the same idea in a slightly different
way: ‘Because IQ is merely a way of scaling measures of general intelligence, the
burden of proof in claiming to move beyond IQ is to demonstrate empirically that
. . . test scores tap something more than or different from general intelligence by, for
example, demonstratingdifferentialcorrelates withother variables(which istheexter-
nal aspect of construct validity).’ Just as incremental validity is important when
appraising innovative measures (Dawis 1992, Lykken 1991, Sanders et al 1995), the
same holds for the dimensional products of factor analysis. Innovative measures and
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
variables worthy of scientific attention provide information not already available;
nonincremental sources of variance do not.
Group Differences When Jenkins & Paterson (1961) compiled their classic book
illustrating the historical development of psychological measurement and individual
differences, and searched ‘for a topic to serve as a model problem [they] quickly
settled on intelligence’ (1961:v). Then, like now, intellectual assessment was the
richest vein of differential psychology. Furthermore, not unlike today (Herrnstein &
Murray 1994), their preface stressed how controversial this area is. Campbell (1996)
provides an excellent contemporary overview (see also Coleman 1990–1991). Hum-
phreys (1995), moreover, has maintained that it is because of the magnitude of group
differences on ability measures, and the real world performances that these measures
are able to forecast, that differential psychology has been a neglected area in
At the apex as well as at the lower tiers of cognitive abilities, attention toward
group contrasts has arisen for several reasons. Before proceeding, however, the mag-
nitude of overlap between various groupings of human populations should be empha-
sized. In standard deviation units, the range within any given population (race, sex)
is many times the range between population means. One noteworthy achievement of
differential psychology is that it has moved human psychological appraisals from
crude nominal categories (group membership) to more refined ordinal and interval
measurement (continuous dimensions of human variation), and experimental proce-
dures for ratio measurement are underway (Deary 1996). As a result of these refine-
ments, all human populations have revealed exceptional talent (comparable ranges).
Ordinal and interval assessments of individual differences illuminate the diversity of
talent within all demographic groupings, which nominal scaling systems are ill
equipped to do.
Since the onset of psychometric inquiry, however, differences among various
racial groupings (sometimes reaching one standard deviation or slightly more) have
been both stubborn and consistent (Cronbach 1975; Jensen 1980, 1998). Furthermore,
the magnitude of these differences has been relatively stable even during periods of
converging educational opportunities (Gottfredson & Sharf 1988). Beyond this, it is
important to understand that, like demonstrable differences, seemingly minor differ-
ences in ability level (mean) and dispersion (variability) warrant critical scrutiny.
Collectively and individually, small group differences in level and dispersion fre-
quently create huge upper tail ratios when stringent cutting scores are implemented
(e.g. for select educational and training opportunities). Asian and Jewishpopulations,
for example, typically manifest superior test scores, relative to the generalpopulation,
and are overrepresented when stringent selection is applied to test scores. Feingold
(1995) presents data on sex differences and considers implications for group differ-
ences more generally. For example, meta-analytic reviews focus on level or aggre-
gating effect sizes (differences in standard deviation units), but groups may also differ
in variability, which meta-analyses typically do not address. Feingold highlights the
importance of examining both.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Sex Differences Most investigators concur on the conclusion that the sexes mani-
fest comparable means on general intelligence (Halpern 1992); yet, there is some
evidence for slightly greater male variability (Eysenck 1995, Jensen 1998, Lubinski
& Dawis 1992). With respect to level, Jensen (1998) has provided a particularly
detailed presentation of this topic, including an innovative methodology for arriving
at this conclusion. However, a number of investigators—including Jensen (1998),
using his new method—have reached a different consensus about sex differences in
strengths and relative weaknesses on specific abilities (Benbow 1988, Geary 1998,
Halpern 1996, Hedges & Nowell 1995, Stanley et al 1992). Females appear to excel
in certain verbal abilities, males in certain mathematical and spatial abilities. Hedges
& Nowell (1995) published probably the most compelling contemporary analysis on
this topic. They analyzed data from six large-scale studies collected between 1960
and 1992. Their analysis is important because, as they point out, many studies on
sex differences are based on nonrandom samples, whereas their probability samples
consisted of stratified random samples of U.S. populations. This study compiled data
from Project Talent, National Longitudinal Study of the High School Class of 1972,
National Longitudinal Study of Youth, High School and Beyond 1980, National
Educational Longitudinal Study 1988, and National Assessment of Educational Pro-
gress. Means, variances, and upper tail ratios .90% and .95% were computed.
Findings were consistent with other reports: Females tend to score higher on several
verbal/linguistic measures, while males score higher in certain quantitative and spa-
tial/mechanical abilities. Moreover, with respect to spatial/mechanicalabilities,males
display higher means and larger variances on nonverbal reasoning tests, which, again,
generate huge upper tail ratios. Hedges & Nowell (1995) discuss implications of
these findings for male/female proportions in math/science domains.
Race Differences Clearly, the most contentious area of contemporary research on
individual differences is found in Black/White contrasts (Gordon 1997). The most
noteworthy group difference in this regard is the approximately one standard devi-
ation difference on the general factor mean, with Whites scoring higher than Blacks.
There are other group differences as well. For example, Hispanic populations tend
to score intermediately between Blacks and Whites, whereas Asian and Jewish pop-
ulations are score slightly higher than Whites (Gottfredson 1997). Nevertheless,
Black/White contrasts have generated the best data (Humphreys 1988, 1991) and, by
far, the most attention (Campbell 1996). Over the years, these differences have moti-
vated intense study of test bias (especially underestimating the performance of under-
represented groups). Somebenchmarks are found in an AmericanPsychologist(1965)
special issue and two APA task force reports (Cleary et al 1975, Neisser et al 1996).
Jensen (1980) is still an excellent source on test bias. Given that these reports, com-
piled overfour decades, reached the sameconclusion foundin two NationalAcademy
of Science reports (Wigdor & Garner 1982, Hartigan & Wigdor 1989), an empirical
generalization can be ventured: Professionally developed general ability measuresdo
not underpredict performance of underrepresented groups.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Flynn Effect Observed scores on intelligence tests have been steadily rising cross-
culturally over this century. These raw-score increases on measures of general intel-
ligence have been labeled the ‘Flynn effect,’’ after the investigator who documented
their occurrence (Flynn 1999). Whether these increases reflect genuine gains in g is,
however, unclear. Increases can occur due to increases on a measure’s construct
relevant or construct irrelevant (nonerror unique) variance, or both. The problem is
complex and has generated considerable discussion (Neisser 1998). As yet, a final
answer is not available. However, evidence that changes are due, at least in part, to
construct irrelevant aspects of measuring tools is available.
Across various g indicators, the Flynn effect is positively correlated with the
amount of nonerror uniqueness. For example, gains on the Raven matrices aregreater
than gains on verbal reasoning composites of heterogeneous verbal tests, which, in
turn, aregreater than gains on broadly sampled tests of g (aggregatesofheterogeneous
collections of numerical, spatial, and verbal problems). The Raven matrices consist
of approximately 50% g variance, whereas heterogeneous collections of cognitive
tests aggregated to form a measure of g approach 85% (Lubinski & Humphreys
1997). (Broad verbal reasoning tests are intermediate.) Complexities are added by
considering that test scores have probably increased (especially at the lower end of
general intelligence) due to advances in medical care, dietary factors,andeducational
opportunities (Jensen 1998). Yet, at high levels of g, the gifted appear to havesuffered
some setbacks as a consequence of being deprived of developmentally appropriate
opportunities—a challenging curriculum at the appropriate time (Benbow & Stanley
1996). This topic deserves intense study for a number of reasons (Moffitt et al 1993,
Schaie 1996), one of which is especially noteworthy. Sorting out the complexities
involved in assessing dysgenic trends (Loehlin 1997, Lynn 1996, Williams & Ceci
1997) is predicatedon understanding thecausal determinantsof rawscorefluctuations
on measures of g.
Whatever these raw score gains are ultimately attributed to, they do not, as some
have indicated, appreciably detract from the construct validityof measures of g. Mean
gains on construct valid measures do not speak to changes in internal or external
covariance structure (Hunt 1995). Populations at contrasting levels of development,
for example, typically manifest the same covariance structure with respect to the trait
indicators under analysis (Rowe et al 1994, 1995).
Horizontal and Vertical Inquiry The idea that constructs may be analyzed at dif-
ferent levels of analysis is well known. For example, Embretson (1983) has contrib-
uted an important distinction to the construct validation process. She suggests a
parsing of the nomological network into two regions: construct representation versus
nomothethic span. The latter denotes the network of empirical relationshipsobserved
with measures at the behavioral level, whereas the former is aimed at underlying
processes or mechanisms responsible for generating these phenotypicmanifestations.
Jensen (1998) has likewise pointed to two lines of empirical research on g, one
vertical and the other horizontal. Both lines dovetail with MacCorquodale & Meehl’s
(1948) distinction between hypothetical constructs (HC) and intervening variables
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
(IV). Although both concepts carry denotative and explicative meaning, hypothetical
constructs stress explanation, whereas intervening variables are more restricted to
description. Spearman’s (1927) initial formulation of g as ‘mental energy,’ was a
HC, whereas the parameters describing the functional relationships between concep-
tually equivalent measures of g and external criteria were IVs. When cross-
disciplinary linkages are drawn, the HCs of one discipline can become the IVs of
another, but that discussion is beyond the scope of this review (see Maxwell 1961).
What is important for our purposes is that connecting threads have been established
between g and several biological phenomena. Ultimately, the causal paths of these
interrelationships will need to be traced.
Pooling studies of a variety of kinship correlates on IQ (e.g. MZ and DZ twins
reared together and apart and a variety of adoption designs), the heritability ofgeneral
intelligence in industrialized nations has been estimated to be between 60%–80%
(Hetherington et al 1994, McGue & Bouchard 1998). Using magnetic resonance
imaging (MRI) technology, brain size controlled for body weight covaries 0.30–0.40
with general intelligence (Bouchard 1999, Jensen 1998, Willerman et al 1991). Haier
(1993) reports that glucose metabolism is related to problem-solving behavior, and
that the gifted appear to engage in more efficient problem solving behavior that
expends less energy. Also, highly intellectually gifted individuals evince enhanced
right hemispheric functioning (Haier & Benbow 1995, O’Boyle et al 1995). The
complexity of electroencephalograph (EEG) waves is positively correlated with g, as
are amplitude and latency of the average evoked potential (AEP) (Lutzenberger et al
1992). Some investigators have determined the negative correlation between g and
inspection times, assessed through chronometric procedures, to be a biological phe-
nomenon (Deary 1996). Anderson (1993) suggested that dendritic arborization is
correlated with g. Although Anderson typically examines histological data across
groups of individuals with documented IQ differences, he also has conducted an
intriguing case study involving Albert Einstein’s brain (Anderson & Harvey 1996).
In contrast to a control group of autopsied men, the frontal cortex of Einstein’s brain
possesses a significantly greater neuronal density (cf. Diamond et al 1985). Given
this, the following was perhaps inevitable: A multidisciplinary team appears to have
uncovered a DNA marker associated with g (Chorney et al 1998).
It is virtually guaranteed that more biological linkages will be made to g (Vernon
1993). Like those already uncovered, they are likely to be heterogeneous and to vary
in strength of association with g. These biological phenomena are in no way mutually
exclusive and can be complementary to one another. Some may transcend phylo-
genetic orders and thus enhance our comparative understanding of general learning
phenomena (Anderson 1993, 1994a,b, 1995). One provocative conjecture is the mye-
lination hypothesis (Miller 1994): Individual differences in cognitive efficiency are
a function of individual differences in the amount of myelin (the fatty substance
coating the neurons).
Proximal and Ultimate Examinations of g Given the biological connections to g
(Vernon 1993), some researchers have gone beyond these proximal associations to
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
speculate on their ultimate evolutionary basis. Bouchard et al (1996) have revised
experience producing drives (EPD) theory, which speaks to human intellectualdevel-
opment. EPD theory-revised is a modification of an earlier formulation by Hayes
(1962), a comparative psychologist and pioneer in language and socialization capa-
bilities of nonhuman primates. His idea was that, like all organisms, humans were
designed to do something, and that they possess EPDs to facilitate ability and skill
acquisition through inherited dispositions that motivate individuals toward particular
kinds of experiences and developmental opportunities. Such evolutionary selective
sensitivities can operate, moreover, in a wide range of functionally equivalent envi-
ronments (which fits with the idea that humans evolved in a highly fluctuating
Other investigators have sought a synthesis between evolutionary psychology and
chronometric procedures formeasuring inspectiontime (Deary1996).Inspectiontime
is a measure of speed of perceptual discrimination on ‘‘simple’ elementary cognitive
tasks (responses to stimulus configurations that typically take less than one second
for average adults to perform with essentially zero errors).Theoretically,performance
on elementary cognitive tasks indexes the time course of information processing in
the nervous system. There are a variety of technical measurement issues surrounding
this area of research, but it does appear that the temporal dynamics of performance
on elementary cognitive tasks covaries negatively with g (faster processing is asso-
ciated with higher g levels). Washburn & Rumbaugh (1997) used inspection time
measures to successfully assess individual differences in cognitive sophistication
among nonhuman primates.
This intriguing line of research might provide a vehicle for comparative psycho-
logical inquiry into the biological underpinnings of general cognitive sophistication,
comparable with what the sign-language modality fostered for language learning in
nonhuman primates. This is certainly not far-fetched. Investigators have long
remarked on the range of individual differences within primate conspecifics. For
example, Premack (1983:125) noted in his discussion of language versus
nonlanguage-trained groups of chimpanzees, ‘Although chimpanzees vary in intel-
ligence, wehave unfortunately never had any control over this factor, having to accept
all animals that are sent to us. We have, therefore, had both gifted and nongifted
animals in each group. Sarah is a bright animal by any standard, but so is Jessie, one
of the non-language trained animals. The groups are also comparable at the other
end of the continuum, Peony’s negative gifts being well matched by those of Luvy.’
Individual differences in processing stimulus equivalency (verbal/symbolic) rela-
tionships have been postulated by some experimentalists to index generalintelligence
(Sidman 1986). If such individual differences are ultimately linked to individual
differences in central nervous system microstructure within and between the primate
order, and these in turn are linked to observations like Premack’s ‘teacher ratings,’
all of the ingredients are assimilated for advancing primate comparative psychology.
The language-communicative performances now routinely displayed by chimpanzees
and, especially, pigmy chimpanzees are truly remarkable (Savage-Rumbaugh et al
1993, Savage-Rumbaugh & Lewin 1994, Wasserman 1993). They encompass sign-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
language reports of emotional states and conspecific tutoring (Lubinski & Thompson
1993). Savage-Rumbaugh et al (1993) have connected these nonhuman primate find-
ings with those from child language-development research. Will primate
comparative-examinations someday provide clues to human individuality? If indi-
vidual differences in acquiring cognitive skills could be linked to more fundamental
biological mechanisms (like the phenomena discussed above), we might have an
especially powerful lens through which to view common phylogenetic processes
involved in cognitive development. Research developments on this front will be
interesting to follow. Perhaps they might even obviate Wilson’s (1998:184) recently
expressed concern: ‘[S]ocial scientists as a whole have paid little attention to the
foundations of human nature, and they have had almost no interest in its deep
Interests have played a large role in differential psychology since the 1920s. Longi-
tudinal inquiry comprising both temporal stability analyses (reliability) and forecasts
of occupational group-membership (validity) established these measures as among
the most important in applied psychology (Harmon et al 1994, Savickas & Spokane
2000). Going beyond adult populations, assessments conducted at more develop-
mentally inchoate stages revealed that interests begin to crystallize during adoles-
cence. They can forecast antecedents to occupational choice (e.g. college major) and,
as such, serve as important tools in educational contexts (Dawis 1992). An especially
critical aspect of these longitudinal studies is their incremental validity (Austin &
Hanisch 1990, Humphreys et al 1993): Interests contribute important information
relative to abilities. Further, the validity generalization of the unique contribution of
interests has been extended to special populations. For example, Achter et al 1999
recently reported that age 13 interest assessments, among intellectually gifted stu-
dents, forecast educational choice (four-year degree) over a 10-year temporal gap
and add incremental validity to ability assessments.These arescientificallysignificant
tools, which (like cognitive abilities) are predictive of a broad spectrum of criteria
ranging from (core) educational/vocational settings to (more peripheral) activities in
everyday life (Dawis 1992, Hogan et al 1996).
Although early research on interests was atheoretical, using empirical keying
(group contrast) methodology to literally form a scale for every occupation, over the
past few decades the push for deriving a general model of interest dimensions has
intensified. A hexagonal structure of interest dimensions emerged (Holland 1996),
which is helpful for understanding how people approach and operate within learning
and work environments. Holland’s model is defined by six general interest themes
known as RIASEC: realistic [working with things and gadgets], investigative [sci-
entific pursuits], artistic [aesthetic pursuits and opportunities for self-expression],
social [people contact and helping professions], enterprising[corporateenvironments:
buying, marketing, selling], and conventional [office practices and well-structured
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
While RIASEC is not embraced by everyone (Gati 1991), it is the most popular
model available and, like the hierarchical organization of human abilities and per-
sonality’s five-factor model (discussed below), innovative frameworks will need to
be measured against it. RIASEC has emerged repeatedly in large samples (Rounds
& Tracey 1993, Tracey & Rounds 1993), and its generalizability has held up cross-
culturally (Day & Rounds 1998). RIASEC is organized around Holland’s (1996)
calculus assumption, which states that adjacent themes are most highly correlated,
and opposite themes leastcorrelated. Prediger (1982)has argued that Holland’smodel
can be reduced to two relatively independent dimensions: people versus things, and
data versus ideas. The former runs from Holland’s social (people) to realistic (things)
themes, whereas the latter runs perpendicular to people versus things splitting enter-
prising and conventional (data) and artistic and investigative (ideas). Predigers two-
dimensional model fits, as he maintains, within RIASEC, but most investigators feel
that the parsimony achieved through this two-dimensional collapse does not offset
the richness that is lost. Nevertheless, Predigers work is important.
While the sexes do not appear to differ appreciably on data versus ideas, they
routinely differ by a full standard deviation on people versus things (females tend to
gravitate toward the former, males toward the latter). For example, Lippa (personal
communication) computed all the effect sizes (female minus male) in his interesting
multi-study article on the people versus things dimension (Lippa 1998). For all three
studies, effect sizes were $ 1.20 on people versus things. This is typical, reflecting
perhaps the largest of all sex differences on major psychological dimensions.
To be sure, there are more specific interest dimensions beyond RIASEC that carry
psychologically significant import [religiosity being a noteworthy example (Waller
et al 1990; see also Harmon et al 1994, Savickas & Spokane 2000)]. Nevertheless,
RIASEC constitutes a cogent outline of this important arena of psychological diver-
sity. Interestingly, like the constituents found in the hierarchy of human cognitive
abilities, antecedents to RIASEC may be traced over many decades. RIASEC exem-
plifies how, through careful research (including cross-cultural inquiry), the nature and
organization of an important domain can be successively clarified. Guilford (1954),
for example, examined and discussed very similar structures: mechanical, scientific,
aesthetic expression, social welfare, business, and clerical. Holland’s (1996) model
stands on the shoulders of much that has gone before it.
As in our earlier discussion of range truncation (Olympic athletes), the most
important dimensions for steering individuals to specific opportunities and settings
are often uniformly high. With respect to forecasting continuouswork-relatedcriteria,
range truncation among incumbents may generate equivocal empirical findings. So,
with respect to predicting job satisfaction,
A number of explanations can be advanced to account for the mixed results
found for interests. If . . . subjects of follow-up studies were the survivors of a
selection process, one might infer that in this process, the dissatisfied would
have tended to leave, whereas the satisfied—and satisfactory—would have
tended to remain. The restriction of range that would result could contribute to
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
the lowering of the true correlations. Unfortunately, the means and standard
deviations of variables frequently go unreported so that a straightforward
check on this simple explanation is often thwarted (Dawis 1991:851–52).
Indeed, psychological research would be more informative ifit routinelydescribed
samples with means and standard deviations on major dimensions of abilities, inter-
ests, and personality for purposes of ecological validity. Doing so would reveal that
some perplexing findings stem from nonrepresentative sampling.
A consensus has emerged on the major personality dimensions, but it is more opaque
than for cognitive abilities and interests. Although the dimensions reviewed here
appear relatively independent of abilities and interests (Ackerman 1996, Ackerman
& Heggstad 1997), it is something of a misnomer to reserve the term ‘personality’
for them. One could argue that abilities and interests are salient aspects ofpersonality.
(Cattell [1971], for example, thought so.) Like garden-variety personality measures,
abilities and interests are enduring features of one’s psychological make-up (Bou-
chard 1997, Rowe 1994, Scarr 1996). A complete understanding of one’s character
or reputation (Hogan et al 1996) is incomprehensible without them. Thus, while
thinking about personality, it is important to keep in mind the wisdom of the great
counseling psychologist, Roe, whose words are as true today as they were when she
published them:
I have become more and more convinced that the role of occupation in the
life of the individual has much broader psychological importance than has
generally been appreciated. I believe that psychological theory could profit
greatly from the kinds of satisfactions that can be found in work. This is as
true for developmental theory as it is for motivational theory ...Ifonewishes
to understand the total psychology of any person, it is at least as important to
understand . . . occupational behavior as it is to understand . . . sexual behav-
ior. (They are not unrelated.) . . . The fact is, of course, that one can start with
any facet of human behavior and work through it to the ‘total personality’
With this in mind, and acknowledging that some of the best contemporary evi-
dence for the scientific significance of broad dimensions of personality is found in
predicting vocational criteria (Hogan et al 1996, Hough 1997), an examination of
recent advances in personality follows.
The Big Five The intensity of work on the dimensionality of personality during
the 1980s and 1990s is comparable to that of validity generalization inabilities during
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
the 1970s and 1980s. This work has been productive. For the most part, examinations
of personality have followed the ‘lexical’ approach suggested by Galton (1884),
namely, that important dimensions of human behavior will be encoded in natural
language for economy of thought. Hence, the dictionary, when systematically exam-
ined, should prove an invaluable source for identifying personality characteristics
(Allport & Odbert 1936). A working model of descriptors from the dictionary is
available: the ‘big five’ (McCrae & Costa 1997) [but see Block’s (1995a) ‘con-
trarian view’ and replies from Costa & McCrae (1995) and Goldberg & Saucier
(1995), and Block’s (1995b) rejoinder].
Labels for the big five have varied, but include Extraversion (surgency, positive
emotionality), Neuroticism (anxiety, negative emotionality), Agreeableness (antago-
nism reversed), Conscientiousness (will to achieve), and Openness(culture, intellect).
Like abilities and interests, these five generic factors have a long history in psychol-
ogy. For years, they were simply referred to as ‘‘Norman’s five,’’ followingNorman’s
(1963) seminal treatment. However, the same dimensions surfaced at least 50 years
ago (Fiske 1949) and were subsequently supported by large-scale analysis of military
samples (cf Tupes & Christal 1992, initially published in 1961). It should be noted
that Eysenck (1995) felt that conscientiousness and agreeableness can be combined
to form his psychoticism (reversed) dimension, thus supporting his preference for a
three-dimensional model (the ‘big three’’): extraversion, neuroticism, and psychoti-
The Big Seven Waller (1999) has traced decisions concerning the item pool that
Allport & Odbert (1936), Cattell, and Norman considered relevant to ‘authentic
traits.’’ Subsequent investigators who consulted Allport and Odbert’s categoricallists
apparently excluded practically all evaluative terms from efforts to develop scales of
the basic dimensionality of personality. Terms such as special, important, immoral,
disloyal, and nasty were not routinely examined in attempts at mapping personality.
For several years now, Tellegen and Waller have studied evaluative terms by
systematically sampling from the dictionary (Tellegen 1993, Tellegen & Waller 2000,
Waller 1999). They have a questionnaire purporting to assess evaluative traits and
the Big Five dimensions (Tellegen et al 1991). Their analysis appears to warrant
seven dimensions: the big five and Positive and Negative Valence. Positive Valence
depicts a dimension with positive loadings on ‘outstanding,’ ‘first-rate,’ ‘excel-
lent,’’ ‘‘remarkable,’’ which form a continuum from ordinary-to-exceptional, or com-
mon-to-impressive. Negative Valence is captured by terms such as ‘cruel,’ ‘evil,’
‘wicked,’ and ‘sickening,’ which portray a continuum from worthy-to-evil, or
decent-to-awful. These two dimensions have held up cross-culturally (Almagor et al
1995, Benet & Waller 1995). Becausethese highly evaluativeterms were prematurely
jettisoned from empirical analyses until recently, there has not been an opportunity
to demonstrate their importance.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Interpretation and Future Directions Tellegen (1993) has suggested that major
dimensions of personality have adaptability import. Individual differences reflect
one’s ‘preparedness’ or ‘tuning’ to affordances in the social landscape (see also
Snow 1991). Tellegen’s (1993) big seven studies motivated him to adopt somewhat
different labels (with the following interpretations). ‘Positive and Negative Valence
reflect primal readiness to encode power and evilness; Positive Emotionality and
Negative Emotionality reflect built-in responsiveness to signals of emotion and emo-
tional-temperamental dispositions; and Dependability, Agreeability, and Convention-
ality (vs. Unusualness) reflect protoscientific propensities to encode a person’s
predictability, controllability, and comprehensibility, respectively’ (1993:126). Tel-
legen also has advanced the idea that we consider these ‘folk concepts’ as distin-
guished from psychological concepts advanced to describe or explain psychological
phenomena and processes. Recent advances have placed personological inquiry into
the broader context of evolutionary theory (Hogan & Hogan 2000).
Hogan et al (1996) have recently cautioned against examiningpersonality dimensions
individually because the manner in which each operates depends on the full con-
stellation of personal characteristics. Two extroverts will operate quite differently, for
example, if their standings on conscientiousness are diametrically opposed. The point
is well-taken, but the evidence indicates that we should move beyond Hogan et al’s
(1996) recommendation (sound as it is) and intermingle cross-domain attributes.Like
contrasting constellations of personality attributes, similar interest and abilitypatterns
often produce markedly different phenotypes as a result of differences on dimensions
from other classes. The paths traveled by two spatially gifted students are likely to
be quite distinct if, for example, they occupy contrasting locations on ‘‘people versus
things.’’ Assuming that more comprehensive assessments willenhancepsychological
theory and practice, some approaches that go beyond domain-constrained treatments
Intellectual Development
Ackerman (1996, Ackerman & Heggstad 1997) has proposed an intriguing model of
adult intellectual development that orchestrates abilities as process, personality, and
interest dimensions simultaneously to describe changes in cognitive content and pro-
cesses throughout the life span. Content denotes the pedagogical aspects of learning
(knowledge), whereas process is more restricted to power of intellect [or e.g. working
memory capacity(Carpenter et al 1990,Kyllonen & Christal 1990), perhaps a modern
conceptualization of Spearman’s (1927) mental energy]. Ackerman’s theory is called
PPIK, for intelligence-as-process, -personality, -interests, and -knowledge. Interests
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
and personality attributes channel the development of knowledge structures down
different paths, for example, CP Snow’s (1967) two intellectual cultures, while intel-
ligence-as-process determines the complexity and density of the knowledge assimi-
lated. Ackerman’s approach is reminiscent of Cattell’s (1971) early formulation of
investment theory, where fluid abilitiesare invested in the development of crystallized
abilities as afunction of nonintellectual personal attributes. Intellectualbodiesdevelop
from a common multidimensional core (abilities, interests, and personality) that are
seemingly quite generic cross-culturally (Carroll 1993, Day &Rounds 1998, McCrae
& Costa 1997).
This model provides an insightful basis for uncovering why individuals withsimi-
lar cognitive profiles can, and frequently do, vary widely in their knowledge base or
‘crystallized abilities.’’ Ackerman (1996, Ackerman & Heggstad1997) hascompiled
ability/interest, ability/personality, and interest/personality correlatesto support PPIK.
Analysis has distilled four across-attribute (ability/interest/personality) trait com-
plexes. They are social, clerical/conventional, science/math, and intellectual/cultural.
Intellectual/cultural, for example, reflects light correlations between verbal abilityand
aesthetic and investigative interests, whereas science/math reflects light correlations
between math/spatial abilities and realistic, investigative and social (reversed) inter-
ests. The psychological import behind these trait complexes is similar to Snow’s
(1991) aptitude complexes (ability ` interest constellations for classifying educa-
tional treatments), and Dawis & Lofquist’s (1984) taxons (ability ` preference con-
stellations for conceptualizing transactions between individuals and work
environments; see below).
PPIK might be especially relevant to contexts where knowledge is more important
than intellectual processing abilities for predicting performance (Ericsson 1996).
Examinations of expert performance (Rolfhus & Ackerman 1996), for example, have
often revealed that the greatest difference between experts and nonexperts is in the
richness and depth of the knowledge structures of the former. Ackerman also has
developed a typical intellectual engagement (TIE) measure for assessing how much
an individual is likely to invest in developing hisor her intellectualabilities.However,
this measure tends to covary more deeply with humanistic than scientific knowledge
domains (Ackerman 1996). Therefore, multiple TIE measures might be required to
capture the multiple motives involved in developing intellect. Perhaps distinct TIE
should be developed for each PPIK trait complex. Given that the current TIE is
primarily relevant to the humanities, a more descriptively apt label might be ‘TIE-
verbal/humanistic’ (for trait complex: intellectual/cultural). A TIE measure focusing
more on nonverbal ideation might better forecast development in more technical
domains: ‘TIE-science/math’’ (for trait complex: science/math).
What one knows (knowledge) and how sophisticated one is at manipulating what
one knows (thinking) are ostensibly two different things. Yet, with respect to mea-
surement operations, content andprocess (knowledge and thinking) always have been
inextricably intertwined (Roznowski 1987). As Ackerman (1996:245) remarks:
‘[A]n individual can strive for breadth of knowledge or depth of knowledge, but
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
there is a trade off between these two orientations. Only the most exceptional intel-
lectual talent will allow for high levels of knowledge domain depth and breadth.’
Does ‘exceptional intellectual talent’ primarily stem from one dimension or two?
Perhaps breadth and depth combine to map g in a manner analogous to area; or
perhaps speed should be added to assess this central dimension akin to measuring
volume? It seems as though we always return to Spearman’s g in one way or
another—a dominant dimension whose scientific significance is central. These obser-
vations notwithstanding, PPIK clearly takes an important step forward in conceptu-
alizing the nature of intellectual development.
Vocational Adjustment
Are you able to do it? Are you happy doing it? Throughout most of this century, in
one form or another, vocational psychologists have been asking clients these two
questions. Often data were collected to help clients whose reactions were initially
uncertain or unclear. Dawis & Lofquist (1984, Lofquist & Dawis 1991) developed
a system to conceptualize vocational adjustment and counseling, the theory of work
adjustment (TWA). TWA ishelpful for understanding whyabilities and interestsshow
incremental validity relative to each other in learning and work settings. Katzell
(1994), reviewing volumes one through three of the Handbook of Industrial and
Organizational Psychology, used TWA as an integrative framework to synthesize
research literature in I/O psychology. TWA has been applied to designing learning
environments throughout the life span (Lubinski & Benbow 2000), and the Journal
of Vocational Behavior (1993) has a special issue on TWA.
TWA is predicated on two dimensions: satisfaction and satisfactoriness. Satisfac-
tion is a function of the correspondence between a person’s preferences (needs, inter-
ests, and values) and the rewards offered in a particular occupational setting or career
path. Satisfactoriness is determined by the correspondence between one’s abilities
and the competency requirements needed for effective performance in a given occu-
pation. Equal emphasis is placed on assessing the individual and the environment;
both are assessed in commensurate terms; and, when ahigh degree ofcorrespondence
is achieved across both dimensions (i.e. the individual is feeling satisfied and is
performing satisfactorily), a symbiotic relationship develops to sustain the joint
person-environment interaction. When satisfaction is high but satisfactoriness is low,
the environment is likely to terminate the relationship; when the inverse occurs, the
person is more likely to break off the relationship.
TWA uses the term ‘taxon’’—akin to Ackerman’s ‘‘trait complexes’’ and Snow’s
‘aptitude complexes’’—to depict ability-preference constellations related to differ-
ential performance and enjoyment outcomes within the world of work. Supporting
data are found in two books (Dawis & Lofquist 1984, Lofquist & Dawis 1991), as
well as throughout the applied psychological literature examining how ability/
preference constellations fit into relatively well-defined ecological settings (e.g. edu-
cational tracks, military classification systems, occupations). As Katzell (1994:13)
noted, ‘[a]though not derived specifically from the theory, there have been many
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
practical applications of parts of it in industry, such as the prediction of turnover from
job satisfaction and the matching of ability with job requirements to predict perfor-
mance.’ Support for TWAs validity is seen in positive results for Schneider et al’s
(1995) attraction-selection-attrition (ASA) model and the gravitational hypothesis
(Dunnette 1998, Wilk et al 1995, Wilk & Sackett 1996). The basic idea is that people
select environments congenial to their personal attributes and style of life and migrate
from those that are not good fits.
Work Performance
The opening sentence of Schmidt & Hunters (1998:262) review of 85 years of
research on selection methods in personnel psychology is consistent with the desire
of appliedpsychologists to uncover longitudinallystable dimensions: ‘From thepoint
of view of practical value, the most important property of a personnel assessment
method is the predictive validity.’
Work performance is an important area of applied psychology, not only in terms
of a society’s economic well-being in internationally competitive markets, but also
in terms of the emotional and physical well-being of citizens within a society (Hunter
& Schmidt 1996, Schmidt & Hunter 1998). For a poignant example, see Hunter &
Schmidt’s (1996) powerful and compelling illustration of factors associated with the
time it takes to catch a rapist (measured in number of crimes committed). Huge
individual differences are found between competent and excellent police officers, in
the effectiveness of their work and how expeditiously justice is served. When con-
sulting with legal officials, Hunter & Schmidt point out that lawyersfrequently appre-
ciate individual differences between competent and poor workers, but they have a
rather poor appreciation of differences between competent and exceptional workers.
Laypersons are unaware of the two primary ways to assess individual differences
in performance: dollar value of output and percent of mean output. At minimum, the
standard deviation of the dollar value of output across individuals has been found to
be 40% of the mean salary of the job. Hence, if the average salary for a job is
$50,000, the standard deviation of employees’ dollar-value output is $20,000. The
difference, therefore, between above-average workers (e.g. one standard deviation
above the mean) and below-average workers (e.g. one standard deviation below the
mean) would be: $70,000 1 $30,000 4 $40,000. Work performance measured as
a percentage of mean output would be estimated as follows: An employee’s output
would be divided by the output of workers at the 50th percentile and then multiplied
by 100. The standard deviation of output as a percentage of average output is mod-
erated by job level. Schmidt & Hunters (1998) review found that percentage to be
around 19% for unskilled and semi-skilled jobs, 32% for skilled jobs, and 48% for
managerial and professional jobs. There is an old saying in applied psychology: For
a difference to be a difference it must make a difference.
In view of these important differences, uncovering predictors to model work per-
formance has attracted much attention. This was anticipated in Lerners (1983) dis-
cussion of ‘human capital.’ Research has added to validity generalization studies of
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
the past two decades by combining personality measures with abilities. Conscien-
tiousness, for example, adds incremental validity with probably as much breadth (but
not quite as much depth) as general ability measures to predictions for many occu-
pations. The longstanding belief that personality measures do not contribute to indi-
vidual differences in work performance is not true. Increments for personality
measures typically range between 0.05 and 0.15, which may seem small when con-
trasted with what ability constructs offer, but their economic and social gains are
huge. Moreover, the troubling group differences on abilities reviewed earlier are not
found on these measures, so personality measurement has the potential to minimize
adverse impact. There are, however, differences in opinion on how best to carve up
personality for predicting work performance (Hough 1997). Nevertheless, there is
widespread agreement that increments in predicted performance beyond ability are
achievable through personality assessment. These increments are especially evident
when studying peak performance.
Creativity and Eminence
A number of dimensions relevant to creativity have been identified. Interestingly,
they are similar to Galton’s (1869) necessary ingredients for eminence. Investigators
operating within frameworks distinct from differential psychology have confirmed
many of these (Gardner 1993). A deeper appreciation of this area is gleaned by
combining the differential psychology of Eysenck (1995) with the work of Gardner
(1993) and Simonton (1990, 1994). These treatments are not incompatible and, in
many respects, the latter two attach idiographic flesh to the normative skeleton out-
lined by Eysenck (1995). They also enlarge classics such as Roe’s (1953) The Making
of a Scientist and Zuckerman’s (1977) Scientific Elite.
Galton defined genius (the ultimate label for one’s track record of creative accom-
plishments leading to eminence) in terms of reputation: ‘those qualities of intellect
and disposition, which urge . . . acts that lead to reputation, I do not mean capacity
without zeal nor zeal without capacity, nor even a combination of both of them
without an adequate power of doing a great deal of very laborious work. But I mean
a nature which, when left to itself, will, urged by an internal stimulus, climb the path
that leads to eminence, and has the strength to reach the summit—one which, if
hindered or thwarted, will fret and strive until the hindrance is overcome . . .
For criterion measurement, Eysenck (1995) and Simonton (1990, 1994) have
adopted Galton’s view for calibrating eminence. In Simonton’s (1990) investigations
into the psychometric properties of reputation assessments (using informed judges),
he has reported internal consistency reliability coefficients .0.85 for artistic distinc-
tion, philosophical eminence, and scientific fame.
The dispositionalpackage that Galtonoutlined is in agreementwith modern views,
although Galton went too far in attributing eminence almost exclusively to personal
attributes. Today, spectacular forms of creativity, like lesser forms, are seen as con-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
fluences of endogenous and exogenousdeterminants, rather than primarily theformer.
Cultural factors and the zeitgeist play critical roles.
What attributes predict eminence? The personal attributes of individuals at the top
of their respective domains include the anticipated (ability ` interest) constellations
(aptitude complexes, trait complexes, and taxons) that distinguish individuals in their
chosen domain or profession from the general population. However, more intense
abilities are characteristic (and more is better) (Benbow 1992). For example, extraor-
dinary engineers and physical scientists possess pronounced quantitative-spatial abil-
ities and interests in investigative and realistic pursuits, whereas humanists possess
higher verbal abilities, relative to nonverbal abilities, and preferences for artistic and
social arenas. Yet, what appears to move the highly creative apart from their peers
is their passion for work. They are exceptional in their industriousness and perse-
verance; they tend to be almost myopically fixated on work. This is something well-
known among academic scientists who train academic scientists(cf Wilson 1998:56).
(Edison’s familiar 1% inspiration 99% perspiration also comes to mind.) The sheer
amount of time devoted to their area of excellence is one of the most exceptional
things about them. Zuckerman’s (1977) account of the extraordinary efforts that
Nobel Laureates displayed to reach the right teachers (who were almost always
Laureates themselves) supports this.
On the other hand, some antecedents contributing to theenormous energyreserves
of certain individuals are not necessarily positive. For example, Jamison (1993) has
observed a higher incidence of manic-depressive disorders among creative writers
and artists. Jensen (1996) has discussed other endogenous factors pertinent to cortical
stimulation, for example, blood serum urate (a cortical stimulant) level (SUL). Inter-
estingly, SUL covaries positively with achievement. Eysenck (1995) focuses onother
neurochemical underpinnings posited to give rise to ‘zeal’ (Galton 1869).
In part because of the intensity with which these individuals approach their work,
the highly creative, as a group, are also known to be difficult in interpersonal rela-
tionships, socially harsh, and abrasive. Gardner (1993) has discussed the ‘casualties’
surrounding these individuals as they steadfastly focus on their work to the exclusion
of other aspects of life. He discusses the ‘mixed blessings’ associated with being
close to suchindividuals. This supportsEysenck’s(1995) view that the highlycreative
are, on average, high on trait psychoticism (or conscientiousness ` agreeableness in
reverse). Following Eysenck, this, among other things, enables the highly creative to
look at things quite differently (unconventionally).
What appears to draw individuals toward particular environments, people, and
opportunities is, in part, the personal attributes that they possess; but once in these
arenas, what actually happens is contingent on opportunity. It might be helpful to
construe dispositional antecedents to exceptional forms of creativity as ‘emergenic
phenomena’’ (Lykken et al 1992), namely, the proper configuration of personalattrib-
utes [including the psychological endurance necessary for developing and maintain-
ing exceptional performance (Ericsson 1996)]. When such constellations find
supportive environments, then, and only then, does Galton’s depiction hold. Jensen
(1996) maintains that: genius 4 high ability 2 high productivity 2 high creativity.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Underpinning this equation, ability 4 information processing efficiency,productivity
4 endogenous cortical stimulation, and creativity 4 trait psychoticism (unconven-
tional ideation). This suggests kinds of inquiry that must at least be entertained for
understanding how products that change culture develop and, ultimately, how such
achievements are best facilitated, as well as inadvertently suppressed.
When Cronbach & Meehl (1955) introduced construct validation, they exemplified
the process by compiling a heterogeneous collection of findings all related to the
psychopathic deviate (Pd) scale of the Minnesota Multiphasic Personality Inventory.
How, they asked, could a scale developed to isolate criminals and delinquents from
the general population also reveal elevated scores for: Broadway actors, high-school
dropouts, deer hunters who accidentally shoot people, police officers, and nurseswho
were rated by their supervisors as not especially afraid of psychotic patients? (Note
this was before wide use of psychoactive drugs, when patients routinely experienced
frightening psychotic episodes.) Pd also covaries negatively with trustworthiness rat-
ings. What underlying construct representation could possibly support thisnomothetic
Two years later, Lykken (1957) published positive findings for what these groups
have in common: Relatively speaking, they are fearless or in possession of a ‘low
anxiety IQ.’’ Using a Pavlovian paradigm, Lykken showed that, as a group, hardened
criminals, when contrasted with random samples of inmates, were ‘retarded’’ when
it came to developing conditioned responses to neutral stimuli paired with shock.
This has been replicated and studied in several laboratories, albeit with somewhat
different labels and measures: ‘‘socialization’’ (reversed), ‘‘danger seeking,’ or ‘sen-
sation seeking’ (Wilson & Herrnstein 1985, Lykken 1995).
As Lykken (1995) points out, however, being fearless does not prescribe a partic-
ular developmental path. This is the stock from which astronauts, poised law enforce-
ment officials, firefighters on elite rescue teams, war heroes, and fighter pilots are
grown. When coupled with other attributes and opportunities, being fearless can be
an asset. However, it can also be a liability because it makes conventional sociali-
zation procedures difficult. For instance, when low fearfulness is combined with
agreeableness ` conscientiousness (reversed), a 75–90 IQ range, mesomorphic body
build, and reared in an abusive crime-ridden environment, a high-risk liability
emerges (Lykken 1995, Wilson & Herrnstein 1985). One of the handicaps faced by
individuals within lower IQ ranges is a limited temporal horizon, a deficit in fore-
seeing temporally remote consequences of actions.
Fortunately, however, if Lykken (1995) is correct, a ‘type like’ psychopath is
relatively infrequent, relative to the proportion of individuals engaged in criminal
behavior. He suggests that most criminal behavior stems from a larger group of
individuals—sociopaths who, withproper parenting, could havebeen socializedaway
from a life of crime. He adds that, while the behavior genetic data are compelling
(for the major individual-differences dimensions ofhis model), typical twinand adop-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
tion studiesdo not includefamilies deeply enmeshed inillegal activities. Thebehavior
genetic studies are restricted to environmental ranges not abnormally deviant from
the population norm. Lykken (1997) argues, however, that interventions are most
likely to be effective in these maladaptive environments (but see Rowe 1997).
Health Risk Behavior
‘If public health officials understood the characteristic behaviors, thoughts, and feel-
ings of those young persons who engage in health-risk behaviors, they could be in
a better position to design health campaigns and educational programs that would
appeal to their target audience’ (Caspi et al 1997:1053). Repeatedly, longitudinal
inquiry has uncovered the significance of individual differences in channeling the
development of harmful maladies not only to the individual at risk (Gordon 1997,
Lubinski & Humphreys 1997, Schaie 1996) but to others occupying their purview.
Like contemporary treatments of creative achievement and crime, contemporary
discussions of health risk behaviors are related to‘delayof gratification’phenomena,
which can have multiple (ability ` personological) antecedents. Outcomes emanat-
ing from both wise and unwise actions, and conscientious versus risky behavior, are
often not precipitous. They frequently develop slowly over time to result in a life
threatening condition, an ostensibly discrete arrest, or a seemingly effortless master-
piece. Short-lived behavioral episodes are often products of years of development.
Just as the development of excellence is in part traceable to comprehending the
temporally remote consequences of immediate practice, aspects of maladaptive
behavior are due to a limited temporal horizon.
Life Span Development
Scarr (1992, 1996; Scarr & McCartney 1983) has drawn on three kinds of genotype-
environment (GE) correlations distinguished by behavioral geneticists—active, pas-
sive, and reactive—to build a developmental theory of individuality. Her formulation
builds on what differential psychologists have uncovered about the normativedimen-
sionality of human variation (abilities, interests, and personality) to gain a purchase
on the development of the idiographic particulars of each individual. Scarrs for-
mulation fits well with treatments of how personal-attribute constellations (aptitude
complexes, trait complexes, or taxons) serve to guide behavioral development down
distinctive paths (Harris 1995).
Personal dispositions interact with the environment in three ways: (a) Passive GE
correlations are in operation, for example, when the genetic antecedents forthedevel-
opment of verbal reasoning ability covary with the vocabulary size of rearing envi-
ronments. Above average parents, for example, provide the genetic basis for complex
verbal reasoning as well as a stimulating learning environment for its development.
(b) Reactive GE correlations come about when children, because of their genetic
differences, evoke different responses from their environment (e.g. when a painfully
shy child attenuates the likelihood of spontaneous social/verbal engagement). (c)
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Active GE correlations are produced when a person takes an active role in seeking
out particular environments—for example, when children, at promise for achieving
excellence in athletics or the performing arts seek out, through their own initiative,
opportunities for athletic participation or musical instruction. This kind of GE inter-
action has been especially prominent in Scarrs recent writings, in which she has
explicated how this mechanism operates in niche building.
Scarr maintains that people (especially as they mature) seek out or strive to create
environments for themselves—environments that are congruent with their personal
point of view and which, in large part, reflect their abilities, interests, and personality.
Finding appropriate niches facilitates positive development, an idea that has been a
longstanding supposition in differential psychology (Lubinski 1996). Scarr (1996)
drew on this literature to offer recommendations for parenting. She suggests that
children need and deserve supportiveloving environmentsto ensure thatthey become
happy individuals adjusted to the complexities associated with the demands of soci-
etal roles. However, she cautions parents against trying to shape children’s enduring
characteristics reflexively; instead, parents should tailor educational curricula and
opportunities for positive development to the unique assets of each child’s
Causal Modeling
In structural equation modeling, designs that omit key determinants of phenomena
under analysis are called mis-specified (the term used to depict errors of omission)
or neglected aspects. One compelling aspect of Herrnstein & Murray (1994) worth
underscoring is their simultaneous examination of two putative causal sources (viz.
general intelligence and SES). Many social scientists found the concurrent compe-
tition of these two factors unfamiliar, as the social science literature is replete with
causal inference stemming from correlations between SES and important outcome
measures (Bouchard et al 1996), but g is seldom assessed concomitantly in such
Causality and Confounds
Removing (partialing out) SES from ability-performance correlations has been
repeatedly criticized because general intelligence and SES share common antecedents
(Bouchard 1997, Bouchard et al 1996). Meehl’s (1970) ex post facto design is the
general rubric for this methodological shortcoming. Yet, Murray (1998) has offered
a clever methodology for untangling the causal influence of SES on ability-perfor-
mance and ability-outcome functions. Using 15-year longitudinal data, Murray stud-
ied income differences between biologically related siblings (reared together) who
differed in general intelligence. As ability differences between siblings increased, so
did their income differences; moreover, these income differences mirrored those in
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
the general population at similar ability ranges. This investigation corroborates a
handful of studies using a similar control for family environment (Waller 1971).
Total Evidence Rule
The same year Burks (1928) published her landmark treatment on decomposing
environmental and hereditary sources of variation, Ellis (1928) introduced psychol-
ogists to a more general refinement. ‘The logicians point out that a cause of much
incorrect thinking is what is known as the fallacy of the neglected aspect. Early
students of certain diseases considered them to be due to hot weather or excessive
rain—neglecting the activities of the fly or the mosquito in spreading the bacteria.
Neglecting aspects of problems often hide variable agencies that must be understood
before the problem can be solved’ (Ellis 1928:9). Subsequently, Carnap (1950) for-
malized this fallacy as the total evidence rule, which maintains that, when evaluating
the plausibility of a particular hypothesis, it is imperative to take into account all of
the relevant information (Bouchard 1997, Lubinski & Humphreys 1997). As com-
monsensical as this seems, it frequently is not done.
For example, investigators readily assume that the covariation between parent and
child in abilities, interests, and personality is due to parent nurturing (cf Thompson’s
1955 review of Hart & Risley 1994). Yet, biometric analyses reveal that covariation
among broad individual differences approaches zero as adulthood is reached among
biologically-unrelated siblings reared together. As unrelated individuals who were
reared together grow older, they appear to ‘grow apart’ (McCartney et al 1990),
with respect to the attributes examined here. It appears that an inconspicuous cause,
namely shared genetic make-up, is responsible for the phenotypic covariation
between biologically related parents and children. Parents do, indeed, have an influ-
ence ontheir children with respect tothe major dimensions reviewed herein;however,
this influence is transmitted through a different reared-in mechanism than many pre-
supposed. This is also supported by a variety of kinship correlates, such as finding
that, on ‘environmental measures’’ (e.g. Home Observation for Measurement of the
Environment (HOME): Plomin & Bergeman 1991), identical twins reared apart
assess their reared-in home environments as similarly as fraternal twins reared
together do (Scarr 1996). This is not to say that abusive environments are not det-
rimental to optimal development; recall Lykken’s (1997) point about the kinds of
families that are typically not found in biometrically informed psychological studies.
What these studies do speak to, however, is that, overall, many families are func-
tionally equivalent in terms of fostering the development of broad individual differ-
ences (Harris 1995, Hetherington et al 1994).
In Consilience: The Unification of Knowledge, Wilson writes:
Today, the greatest divide within humanity is not between races, or religions,
or even, as widely believed between literate and illiterate. It is the chasm that
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
separates scientific from prescientific cultures ....Without the instruments
and accumulated knowledge of the natural sciences—physics, chemistry, and
biology—humans are trapped in a cognitive prison. They invent ingenious
speculations and myths about the origin of the confining waters, of the sun
and the sky and the stars above, and the meaning of their own existence. But
they are wrong, always wrong, because the world is too remote from ordinary
experience to be merely imagined.’ (1998:45)
By consilience, Wilson (1998) means the joining together of ideas across disci-
plines in order to paint a more comprehensive picture of the nature of the universe.
He bemoans how over-specialization among the educated elite makes important con-
ceptual syntheses unlikely, and suggests that professional misunderstandings often
arise from ignorance of other disciplines, ‘not from a fundamental difference in
mentality’’(1998:126). He cites Snow’s (1967) ‘‘two cultures’’ as a familiar example,
while also noting that social scientists frequently neglect modern biological findings.
Dawis (1992) has remarked that psychometric tools for assessing the attributes
reviewed here provide unparalleled windows on humanity—akin to the microscope
in biology and the telescope in astronomy. Over psychology’s short history much has
been learned about human diversity, especially for understanding the niches people
seek out, as well as those that people attempt to avoid, build, or change. This body
of knowledge has interconnected beautifully with other disciplines, yet frequently,
individual differences are neglected in research design and interpretation. Whereas
biologists interested in protein molecules are unlikely to say: ‘‘But I am not interested
in carbon atoms. I’ll leave that for others,’ many psychologists appear content in
examining human behavior while neglecting relevant scientific information.
Kimble (1994) has scolded psychologists in ‘Anti-Intellectualism Masquerading
as Human Sensitivity,’ for their use of huge jargon-to-substance and feeling-to-
thinking ratios on politically correct topics. ‘How you feel about a finding has no
bearing on its truth’ (Kimble 1994:257). In reviewing Sternberg & Grigorenko’s
(1997) Intelligence: Heredity and Environment, Hunt (1997) sees certain chapters as
excellent overviews, but others as nonscientific ‘cultural perspectives.’ Is there a
way to render this variance in psychological discourse more understandable?
If psychological practice is the application of scientific principles to individuals
and groups, perhaps the psychology of individual differences, combined with the
history of psychology, can illuminate such contrasting points of view. Here, I suggest
that contrasting points of view held by certain groups of psychologists reflect the
individual differences that they possess, the niches they selected for professional
development, and the scientific standards (role models) found therein. With respect
to selection for professional training, for example, psychology, relative to other dis-
ciplines, clearly draws on multiple attribute patterns (aptitudecomplexes, taxons, trait
complexes). Psychologically speaking, APA is a heterogeneous lot, relative to other
disciplines. Some psychologists work with people, others do not, and yet, this seems
reasonable. Distinctive sets of skills and interests are needed for psychology’s mul-
tifaceted roles. Boring’s (1950) familiar distinction between clinicaland experimental
psychologists (viz the former ‘like people’’) comes to mind. But are some psychol-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
ogists becoming too specialized? Is some practice and writing drifting away from a
scientific base only to result in a House of Cards (Dawes 1994)?
Although marked group differences in ‘people interests’ across psychological
specialties are conspicuous and familiar, ability profiles often vary too, sometimes in
level (in contrast to Wilson’s view), but more often in pattern (something not con-
sidered by Wilson). Indeed, it appears thatgroup differences withinthemain historical
branches of psychology, across people and their intellectual products, become more
understandable by considering individual-differences profiles. Further, it is suggested
that people attracted to certain specialties tend to approach problems with different
criteria for what constitutes a satisfying explanation. This has intensified to the point
of becoming scientifically problematic. Consider the following.
The history of systems of psychology may be traced to England (differential),
Germany (experimental), and France (clinical); the three systems differentially
emphasize quantitative, spatial, and verbal reasoning, respectively. They can be seen
as subtle divides across Snow’s (1967) two cultures. Over time, specialization in
these areas increased. Clinical psychologists slowly drifted more toward people con-
tact and intellectual content restricted to verbal reasoning and literary skill, whereas
differential psychologists and experimentalistsdrifted toward quantitativemodelsand
technical instrumentation. These all depict niche building (Scarr 1996), a powerful
tool for conceptualizing ‘‘climate’ and organizational change (Bouchard 1997, Dun-
nette 1998).
It is important to keep in mind, however, that psychological diversity is not
necessarily problematic. Individual differences can be enriching. When individual
differences are anchored by common ground, they routinely give rise to effective
solutions through different strategies. Consider Thurstone’s (1935) Vectors of the
Mind and Burt’s (1941) Factors of the Mind. Thurstone, a former engineer, chose
to highlight concepts with line drawings (spatial configurations: ‘vectors’’),
whereas Burt, a brilliant algebraist, used copious formulas (‘‘factors’’) but few
diagrams. Factor analysis can be presented either algebraically or geometrically, so
both approaches are complementary (and one may serve some students better than
others). Either is fine. But when disciplines become more complex and criticaltools
are difficult to master, psychological diversity can be problematic. Sometimes
migration is necessary. As Terman (1954:222) noted: ‘Thorndike confessed to me
once that his lack of mechanical skill was partly responsible for turning him to
mental tests andto the kindsof experiments on learningthat requirednoapparatus.’
Whereas of BF Skinner, it was said that all he needed to build an apparatus was
cardboard, string, and a piece of chewing gum. Clearly Freud and William James
were primarily literary in their approach (Freud’s literary skills earned him the
Goethe Prize, and James was arguably a better writer than his brother Henry). All
were excellent psychologists, yet most certainlythey possessed differentindividual-
differences profiles. Nevertheless, along with their uniqueness, they were all highly
scientific (investigative) in orientation. Has this latter interest changed in some
modern psychological specialties?Have certain segments of the psychologicalcom-
munity become scientifically problematic?
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
These considerations may shed light on contentious debate beyond the bio-
logical bases of human behavior. If this analysis has verisimilitude, marked group
(individual-differences) profiles should be observed among individuals with con-
trasting views on facilitated communication, recovered memory, qualitative (ver-
sus quantitative) methods, clinical (versus statistical) prediction, and numerous
alternative formulations of human intelligence (emotional, multiple, etc.).Perhaps
one group places a premium on characterizing unique nuances attendant with all
psychological phenomena (measured against criteria of eloquence and verbal
cohesiveness), while another emphasizes the communality cutting across scien-
tifically significant dimensions of human behavior and their external linkages
(graphed or measured quantitatively). If so, the seeds for communicating at cross-
purposes are planted and germinate from deep differences in fundamental quali-
ties—nurtured and supported by distinct niches.
Differential psychology not only fosters consilience, it offers understanding
for some of the most critical social issues of our time. A coherent picture of the
human condition is incomprehensible without individual differencesconcepts and
methods. Finally, and perhaps most profoundly, differential psychology might
point to ways to enhance the scientific integrity of psychology, and the social
sciences more generally, by revealing (through multidimensional models) ways
to develop, select, and train students for coming to terms with human behavior
from a scientific point of view.
I am grateful to several colleagues and friends for providing me with invaluable
discussions and feedback on this chapter: Britt Anderson, Camilla P Benbow,
Thomas J Bouchard Jr, John B Carroll, Rene V Dawis, Lewis Goldberg, Robert
A Gordon, Linda S Gottfredson, Robert Hogan, Lloyd G Humphreys, Douglas
N Jackson, Arthur R Jensen, David T Lykken, Paul E Meehl, Martha Morelock,
Robert Plomin, James Rounds, Frank L Schmidt, Lynne Schoenauer, Daniel L
Shea, Julian C Stanley, Auke Tellegen, Mary L Tenopyr, Niels G Waller, and
Rose Mary Webb.
Visit the Annual Reviews home page at
Achter JA, Lubinski D, Benbow CP. 1996.
Multipotentiality among intellectually
gifted: ‘It was never there and already it’s
vanishing.’ J. Couns. Psychol. 43:65–76
Achter JA, Lubinski D, Benbow CP, Eftekhari-
Sanjani H. 1999. Assessingvocational pref-
erences among gifted adolescents adds
incremental validity to abilities. J. Educ.
Psychol. 91:777–86
Ackerman PL. 1988. A review of Linda S Gott-
fredson ed. The g factor in employment.
Journal of Vocational Behavior. Educ. Psy-
chol. Meas. 48:553–58. Special issue
Ackerman PL. 1996. A theory of adult intel-
lectual development: process, personality,
interests, and knowledge. Intelligence 22:
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Ackerman PL, Heggestad ED. 1997. Intelli-
gence, personality, and interests: evidence
for overlapping traits. Psychol. Bull.
Allport GW, Odbert HS. 1936. Trait names: a
psycho-lexical study. Psychol. Monogr:
Gen. Appl. 47 (whole No. 211: 1–171)
Almagor M, Tellegen A, Waller NG. 1995. The
big seven model: a cross-cultural replica-
tion. J. Pers. Soc. Psychol. 69:300–7
Amer. Psychol. 1965 Testing and public policy.
Am. Psychol. 1998. Comments. 53:566–77
Anastasi A. 1937. DifferentialPsychology. NY:
Anderson B. 1993. Evidence from the rat for a
general factor that underlies cognitive per-
formance that relates to brain size: intelli-
gence? Neurosci. Lett. 153:98–102
Anderson B. 1994a. Role of animal research in
the investigation of human mental retarda-
tion. Am. J. Ment. Retard. 99:50–59
Anderson B. 1994b. The volume of the cere-
bellar molecular layer predicts attention to
novelty in rats. Brain Res. 641:160–62
Anderson B. 1995. G explained. Med. Hypoth-
eses. 45:602–4
Anderson B, Harvey T. 1996. Alterations in
cortical thickness and neuronal density in
the frontal cortex of Albert Einstein. Neu-
rosci. Lett. 210:161–64
Austin JT, Hanisch KA. 1990. Occupational
attainment as a function of abilities and
interests. J. Appl. Psychol. 75:77–86
Barrett GV, Depinet RL. 1991. A reconsidera-
tion for competence rather than for intelli-
gence. Am. Psychol. 46:1012–24
Benbow CP. 1988. Sex differences in mathe-
matical reasoning ability among the intel-
lectually talented. Behav. Brain Sci.
Benbow CP. 1992. Academic achievement in
mathematics and science of students
between ages 13 and 23: Are there differ-
ences among students in the top one percent
of mathematical ability? J. Educ. Psychol.
Benbow CP, Lubinski D, eds. 1996. Intellec-
tual Talent. Baltimore: Johns Hopkins
Univ. Press. 428 pp.
Benbow CP, Stanley JC. 1996. Inequity in
equity: how ‘equity’ can lead to inequity
for high-potential students. Psychol. Public
Policy. Law 2:249–92
Benet V, Waller NG. 1995. The big seven fac-
tor model of personality description: evi-
dence for its cross-cultural generality in a
Spanish sample. J. Pers. Soc. Psychol.
Berliner DC, Calfee RC, eds. 1996. Handbook
of Educational Psychology. New York:
Macmillan. 1071 pp.
Block J. 1995a. A contrarian view of the five
factor approach to personality description.
Psychol. Bull. 117:187–215
Block J. 1995b. Going beyond the five factors
given: rejoinder to Costa and McCrae
(1995) and Goldberg and Saucier (1995).
Psychol. Bull. 117:226–29
Bouchard TJ. 1997. Genetic influence on men-
tal abilities, personality, vocational inter-
ests, and work attitudes. Int. Rev. Indus.
Organ. Psychol. 12:373–95
Bouchard TJ. 1999. The definitive case for g.
Contemp. Psychol. 44:133–35
Bouchard TJ, Lykken DT, Tellegen A, McGue
M. 1996. Genes, drives, environment, and
experience: EPD theory revisited. See Ben-
bow & Lubinski 1996, pp. 5–43
Brand C. 1987. The importance of general
intelligence. In Arthur Jensen: Consensus
and Controversy, ed. S Magil, C Magil, pp.
251–65. NY: Falmer. 420 pp.
Brody N. 1992. Intelligence. San Diego: Aca-
demic. 395 pp.
Brody N. 1996. Intelligence and public policy.
Psychol. Public Policy Law 3/4:473–85
Boring EG. 1950. A History of Experimental
Psychology. NY: Appleton-Century-Crofts.
777 pp.
Burks BS. 1928. The relative influence of
nature and nurture upon mental develop-
ment. 27th Yearbook of the National Society
for Education, pp. 219–316. Bloomington,
IN: Public School
Burt C. 1941. The Factors of the Mind. NY:
Campbell DT, Fiske DW. 1959. Convergent
and discriminant validation by the
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
multitrait-multimethod matrix. Psychol.
Bull. 93:81–105
Campbell JP. 1990. The role of theory in indus-
trial and organizational psychology. In
Handbook of I/O Psychology, ed. MD Dun-
nette, LM Hough, pp. 39–74. Palo Alto,
CA: Consulting Psychologists. 755 pp.
Campbell JP. 1996. Group differences and per-
sonnel decisions: validity, fairness, and
affirmative action. J. Vocat. Behav. 49:122–
Carpenter PA, Just MA, Shell P. 1990. What
one intelligence test measures: a theoretical
account. Psychol. Rev. 97:404–31
Carnap R. 1950. Logical Foundations of Prob-
ability. Chicago: Univ. Chicago Press
Carroll JB. 1993. Human Cognitive Abilities.
Cambridge: Cambridge Univ. Press.
819 pp.
Carroll JB. 1994. An alternative Thurstonian
view of intelligence. Psychol. Inq. 5:
Carroll JB. 1997. Psychometrics, intelligence,
and public perception. Intelligence 24:
Caspi A, Begg D, Dickson N, Harrington H,
Langley J, et al. 1997. Personality differ-
ences predict health-risk behaviors in
young adulthood. J. Pers. Soc. Psychol.
Caspi A, Moffitt TE. 1993. When do individual
differences matter? A paradoxical theory of
personality coherence. Psychol. Inq. 4:247–
Cattell JM. 1890. Mental tests and measure-
ments. Mind 15:373–80
Cattell RB. 1971. Abilities: Their Structure
and Growth. Boston: Houghton Mifflin
Ceci SJ. 1996. IQ in society. Psychol. Public
Policy Law 3/4:403–645. Special issue
Chapman PD. 1988. School as Sorters. NY:
New York Univ. Press
Chorney MJ, Chorney K, Seese N, Owen MJ,
McGuffin P, et al. 1998. A quantitative trait
locus (QTL) associated with cognitive abil-
ity in children. Psychol. Sci. 9:159–66
Cleary TA, Humphreys LG, Kendrick SA,
Wesman A. 1975. Educational uses of tests
with disadvantaged students. Am. Psychol.
Coleman JS. 1990–1991. The Sidney Hook
Memorial Award Address: on the self-
suppression of academic freedom. Acad. Q.
Cook TD, Campbell DT. 1979. Quasi-
Experimentation. Chicago: Rand McNally.
405 pp.
Costa PT, McCrae RR. 1995. Solid ground in
the wetlands of personality: a reply to
Block. Psychol. Bull. 117:216–20
Cronbach LJ. 1970. Essentials of Psychologi-
cal Testing. NY: Harper & Row. 752 pp. 3rd
Cronbach LJ. 1975. Five decades of public
controversy over mental testing. Am. Psy-
chol. 30:1–14
Cronbach LJ, Meehl PE. 1955. Construct
validity in psychological tests. Psychol.
Bull. 52:281–302
Dawes RM. 1994. House of Cards. NY: Free.
338 pp.
Dawis RV. 1991. Vocational interests, values,
and preferences. In Handbook of I/O Psy-
chology, Vol. 2, ed. MD Dunnette, LM
Hough, pp. 833–71. Palo Alto, CA: Con-
sulting Psychologists. 957 pp.
Dawis RV. 1992. The individual differences
tradition in counseling psychology. J.
Couns. Psychol. 39:7–19
Dawis RV, Lofquist LH. 1984. A Psychological
Theory of Work Adjustment. Minneapolis:
Univ. Minn. Press. 245 pp.
Day SX, Rounds J. 1998. The universality of
vocational interest structure among racial/
ethnic minorities. Am. Psychol. 53:728–36
Deary IJ. 1996. Intelligence and inspection
time. Am. Psychol. 51:599–608
Diamond MC, Scheibel AB, Murphy GM, Har-
vey T. 1985. The brain of a scientist: Albert
Einstein. Exp. Neurol. 88:198–204
Dunnette MD. 1998. Emerging trends and vex-
ing issues in I/O psychology. Appl. Psy-
chol: Int. Rev. 47:129–53
Ellis RS. 1928. The Psychology of Individual
Differences. NY: Appleton. 533 pp.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Embretson S. 1983. Construct validity: con-
struct representation versus nomothetic
span. Psychol. Bull. 93:179–97
Ericsson KA. 1996. The Road to Excellence.
Mahwah, NJ: Erlbaum. 369 pp.
Eysenck HJ. 1995. Genius. Cambridge: Cam-
bridge Univ. Press. 344 pp.
Feingold A. 1995. The additive effects of dif-
ferences in central tendency and variability
are important in comparisons between
groups. Am. Psychol. 50:5–13
Fiske DW. 1949. Consistency of the factorial
structures of personality ratings from dif-
ferent sources. J. Abnorm. Soc. Psychol.
Flynn JR. 1999. Searchingfor justice. Am.Psy-
chol. 54:5–20
Galton F. 1869. Hereditary Genius. London:
Macmillian. 379 pp.
Galton F. 1884. Measurement of character.
Fortn. Rev. 36:179–85
Galton F. 1890. Remarks. Mind 15:380–81
Gardner H. 1993. Creating Minds. NY: Basic.
464 pp.
Gati L. 1991. The structure of vocational inter-
ests. Psychol. Bull. 109:209–24
Geary DC. 1998. Male, Female: The Evolution
of Human Sex Differences. Washington,
DC: Am. Psychol. Assoc. 397 pp.
Ghiselli EE. 1972. Comment on the use of
moderator variables. J. Appl. Psychol.
Goldberg LR, Saucier G. 1995.So what do you
propose we use instead? A reply to Block.
Psychol. Bull. 117:221–25
Gordon RA. 1997. Everyday life as intelli-
gence test. Intelligence 24:203–320
Gottfredson LS. 1997. Intelligence and social
policy. Intelligence 24:L-320. Special issue
Gottfredson LS, Sharf JC. 1988. Fairness in
employment testing. J. Vocat. Behav.
33:225–477. Special issue
Grove WM, Meehl PE. 1996. Comparative
efficiency of informal (subjective, impres-
sionistic) and formal (mechanical, algorith-
mic) prediction procedures: the
clinical-statistical controversy. Psychol.
Public Policy Law 2:293–323
Guilford JP, et al. 1954. A factor analytic study
of human interests. Psychol. Monogr. 68(4,
whole No. 375)
Gustafsson J, Undheim JO. 1996. Individual
differences in cognitive functions. See Ber-
liner & Calfee 1996, pp. 186–242
Haier RJ. 1993. Cerebral glucose metabolism
and intelligence. See Vernon 1993, pp.
Haier RJ, Benbow CP. 1995. Sex differences
in lateralization in temporal lobe glucose
metabolism during mathematical reason-
ing. Dev. Neurobiol. 11:405–14
Halpern DF. 1992. SexDifferencesin Cognitive
Abilities. Hillsdale, NJ: Erlbaum. 308 pp.
Halpern DF. 1996. Public policy implications
of sex differences in cognitive abilities.
Psychol. Public Policy Law 3/4:561–74
Harmon LW, Hansen JC, Borgen FH, Hammer
AL. 1994. Applications and Technical
Guide for the Strong Interest Inventory.
Palo Alto, CA: Consult. Psychol. 392 pp.
Harris J. 1995. Where is the child’s environ-
ment?: a group socialization theory of
development. Psychol. Rev. 102:458–89
Hart B, Risley TR. 1995. Meaningful Differ-
ences. Baltimore: Brookes. 268 pp.
Hartigan JA, Wigdor AK. 1989. Fairness in
Employment Testing. Washington, DC:
Natl. Acad. 354 pp.
Hayes KJ. 1962. Genes, drives, and intellect.
Psychol. Rep. 10:299–342
Hedges LV, Nowell A. 1995. Sex differences
in mental test scores, variability, and num-
bers of high scoring individuals. Science
Herrnstein RJ, Murray C. 1994. The Bell
Curve. NY: Free. 845 pp.
Hetherington EM, Reiss D, Plomin R. 1994.
Separate Worlds of Siblings. Hillsdale, NJ:
Earlbaum. 232 pp.
Hogan J, Hogan R. 2000. Theoretical frame-
works for assessment. In Individual Assess-
ment, ed. PR Jeanneret, R Silzer. San
Francisco: Jossey Bass. In press
Hogan R, Hogan J, Roberts BW. 1996. Person-
ality measurement and employment deci-
sions. Am. Psychol. 51:469–77
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Holahan CK, Sears RR. 1995. The Gifted
Group in Later Maturity. Stanford, CA:
Stanford Univ. Press. 363 pp.
Holland JL. 1996. Exploring careers with a
typology. Am. Psychol. 51:397–406
Hough LM. 1997. The millennium for person-
ality psychology: new horizons or good old
daze. Appl. Psychol.: Int. Rev. 47:233–61
Humphreys LG. 1962. The organization of
human abilities. Am. Psychol. 17:475–83
Humphreys LG. 1988. Trends in levels of aca-
demic achievement of blacks and other
minorities. Intelligence 12:231–60
Humphreys LG. 1990. View of a supportive
empiricist. Psychol. Inq. 1:153–55
Humphreys LG. 1991. Limited vision in the
social sciences. Am. J. Psychol. 104:333–
Humphreys LG. 1994. Intelligence from the
standpoint of a (pragmatic) behaviorist Psy-
chol. Inq. 5:179–92
Humphreys LG. 1995. Foreword. In Assessing
Individual Differences in Human Behavior,
ed. D Lubinski, RV Dawis, p. v. Palo Alto,
CA: Consulting Psychologists. 386 pp.
Humphreys LG, Lubinski D, Yao G. 1993.
Utility of predicting group membership and
the role of spatial visualization in becoming
an engineer, physical scientist, or artist. J.
Appl. Psychol. 78:250–61
Hunt E. 1995. Will We Be Smart Enough? NY:
Russell Sage Found. 332 pp.
Hunt E. 1996. When should we shoot the mes-
senger?: issues involving cognitive testing,
public policy, and law. Psychol. Public Pol-
icy Law 3/4:486–505
Hunt E. 1997. Nature versus nurture: The feel-
ing of vuja de. In Intelligence:Heredityand
Environment, JR Sternberg and E Grigo-
renko ed. pp. 531–51. NY: Cambridge
Univ. Press. 608 pp.
Hunter JE, Schmidt FL. 1996. Intelligence and
job performance: economic and social
implications. Psychol. Public Policy Law 3/
Jamison KR. 1993. Touched With Fire. NY:
Free. 370 pp.
Jenkins JJ, Paterson DG. 1961. Studies in Indi-
vidual Differences: The Search for Intelli-
gence. NY: Appleton-Century-Crofts. 774
Jensen AR. 1980. Bias in Mental Testing. NY:
Free. 786 pp.
Jensen AR. 1996. Giftedness and genius. See
Benbow & Lubinski 1996, pp. 393–411
Jensen AR. 1998. The g Factor. Westport, CT:
Praeger. 648 pp.
J. Vocat. Behav., ed. HEA Tinsley. 1993. Spe-
cial issue on the Theory of Work Adjust-
ment. 43(1 ):1–132 (whole issue)
Katzell RA. 1994. Contemporary meta-trends
in industrial and organizationalpsychology.
In Handbook of I/O Psychology, vol. 4, ed.
HC Triandis, MD Dunnette, LM Hough,pp.
1–89. Palo Alto, CA: Consulting Psychol-
ogists. 869 pp.
Kimble GA. 1994. A new formula for behav-
iorism. Psychol. Rev. 101:254–58
Kyllonen PC, Christal RE. 1990. Reasoning
ability is (little more than)working memory
capacity?! Intelligence 14:389–433
Lerner B. 1983. Test scores as measures of
human capital and forecasting tools. In
Intelligence and National Achievement, ed.
RB Cattell, pp. 70–99. Washington DC:
Lippa R. 1998. Gender-related individual dif-
ferences and the structure of vocational
interests. J. Pers. Soc. Psychol. 74:996–
Loehlin JC. 1997. Dysgenesis andIQ. Am. Psy-
chol. 52:1236–39
Loevinger J. 1954. Effect of distortion on item
selection. Educ. Psychol. Meas. 14:441–48
Lofquist LH, Dawis RV. 1991. Essentials of
Person Environment Correspondence
Counseling. Minneapolis: Univ. Minn.
Press. 171 pp.
Lubinski D. 1996. Applied individual differ-
ences research and its quantitativemethods.
Psychol. Public. Policy. Law 2:187–203
Lubinski D, Benbow CP. 2000. States of excel-
lence. Am. Psychol. 54: In press
Lubinski D, Dawis RV. 1992. Aptitudes, skills,
and proficiencies. In Handbook of I/O Psy-
chology, Vol. 3, ed. MD Dunnette, LM
Hough, pp. 1–59. Palo Alto, CA: Consult-
ing Psychologists. 1095 pp.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Lubinski D, Humphreys LG. 1997. Incorpo-
rating general intelligence into epidemi-
ology and the social sciences. Intelligence
Lubinski D, Thompson T. 1993. Species and
individual differences in communication
based on private states. Behav. Brain Sci.
Lutzenberg W, Burbaumer N, Flor H, Rock-
stroh B, Elbert T. 1992. Dimensional anal-
ysis of the human EEG and intelligence.
Neurosci. Lett. 143:10–14
Lykken DT. 1957. A study of anxiety in the
sociopathic personality. J. Abnorm. Soc.
Psychol. 55:6–10
Lykken DT. 1991. What’s wrong with psy-
chology anyway? In Thinking Clearly
About Psychology, ed. D Chiccetti, W
Grove, pp. 3–39. Minneapolis: Univ. Minn.
Press. 277 pp.
Lykken DT. 1995. The Antisocial Personali-
ties. Hillsdale, NJ: Erlbaum. 259 pp.
Lykken DT. 1997.The Americancrimefactory.
Psychol. Inq. 8:261–70
Lykken DT, McGue M, Tellegen A, Bouchard
TJ. 1992. Emergenesis. Am. Psychol.
Lynn R. 1996. Dysgenics. Westport, CT: Prae-
ger. 237 pp.
MacCorquodale K, Meehl PE. 1948. On a dis-
tinction between hypothetical constructs
and intervening variables. Psychol. Rev.
Macklin ML, Metzger LJ, Litz BT, McNally
RJ, Lasko NB, et al. 1998. Lower precom-
bat intelligence is a risk factor for posttrau-
matic stress disorder. J. Consult. Clin.
Psychol. 66:323–26
Maxwell G. 1961. Meaning postulates in sci-
entific theory. In Current Issues in Philos-
ophy of Science, ed. H Feigl, G Maxwell,
pp. 169–95. NY: Holt, Rinehart, & Win-
ston. 484 pp.
McCartney K, Harris MJ, Bernieri F. 1990.
Growing up and growing apart: a develop-
mental meta-analysis of twin studies. Psy-
chol. Bull. 107:226–37
McCrae RR, Costa PT. 1997. Personalitystruc-
ture as a human universal. Am. Psychol.
McGue M, Bouchard TJ. 1998. Genetic and
environmental influences on human behav-
ioral differences. Annu. Rev. Neurosci.
McNemar Q. 1964. Lost: our intelligence?
Why? Am. Psychol. 19:871–82
Meehl PE. 1970. Nuisance variables and the ex
post facto design. In Minnesota Studies in
the Philosophy of Science IV ed. M Radner,
S Winokur. Minneapolis: Univ. Minn.
Press. 441 pp.
Meehl PE. 1990. Appraising and amending
theories. Psychol. Inq. 1:108–41
Meehl PE. 1998. The Power of Quantitative
Thinking. Washington, DC: American Psy-
chological Society Cattell Award Address.
Messick S. 1992. Multiple intelligences or
multilevel intelligence? Psychol. Inq.
Miller EM. 1994. Intelligence and brain mye-
lination: a hypothesis. Pers. Individ. Differ.
Moffitt TE, Caspi A, Harkness AR, Silva PA.
1993. The natural history of change in intel-
lectual performance. J. Child. Psychol. Psy-
chiatry 34:455–506
Murray C. 1998. Income, Inequality, and IQ.
Washington, DC: Am. Enterprise Inst. 47
Neisser U. 1998. The Rising Curve. Washing-
ton, DC: Am. Psychol. Assoc. 415 pp.
Neisser U. 1999. The great g mystery. Con-
temp. Psychol. 44:131–33
Neisser U, Boodoo G, Bouchard TJ, Boykin
AW, Brody N, et al. 1996. Intelligence:
knowns and unknowns. Am. Psychol.
Norman WT. 1963. Toward a taxonomy of per-
sonality attributes. J.Abnorm. Soc.Psychol.
O’Boyle MW, Benbow CP, Alexander JE.
1995. Sex differences, hemispheric lateral-
ity, and brain activity in the intellectually
gifted. Dev. Neurobiol. 11:415–44
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
O’Toole BJ. 1990. Intelligence and behavior
and motor vehicle accident mortality.
Accid. Anal. Prev. 22:211–21
Plomin R, Bergeman CS. 1991. The nature of
nuture: genetic influence on ‘environmen-
tal’ measures. Behav. Brain Sci. 14:373–
Prediger DJ. 1982. Dimensions underlying
Holland’s hexagon: missing link between
interests and occupations? J. Vocat. Behav.
Premack D. 1983. The codes of man and
beasts. Behav. Brain Sci. 6:125–67
Roe A. 1953. The Making of a Scientist. NY:
Dodd, Mead. 244 pp.
Roe A. 1956. The Psychology of Occupations.
NY: Wiley. 340 pp.
Rolfhus EL, Ackerman PL. 1996. Self-report
knowledge: at the crossroads of ability,
interests, & personality. J. Educ. Psychol.
Rounds JB, Tracey TJ. 1993. Predigersdimen-
sional representation of Holland’s RIASEC
circumplex. J. Appl. Psychol. 78:875–90
Rowe DC. 1994. The Limits of Family Influ-
ence. NY: Guilford. 232 pp.
Rowe DC. 1997. Are parents to blame? A look
at the antisocial personalities. Psychol. Inq.
Rowe DC, Vazsonyi AT, Flannery DJ. 1994.
No more than skin deep: ethnic and racial
similarity in developmental process. Psy-
chol. Rev. 101:396–413
Rowe DC, Vazsonyi AT, Flannery DJ. 1995.
Ethnic and racial similarity in developmen-
tal process. Psychol. Sci. 6:33–38
Roznowski M. 1987. Use of tests manifesting
sex differences as measures of intelligence.
J. Appl. Psychol. 72:480–83
Sanders CE, Lubinski D, Benbow CP. 1995.
Does the Defining Issues Test measure psy-
chological phenomena distinct from verbal
ability?: an examination of Lykken’s query.
J. Pers. Soc. Psychol. 69:498–504
Savage-Rumbaugh S, Lewin R. 1994. Kanzi:
The Ape at the Brink of the Human Mind.
NY: Wiley. 299 pp.
Savage-Rumbaugh S, Murphy J, Sevcik RA,
Brakke KE, et al. 1993. Language compre-
hension in ape and child. Monogr. Soc. Res.
Child. Dev. 58:1–256
Savickas ML, Spokane AR. 2000. Vocational
Interests. Palo Alto, CA: Davies-Black. In
Scarr S. 1992. Developmental theories for the
1990s: development and individual differ-
ences. Child Dev. 63:1–19
Scarr S. 1996. How people make their own
environments: implications for parents and
policy makers. Psychol. Public Policy Law
Scarr S, McCartney K. 1983. How people
make their own environments: a theory of
genotype r environment effects. ChildDev.
Schaie KW. 1996. Intellectual Development in
Adulthood. NY: Cambridge Univ. Press.
396 pp.
Schmidt DB, Lubinski D, Benbow CP. 1998.
Validity of assessing educational-voca-
tional preference dimensions among intel-
lectually talented 13-year olds. J. Couns.
Psychol. 45:436–53
Schmidt FL, Hunter JE. 1998. The validity and
utility of selection methods in personnel
psychology: practical and theoretical impli-
cations of 85 years of research findings.
Psychol. Bull. 124:262–74
Schneider BS, Goldstein AW, Smith DB. 1995.
The ASA framework: an update. Personnel
Psychol. 48:747–73
Scott WD. 1920. Changes in some of our con-
ceptions and practices of personnel. Psy-
chol. Rev. 27:81–94
Sidman M. 1986. Functional analysis of emer-
gent verbal classes. In Analysis and Inte-
gration of Behavioral Units, ed. T
Thompson, MD Zeiler, pp. 213–45. Hills-
dale, NJ: Erlbaum. 367 pp.
Simonton DK. 1990. Psychology, Science, and
History. New Haven, CT: Yale Univ. Press.
291 pp.
Simonton DK. 1994. Greatness. NY: Guilford.
502 pp.
Snow CP. 1967. The Two Cultures and a Sec-
ond Look. London: Cambridge Univ. Press.
107 pp.
Snow RE. 1989. Aptitude-treatment interac-
tion as a framework for research on indi-
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
vidual differences in learning. In Learning
and Individual Differences, ed. PL Acker-
man, RJ Sternberg, RG Glasser, pp. 13–59.
NY: Freedman. 107 pp.
Snow RE. 1991. The concept of aptitude. In
Improving Inquiry in the Social Sciences,
ed. RE Snow, DE Wiley, pp. 249–84. Hills-
dale, NJ: Earlbaum. 343 pp.
Snow RE. 1996. Aptitude development and
education. Psychol. Public Policy Law 3/
Snow RE, Corno L, Jackson D III. 1996. Indi-
vidual differences in affective and conative
functions. See Berliner & Calfee 1996, pp.
Snow RE, Lohman DF. 1989. Implications of
cognitive psychology for educational mea-
surement. In Educational Measurement,ed.
R Linn, pp. 263–331. NY: Collier. 610 pp.
Snyderman M, Rothman S. 1987. Survey of
expert opinion on intelligence and aptitude
testing. Am. Psychol. 42:137–44
Spearman C. 1904. ‘General intelligence,’
objectively determined and measured. Am.
J. Psychol. 15:201–92
Spearman C. 1927. The Abilities of Man. NY:
Macmillan. 415 pp.
Stanley JC, Benbow CP, Brody LE, Dauber S,
Lupkowski A. 1992. Gender differences on
eighty-six nationally standardized tests. In
Talent Development, ed. N Colangelo, SG
Assouline, DL Ambroson, pp. 42–65. NY:
Trillium. 432 pp.
Sternberg RJ. 1997. Intelligence and life long
learning. Am. Psychol. 51:77–101
Sternberg RJ, Detterman DK. 1986. What is
Intelligence? Norwood, NJ: Ablex. 173 pp.
Sternberg RJ, Grigorenko EL. 1997. Intelli-
gence: Heredity and Environment. NY:
Cambridge Univ. Press. 608 pp.
Tanner JM. 1965. Physique and athletic per-
formance: a study of Olympic athletes. In
Penguin Science Survey 1965 B, ed. SA
Barnett, A McLaren. Baltimore: Penguin
Tellegen A. 1993. Folk concepts and psycho-
logical concepts in personality and person-
ality disorder. Psychol. Inq. 4:122–30
Tellegen A, Grove W, Waller NG. 1991. Inven-
tory of Personal Characteristics #7 (IPC7).
Unpublished materials, Dept. Psychology,
Univ. Minnesota, Minneapolis
Tellegen A, Waller NG. 2000. Exploring per-
sonality through test construction: devel-
opment of the Multidimensional
Personality Questionnaire. In Personality
Measures: Development and Evaluation,
ed. SR Briggs, JM Cheek. Greenwich, CT:
JAI. In press
Terman L. 1954. The discovery and encour-
agement of exceptional talent. Am. Psychol.
Thompson T. 1995. Children have more need
of models than critics: early language expe-
rience and brain development. J. Early
Interv. 19:264–72
Thorndike R, Lohman DF. 1990. A Century of
Ability Testing. Chicago: Riverside. 163 pp.
Thorndike RL. 1985. The central role of gen-
eral ability in prediction. Multivar. Behav.
Res. 20:241–54
Thorndike RL. 1994. g (Editorial). Intelligence
Thurstone LL. 1935. Vectors of the Mind. Chi-
cago: Univ. Chicago Press. 266 pp.
Thurstone LL. 1940. Issues in factor analysis.
Psychol. Bull. 37:189–235
Tracey TJ, Rounds J. 1993. Evaluating Hol-
land’s and Gati’s vocational-interest mod-
els: a structural meta-analysis. Psychol.
Bull. 113:229–46
Tupes EC, Christal RE. 1992. Recurrent per-
sonality factors based on trait ratings. J.
Pers. 60:225–51. (Reprinted from USAF
ASD Tech. Rep. No. 61–97, 1961, Lack-
land Air Force Base, TX: US Air Force.)
Tyler LE. 1965. The Psychology of Human Dif-
ferences. NY: Appleton-Century-Crofts.
572 pp.
Vernon PA. 1993.Biological Approachesto the
Study of Human Intelligence. Norwood, NJ:
Vernon PE. 1961. The Structure of Human
Abilities. London: Methuen. 208 pp. 2nd
Waller JH. 1971. Achievement and social
mobility: relationships among IQ score,
education, and occupation in two genera-
tions. Soc. Biol. 18:252–59
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Waller NG. 1999. Evaluating the structure of
personality. In Personality and Psychopa-
thology, ed. CR Cloninger. pp. 155–97.
Washington, DC: Am. Psychiatric Assoc.
524 pp.
Waller NG, Kojetin BA, Bouchard TJ, Lykken
DT, Tellegen A. 1990. Geneticand environ-
mental influences on religious interests,
attitudes, and values. Psychol. Sci. 2:138–
Washburn DA, Rumbaugh D. 1997. Faster is
smarter so why are we slower? Am. Psy-
chol. 52:1147–48
Wasserman EA. 1993. Comparative cognition.
Psychol. Bull. 113:211–28
Wiegman O, Kuttschreuter M, Baarda B. 1992.
A longitudinal study of the effects of tele-
vision viewing on aggressive and prosocial
behavior. Br. J. Soc. Psychol. 31:147–64
Wigdor AK, Garner WR. 1982. Ability Testing.
Washington, DC: National Academy
Wilk SL, Desmarais LB, Sackett PR. 1995.
Gravitation to jobs commensurate with
ability: longitudinal and cross-sectional
tests. J. Appl. Psychol. 80:79–85
Wilk SL, Sackett PR. 1996. Longitudinal anal-
ysis of ability-job complexity fit and job
change. Personnel Psychol. 49:937–67
Willerman L. 1979. The Psychology of Individ-
ual and Group Differences. San Francisco:
Freedman. 531 pp.
Willerman L, Rutledge JN, Bigler ED. 1991.
In vivo brain size and intelligence. Intelli-
gence 15:223–28
Williams WM, Ceci SJ. 1997. Are Americans
becoming more or less alike?: trends in
race, class, and ability differences in intel-
ligence. Am. Psychol. 52:1226–35
Willis SL, Schaie KW. 1986. Practical intelli-
gence in later adulthood. In Practical Intel-
ligence, ed. RJ Sternberg, RK Wagner, pp.
236–68. Cambridge: Cambridge Univ.
Press. 386 pp.
Wilson EO. 1998. Consilience: The Unity of
Knowledge. NY: Knopf. 332 pp.
Wilson JQ, Herrnstein RJ. 1985. Crime and
Human Nature. NY: Simon & Schuster. 639
Zuckerman H. 1977. Scientific Elite. NY: Free.
235 pp.
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
Annual Review of Psychology
Volume 51, 2000
Parenting and its Effects on Children: On Reading and Misreading
Behavior Genetics,
leanor E. Maccob
The Pain of Being Sick: Implications of Immune-to-Brain Communication
for Understandin
Pain, L. R. Watkins, S. F. Maier
Thought Suppression,
ichard M. Wenzlaff, Daniel M. Wegner
Social Cognition: Thinking Categorically about Others, C. Neil Macrae,
Galen V. Bodenhausen
Are There Kinds of Concepts?, Douglas L. Medin, Elizabeth B. Lynch,
aren O. Solomon
New Perspectives and Evidence on Political Communication and
n Effects, Shanto I
ar, Adam F. Simon
Goal Theory, Motivation, and School Achievement: An Integrative
artin V. Covin
Applications of Structural Equation Modeling in Psychological Research,
obert C. MacCallum, James T. Austin
The Environmental Psychology of Capsule Habitats, Peter Suedfeld, G.
aniel Steel
Food Intake and the Regulation of Body Weight, Stephen C. Woods,
ichael W. Schwartz, Denis G. Baskin, Rand
J. Seele
Negotiation, Max H. Bazerman, Jared R. Curhan, Don A. Moore,
athleen L. Valle
Parental and Child Cognitions in the Context of the Family,
aphne Blunt
ental, Charlotte Johnston
Evaluation Methods for Social Intervention, Mark W. Lipsey, David S.
Adult Psychopathology: Issues and Controversies, T. A. Widiger, L. M.
Scientific and Social Significance of Assessing Individual Differences:
""Sinking Shafts at a Few Critical Points, David Lubinski
The Effects of Family and Community Violence on Children, Gayla
olin, Elana B. Gordis
Toward a Psychology of Memory Accuracy, Asher Koriat, Morris
Goldsmith, Ainat Pansk
Attitude Change: Persuasion and Social Influence, Wendy Wood
Cultural Psychopathology: Uncovering the Social World of Mental
Illness, Steven Re
eser Ló
ez, Peter J. Guarnaccia
Memory Systems in the Brain,
dmund T. Rolls
Personnel Selection: Looking Toward the Future--Remembering the Past,
eaetta M. Hou
h, Frederick L. Oswald
Emotion, Regulation, and Moral Development, Nancy Eisenberg
Neural Basis of Hearing in Real-World Situations, Albert S. Feng, Rama
Annu. Rev. Psychol. 2000.51:405-444. Downloaded from
by UNIVERSITY OF ILLINOIS on 10/07/07. For personal use only.
... One of the principles of the identification paradigm is that any variable that predicts future attainment should be used to identify students in need of gifted services in the present. Generally speaking, the best predictor of future achievement is current achievement, followed by ability, interest, and personality factors (Ackerman et al., 2011;Lohman, 2005a;Lubinski, 2000). Achievement is most directly affected by opportunity to learn because it is most closely tied to education (Lohman, 2005b). ...
... Thus, it holds that ability tasks can be meaningfully organized by the content of the tasks and that verbal, spatial, and quantitative reasoning comprise the broad abilities. A wide range of work supports the central role of these three domains to academic success (see Lubinski, 2000, for a review). ...
Education researchers, policymakers, and practitioners are concerned with identifying and developing talent for students with fewer opportunities, especially students from historically marginalized groups. An emerging body of research suggests “universally screening” or testing all students, then matching those students with appropriate educational challenges, is effective in helping marginalized students. However, most tests have focused on two areas: math and verbal reasoning. We leverage three nationally representative samples of the U.S. population at different time points that include both novel cognitive measures (e.g., spatial, mechanical, and abstract reasoning) and non-cognitive measures (e.g., conscientiousness, creativity or word fluency, leadership skill, and artistic skill) to uncover which measures would improve proportional representation of marginalized groups in talent identification procedures. We find that adding spatial reasoning measures in particular—as well as other non-cognitive measures such as conscientiousness, leadership, and creativity—are worthwhile to consider for universal screening procedures for students to narrow achievement gaps at every level of education, including for gifted students. By showing that these nontraditional measures both improve proportional representation of underrepresented groups and have reasonable predictive validity, we also broaden the definition of what it means to be “gifted” and expand opportunities for students from historically marginalized groups.
... These traits influence a major aspect of one's long-term economic prospects in life, sit at the juncture of research between differential psychology and labor economics, and often have important policy implications. One associated and often contentious question concerns the sex difference in occupational interests, which is possibly "the largest of all sex differences on major psychological dimensions" [1]. These sex differences are well established; they have been studied for more than a century and are relatively consistent across nations and across historical periods [2][3][4][5]. ...
Full-text available
We investigated sex differences in 473,260 adolescents' aspirations to work in things-oriented (e.g., mechanic), people-oriented (e.g., nurse), and STEM (e.g., mathematician) careers across 80 countries and economic regions using the 2018 Programme for International Student Assessment (PISA). We analyzed student career aspirations in combination with student achievement in mathematics, reading, and science, as well as parental occupations and family wealth. In each country and region, more boys than girls aspired to a things-oriented or STEM occupation and more girls than boys to a people-oriented occupation. These sex differences were larger in countries with a higher level of women's empowerment. We explain this counter-intuitive finding through the indirect effect of wealth. Women's empowerment is associated with relatively high levels of national wealth and this wealth allows more students to aspire to occupations they are intrinsically interested in. Implications for better understanding the sources of sex differences in career aspirations and associated policy are discussed.
... Su, Stoll és Rounds (2019) ugyanakkor továbbra is úgy látják, a fenti megfogalmazások továbbra sem adnak választ arra, hogyan alakul ki a preferencia az egyénben, mi dönti el, kedves-e számunkra az adott dolog, miért helyezünk valamit előbbre, mint a többit, van-e látens jelentése a válaszainknak, illetve változik-e idővel az érdeklődés. Korábbi kutatásokat áttekintve megkülönböztetik a szituációs érdeklődést mint pszichológiai állapotot, amely egy feladat vagy környezet által generált pillanatnyi kíváncsiság (Ainley, 2007, Hidi és Renninger, 2006, Krapp, 2007, Silvia, 2006, a diszpozíciós érdeklődéstől vagy hajlamtól, amely az egyén által preferált tevékenységekben és környezetekben mutatkozik meg (Ackerman és Heggestad, 1997, Holland, 1959, Lubinski, 2000, Savickas és Spokane, 1999, Rounds és Su, 2014. Előbbi az érdeklődés eredetét és funkcióját, míg utóbbi a fejlődését és megtapasztalását jelzi. ...