Intelligence

Ian J. Deary
Centre for Cognitive Ageing and Cognitive Epidemiology, Department of Psychology,
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

University of Edinburgh, Edinburgh EH8 9JZ, United Kingdom; email: i.deary@ed.ac.uk

Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Annu. Rev. Psychol. 2012. 63:453–82 Keywords

First published online as a Review in Advance on IQ, cognitive ability, psychometrics, behavior genetics, cognitive
September 19, 2011
epidemiology, twins, education, health
The Annual Review of Psychology is online at
psych.annualreviews.org Abstract
This article’s doi: Individual differences in human intelligence are of interest to a wide
10.1146/annurev-psych-120710-100353
range of psychologists and to many people outside the discipline.
Copyright  c 2012 by Annual Reviews. This overview of contributions to intelligence research covers the first
All rights reserved
decade of the twenty-first century. There is a survey of some of the
0066-4308/12/0110-0453$20.00 major books that appeared since 2000, at different levels of expertise
and from different points of view. Contributions to the phenotype of
intelligence differences are discussed, as well as some contributions
to causes and consequences of intelligence differences. The major
causal issues covered concern the environment and genetics, and how
intelligence differences are being mapped to brain differences. The
major outcomes discussed are health, education, and socioeconomic
status. Aging and intelligence are discussed, as are sex differences in
intelligence and whether twins and singletons differ in intelligence.
More generally, the degree to which intelligence has become a part of
broader research in neuroscience, health, and social science is discussed.

453
 PS63CH18-Deary ARI 31 October 2011 12:36

technology begun by Binet (1905). Some key

Contents questions that Spearman and Binet addressed
are still lively topics of research: Along which
INTRODUCTION . . . . . . . . . . . . . . . . . . 454
dimensions of mental abilities do people dif-
RECENT BOOKS ON
fer? Do these differences matter? And what
INTELLIGENCE . . . . . . . . . . . . . . . . 455
are the causes of these differences? These and
The Remarkable 1990s . . . . . . . . . . . . . 455
other questions—such as the effect of aging
Intelligence Books from the Past
on intelligence—are included in the present
Decade . . . . . . . . . . . . . . . . . . . . . . . . . 455
review.
THE PSYCHOMETRIC
One should be explicit about the difficul-
STRUCTURE OF
ties that accompany the topic of human intelli-
INTELLIGENCE . . . . . . . . . . . . . . . . 457
gence differences (which will normally, here, be
COGNITIVE CORRELATES OF
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

shortened to just intelligence). The study of

INTELLIGENCE . . . . . . . . . . . . . . . . 459
individual differences has never been in the
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

THE BIOLOGY OF
mainstream of psychology. With respect to psy-
INTELLIGENCE . . . . . . . . . . . . . . . . 460
chology’s two cultures—experimental and dif-
Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
ferential (Cronbach 1957)—differential is the
Brain Imaging . . . . . . . . . . . . . . . . . . . . . 465
small minority. Not only that, but the statistical
Fluctuating Asymmetry . . . . . . . . . . . . 467
procedures used by differential psychologists
PREDICTIVE VALIDITY OF
are relatively abstruse to those outside the field.
INTELLIGENCE . . . . . . . . . . . . . . . . 468
These factors—and the fact that intelligence-
Education, Occupation,
type tests have been used so widely in prac-
and Social Mobility . . . . . . . . . . . . . 468
tical settings and have been the subject of
Health, Illness, and Death . . . . . . . . . . 469
controversies—contribute to the range of at-
INTELLIGENCE AND AGING . . . . . 470
titudes that intelligence research attracts from
CONTROVERSY OR
lay and professional outsiders. These attitudes
CONSENSUS? . . . . . . . . . . . . . . . . . . . 472
include interest (research reports on intelli-
CONCLUDING REMARKS . . . . . . . . . 473
gence often attract much media interest), indif-
ference (much of mainstream psychology and
wider social science ignores individual differ-
ences in intelligence), and hostility (the emo-
INTRODUCTION tional heat generated by some aspects of in-
Some people are cleverer than others. The ways telligence research is matched by few other
in which this occurs, and the causes and conse- topics in psychology). However, this review
quences of these individual differences, are the also describes how researchers from a num-
topics of this review. ber of other disciplines—e.g., neuroscience and
It is some time since the Annual Review of epidemiology—are newly and keenly includ-
Psychology contained an article that dealt sub- ing intelligence as a topic in their research.
stantially with human intelligence differences. Intelligence is rarely discussed for long be-
Therefore, the period that is covered is broadly fore the word “controversial” appears; this is
the past ten years: the first decade of the twenty- another difficulty. Because there is contro-
first century. Historically, this is neat, because versy attached to some research topics in in-
the foundations of the scientific study of hu- telligence, it is important that there are clear
man intelligence differences were laid in the and even-handed accounts of what is known
first decade of the twentieth century. Then, and what is unknown about it. The present
there were the statistical developments, empir- piece attempts to be both, with examples of
ical discoveries, and conceptual innovations of influential studies and pointers to areas of
Spearman (1904) and the intelligence testing disagreement.

454 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

RECENT BOOKS Intelligence Books

ON INTELLIGENCE from the Past Decade

The Remarkable 1990s Because of space limitations, it is possible only

to select some books about intelligence from the
Books in the past 10 years followed a remarkable previous decade. Inevitably, someone’s favorite
decade, because the 1990s witnessed several book on intelligence will be missing. However,
important publishing events in intelligence. the following have been chosen because they
Carroll’s (1993) Human Cognitive Abilities was provide interesting and still-useful accounts of
his decades-in-preparation analysis of over different aspects of intelligence.
400 intelligence data sets that synthesized With regard to general books on intelli-
human cognitive differences in a three-level gence, there are two books for the lay person
psychometric hierarchy. Jensen’s (1998) The g or beginning student: Deary’s (2000) Intelli-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

Factor was a massive review of the construct of gence: A Very Short Introduction and Cianciolo
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

general intelligence ( g). Brand’s (1996) book of & Sternberg’s (2004) Intelligence: A Brief His-
the same name—The g Factor—was withdrawn tory are both elementary introductions to the
by Wiley after some of the author’s remarks in science of the area, and they also cover some
the press about group differences. The book of the controversies in intelligence research.
was given a postmortem review in Nature Bartholomew’s (2004) Measuring Intelligence is
(Mackintosh 1996), which commented that, a useful view of intelligence research from the
“This seems a singularly cack-handed attempt perspective of a statistician, making it more ob-
at censorship. . . How is it that they found out jective than most accounts. Probably the best
about the repellant nature of Brand’s views only book on human intelligence differences to ap-
after they had printed and distributed copies of pear for many years, Hunt’s (2011) Human In-
the book” (p. 33). Of these books, Carroll’s is telligence is superb. It is written by someone
singled out here for its usefulness in cementing almost equally at home in experimental and
a psychometric structure for intelligence that differential psychology, with an engaging wit,
brought “harmony where there had been comprehensive knowledge, and clear writing
discord,” although similar models had been style. Most of all, it is written with great even-
available for decades. However, the intelligence handedness; even for the controversial areas of
publishing event of the 1990s was Herrnstein & intelligence it is strongly recommended as an
Murray’s (1994) The Bell Curve, which spawned excellent account of the science of intelligence
an industry of mainly hostile reaction that con- research to date. It also has the merit of not
tinues into the twenty-first century. The book avoiding the technicalities involved in intelli-
is an unusual mixture of overview, empirical gence research, and it explains them well.
analyses, statistical tutorials, policy reflections, The previous decade saw some interesting
and appendixes and notes (approximately books on the history of intelligence. Carson’s
300 pages of the latter). For example, Part II of (2007) The Measure of Merit is a detailed look at
the book was almost 150 pages of analyses of the the origins of the idea of merit and mental test-
white people in the National Longitudinal Sur- ing, especially in France and the United States,
vey of Youth 1979. The authors ran regression and begins, historically, well prior to current
models that enquired about the relative im- intelligence testing technology. It is light on
portance of measured intelligence and parental the psychometric aspects and arguments of in-
socioeconomic status for important social telligence, but very careful and painstaking in,
outcomes such as poverty, education, employ- for example, describing how the military were
ment, injury, marriage, divorce, childbearing, won over to mental testing in World War I.
welfare dependency, parenting, crime, and Wright Gillham’s (2001) A Life of Sir Francis
citizenship. Galton: From African Exploration to the Birth of

www.annualreviews.org • Intelligence 455

 PS63CH18-Deary ARI 31 October 2011 12:36

Eugenics is probably the best biography of the areas of psychological research attract such at-
Victorian genius to date and includes his sem- tention from journalists and psychologists from
inal work on intelligence. Remarkably, this pi- outside their fields.
oneer of intelligence research and statistics was There have been several books on the
a significant contributor to many other scien- causes of intelligence differences. Deary’s
tific areas; so much so that, despite the impor- (2000) Looking Down On Human Intelligence:
tance of Galton’s contribution to intelligence From Psychometrics to the Brain was a relatively
and cognate topics, these take up a small section pessimistic survey of what had been con-
of the book. White’s (2006) Intelligence, Destiny tributed by studying cognitive components of
and Education is an intriguing examination of intelligence and also the relatively limited find-
the originators of intelligence testing. His the- ings that had come from biological approaches
sis is that these researchers were largely char- to intelligence during the twentieth century.
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

acterized by a nonconformist religious back- Geary’s (2005) The Origin of Mind: Evolution
ground, which in turn led them to emphasize of Brain, Cognition, and General Intelligence took
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

gifted and “feeble-minded” individuals (and less a more evolutionary approach and tried to fit
so the middle of the continuum), heredity, and intelligence differences within broader ideas in
the usefulness of examinations. Although in- the development of brain and mind. Garlick’s
triguing, the thesis is not convincing. (2010) Intelligence and the Brain: Solving the Mys-
In addition to history books there have also tery of Why People Differ in IQ and How a Child
been several of what might be characterized Can Be a Genius had a particular thesis concern-
as broadly anti-IQ-type books. These include ing white matter and processing speed and how
Murdoch’s (2007) IQ: A Smart History of a Failed they produce intelligence, but it was probably
Idea, Nisbett’s (2009) Intelligence and How to Get at too general a level and with too little empiri-
It, Shenk’s (2010) The Genius In All of Us: Why cal information to validate the interesting ideas
Everything You’ve Been Told About Genetics, Tal- fully. Duncan’s (2010) How Intelligence Happens
ent and Intelligence is Wrong, and Stanovich’s is a welcome account of intelligence from an
(2009) What Intelligence Tests Miss: The Psychol- experimental psychologist who acknowledges
ogy of Rational Thought. All are readable, often the existence and omnipresence of Spearman’s
highly so. Murdoch reviews the history and ma- g, with biological leanings. Flynn’s (2007) What
jor areas of controversy in intelligence research, Is Intelligence? was an interesting culmination
with clear antagonism toward IQ testing and of his work since he had demonstrated the
a preference for theories that emphasize mul- Flynn effect of rising intelligence. I have
tiple mental abilities. Stanovich, in particular, largely avoided edited books in this survey.
emphasizes aspects of reasoning that are poorly However, Wilhelm & Engle’s (2005) Handbook
correlated with intelligence and often lacking of Understanding and Measuring Intelligence and
in people who are intelligent. Shenk empha- Sternberg & Grigorenko’s (2002) The General
sizes gene-environment interactions in human Factor of Intelligence: How General Is It? both
life, downplays main effects of intelligence, and remain useful edited compilations of differ-
argues that the attainment of excellence may ent viewpoints on the causes of intelligence
be found more in effort. Of these three books, differences.
Stanovich’s and Shenk’s stand out as having the Books on intelligence that concern individ-
better empirical bases. However, the empirical ual, long-term cohort studies appeared in the
surveys take them where they will. Stanovich past decade. The latest monograph from Schaie
could easily have emphasized the value of ra- (2005)—Developmental Influences on Adult In-
tional thought without the negative reference telligence: The Seattle Longitudinal Study—was
to intelligence tests in the title. None of the an update on the Seattle longitudinal aging
four authors is a researcher on intelligence, and study. This cross-sequential study started in the
it is interesting to reflect on how many other 1950s, and this is an indispensable book about

456 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

cognitive aging. Deary et al.’s (2009c) A Life- tests is applied to a sample of the population,
time of Intelligence: Follow-up Studies of the some of the between-subject variation is shared
Scottish Mental Surveys of 1932 and 1947 was by all tests, some is shared by tests that have
a summary of 10 years of work on follow-up family resemblances within a cognitive domain,
studies of the population-wide intelligence tests and some is specific to the individual test. g of-
that took place in Scotland. Because these two ten accounts for nearly half the variance when
cover aspects of aging and intelligence, it should a broad battery of cognitive tests is applied to a
be noted that Salthouse’s (2010) Major Issues representative sample of the adult population.
in Cognitive Aging appeared recently and is a Relatively little of the variance lies at the do-
highly recommended source of information and main level. Researchers do not always agree on
toolkit for those concerned with aging aspects the nature of the domains—they can vary in
of intelligence. number, name and content between samples
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

Lynn & Vanhanen’s (2002) IQ and the depending on the battery applied—and there
Wealth of Nations attempted to analyze pros- have long been worries about whether the na-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

perity at the national level and IQ. This ture of g might vary between cognitive batteries.
type of global, country-level epidemiology has The latter worry was addressed directly us-
many critics. Lynn followed this up with three ing over 400 subjects from the Minnesota Study
other books on international aspects of intelli- of Twins Reared Apart ( Johnson et al. 2004).
gence and productivity. At the personal level The subjects had taken three large cognitive test
of success, Saunders’s (2010) Social Mobility batteries, originating from different theoretical
Myths summarized his and others’ research, orientations: the Hakstian and Cattell Compre-
with a conclusion that social mobility is far hensive Ability Battery (14 tests); the Hawaii
greater and more meritocratic in U.K. society Battery, with Raven’s Matrices added (17 tests);
than most people—especially politicians and and the Wechsler Adult Intelligence Scale (11
sociologists—think, and that mobility substan- tests). Each battery had a strong g factor, and the
tially depends on intelligence and effort. correlations among the three g factors—from a
hierarchical confirmatory factor analysis—were
0.99, 0.99, and 1.00. That is, the individual dif-
THE PSYCHOMETRIC ferences in g were identical from the three dif-
STRUCTURE OF INTELLIGENCE ferent batteries, leading the authors to title the
Following Carroll’s (1993) synthetic account paper “Just one g.” The result was replicated in
of the psychometric structure of intelligence, a sample of over 500 Dutch seamen who had
there has been a broad consensus that meaning- been tested on five different cognitive batter-
ful variance among people exists at three levels: ies ( Johnson et al. 2008c). These two papers
third-level general cognitive ability ( g), second- are important contributions to the psychome-
level broad domains of cognitive functioning tric structure of intelligence. As the authors
(group factors), and first-level test-specific vari- stated, “our results provide the most substan-
ation. To explain these levels simply, consider tive evidence of which we are aware that most
the answer to the following question: Why are psychological assessments of mental ability of
some people good at explaining the meanings any breadth are consistently identifying a com-
of words in their first language? The answers mon underlying component of general intelli-
are that people who are good at one mental gence. These results provide evidence both for
task tend to be good at other types of men- the existence of a general intelligence factor and
tal task (third level; g); people who are good for the consistency and accuracy of its measure-
within one domain (e.g., verbal ability) tend to ment” ( Johnson et al. 2008c, p. 91).
be good at other tasks in that domain; and peo- The subjects’ mental test data from the
ple have strengths in specific, narrow mental Minnesota Study of Twins Reared Apart were
skills. Thus, when a diverse battery of mental also used to tweak Carroll’s (1993) proposed

www.annualreviews.org • Intelligence 457

 PS63CH18-Deary ARI 31 October 2011 12:36

three-level hierarchy of intelligence differences operationalize Gardner’s (1983, 1993) multiple

( Johnson & Bouchard 2005a). The authors in- intelligences as tests and to examine their inter-
quired whether there might usefully be pro- correlations and correlations with a standard
posed some more general—but not as gen- psychometric intelligence test (Wonderlic
eral as g—latent traits between the several Personnel Test) (Visser et al. 2006). There
second-level factors (group factors, or cogni- were clear results. As operationalized by these
tive domains) and g, thus giving a four-level authors, most of the Gardner mental skills were
hierarchy. Carroll’s results had suggested no. correlated substantially with psychometric in-
The second-level factors were called crystal- telligence; formed a substantial g factor; and
lized (12 tests), fluency (4), fluid (12), mem- musical and body-kinesthetic intelligence were
ory (6), perceptual speed (11), and visualization more separate and intrapersonal intelligence
(14). Johnson and Bouchard’s proposed candi- harder to measure. One group of investigators
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

dates for new, third-level general factors were claimed to have found a “group intelligence”
the Cattell-Horn fluid and crystallized fac- that can explain how well a group performs
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

tors and Vernon’s verbal-educational (v:ed) and on tasks but is not just the mean or maximum
perceptual (k:m; spatial:mechanical) factors. of the individual general intelligence of the
Vernon’s factors arose because, contra Spear- members of the group (Woolley et al. 2010).
man, he thought that, “It may be concluded that Group intelligence was higher in groups
no test measures nothing but g and a specific where turn-taking in speaking was relatively
factor, since the type of test material employed evenly distributed among members and in
always introduces some additional common el- groups whose members had higher mean social
ement” (Vernon 1956, p. 144). Therefore, ac- sensitivity. Their practical suggestions were
cording to Vernon, the v:ed domain influenced that it might be easier to boost the intelligence
all tests involving verbal material and those that of a group than of an individual and that it
required the manipulation of words. The k:m might be useful to introduce group intelligence
domain influenced tests that required the men- testing for teams of workers. Replication of
tal manipulation of shapes, spatial imagination, this finding is necessary. The study has possible
and mechanical knowledge. However, better limitations in that the individual intelligence
fitting than any of these three models was one test was a single test, versus a range of tests
that included three factors at the third level: ver- for group intelligence, which, therefore, would
bal, perceptual, and image rotation. All three have more general variation and be more likely
loaded very highly on g. The authors called to be correlated with a criterion measure.
this the VPR model of intelligence. The ver- Furthermore, there was more overlap in
bal and perceptual, and perceptual and image content between the group intelligence test
rotation, factors correlated very highly, and ver- and the criterion tasks than looked likely with
bal and image rotation less so. The same four- the individual intelligence measure.
level model was tested versus the other three It is often stated that the major historical
using Thurstone’s data on 60 mental tests from challenge to Spearman’s conception of intelli-
1941 and again was found to fit best ( Johnson gence differences as being largely based on g
& Bouchard 2005b). was from Thurstone and his Primary Mental
Whereas the hierarchical structure of Abilities (PMAs). This is only partly correct.
intelligence differences does allow researchers It was clear from the late 1930s and certainly
and interested others to focus on the variance by the 1940s that the PMAs were not indepen-
accounted for in cognitive abilities at different dent and that Thurstone’s own data contained
levels of generality, there is still some impetus a statistical g factor. Probably the strongest
from those looking for an alternative formula- psychometric challenge to Spearman’s account
tion of intelligence or for additions to individual of intelligence differences was from Godfrey
general intelligence. There was an attempt to Thomson (Bartholomew et al. 2009). Thomson

458 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

never denied Spearman’s positive manifold of ies of intelligence. That is, a study may be
correlations among mental tests, but he sug- potentially more informative if it includes in-
gested a radically different reason for its oc- telligence phenotypes, a biological marker, and
curring. Instead of g—perhaps, according to an additional, potentially explanatory psycho-
Spearman, the result of people having gener- logical construct.
ally more or less of mental energy or power— Odd as it is to report, and in spite of there be-
Thomson found that the universally positive ing a hundred years of research in these areas,
correlations among tests could also arise from an indication of the true correlation between
each test’s sampling a subset of numerous, intelligence and sensory discrimination and re-
independent mental bonds; thus his “bonds” action time appeared only in the past decade.
or “sampling” theory of intelligence. The Spearman (1904) proposed what he termed a
Spearman-Thomson debates lasted from the functional correspondence between general in-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

First World War until almost the end of World telligence and sensory discrimination. Ninety-
War II. A fresh look at Thomson’s ideas con- seven years after this, it was reported, in a large
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

cluded that his model of intelligence was not in- sample (N = 899) of healthy adults in the
ferior to Spearman’s, either on statistical or bio- United States, that general intelligence (from
logical grounds, though that was partly because a battery of 13 tests) correlated 0.21 with pitch
both were vague biologically (Bartholomew discrimination and 0.31 with color discrimina-
et al. 2009). A related development is the mutual tion (Acton & Schroeder 2001). These bivari-
interaction model of intelligence, which also ate correlations between general intelligence
posits the emergence of a general factor without and sensory discrimination do not actually test
a general cause (van der Maas et al. 2006). The what Spearman (1904) hypothesized; namely,
basic idea is that a statistical g emerges through that whatever was common to discrimination
the mutual interaction, over the course of their measures was almost perfectly correlated with
development, of several cognitive processes. whatever is common to cognitive test measures.
Acton & Schroeder’s sample was reanalyzed us-
ing structural equation modeling, and the cor-
COGNITIVE CORRELATES relation between general intelligence and gen-
OF INTELLIGENCE eral sensory ability latent traits was 0.68; in a
Those taking a reductionist view of intelligence separate Scottish sample of children, it was 0.92
have not always gone straight down to biol- (Deary et al. 2004a). Spearman was substantially
ogy. Three levels of reduction have been visited correct, although we still do not know what
that stop short of, say, genetics or brain imag- causes the correlation between these two latent
ing: other psychometric measures; measures traits.
from experimental psychology, especially reac- With regard to reaction time, the first large
tion times; and measures from psychophysics, (N = 900) population-representative study of
including inspection time. With the increased its correlation with intelligence found, in a very
accessibility of brain imaging, studies that use narrow age cohort about 56 years old, as fol-
these sorts of tasks and their related constructs lows: four-choice reaction time mean = −0.49;
along with intelligence are probably declining four-choice reaction time intraindividual vari-
from a peak between the 1970s and 2000 (Deary ability = −0.26; simple reaction time mean =
2000). Much interest has focused, in the broadly −0.31; and simple reaction time intraindivid-
psychometric-experimental levels, on process- ual variability = −0.26 (Deary et al. 2001).
ing speed and working memory as potential ex- A series of empirical reviews—but not for-
planatory variables for intelligence. However, mally conducted or presented meta-analyses—
there are new findings to report from each of of processing speed and intelligence correla-
these areas, some of which are being incorpo- tions gathered findings from 172 studies con-
rated within brain imaging and genetic stud- taining over 50,000 subjects in total (Sheppard

www.annualreviews.org • Intelligence 459

 PS63CH18-Deary ARI 31 October 2011 12:36

2008). The correlation was slightly lower than might reflect a correct judgment that explana-
that of Deary et al. (2001) but included con- tions have not been and will not be forthcoming
venience samples; that is, samples of partici- from that research route. On the other hand,
pants where no attempt is made to match to it might be that researchers in intelligence
population characteristics and that often in- have prematurely left this field to explore other
volve students or other relatively cognitively routes that promise more by way of a reduction-
homogeneous groups. Much of Sheppard’s ac- istic account of intelligence differences; for ex-
count concerns cognitive components—such ample, in brain imaging and molecular genetics.
as the reaction time for processing in short-
term memory or speed of retrieval from long-
term memory—but I consider these to be a THE BIOLOGY OF
distraction when the correlation with straight- INTELLIGENCE
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

forward choice reaction time is so high. The Around 2000, there were two overviews of
cognitive components claimed to be isolable the biological correlates of human intelligence
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

from reaction time models tend to be unreli- differences (Deary 2000, Jensen 1998). They
able and not, in any case, to improve the cor- dealt with the same assortment of biological
relation with intelligence (Deary 2000, chap- tools: genetics (behavioral and molecular stud-
ter 6). This was emphasized by Lohman (1994) ies), brain imaging (structural and functional),
in his overview of such attempts: “attempts to the brain’s electrical responses (analyses of
isolate component scores that decompose in- the electroencephalograph [EEG] and evoked
dividual differences on homogeneous tasks into responses), nerve conduction velocity, and an
process measures cannot succeed, and so our ef- assortment of less-studied approaches. Little
forts should be directed elsewhere” (p. 9). The was certain at that time. Two things seemed
mean correlation between general intelligence relatively firm: People with higher measured
and visual inspection time—a psychophysical intelligence tended to have larger brains,
task that does not involve reaction speed but, in- and intelligence differences had a substantial
stead, records correct discriminations based on heritability. An emerging finding was that
a simple stimulus that is presented for different older people with the e4 allele of the gene for
durations—was −0.36 (Sheppard 2008); people apolipoprotein E (APOE) tended to have lower
with higher intelligence test scores were more cognitive ability. Apart from these findings,
efficient in accumulating accurate information most approaches suffered from lack of repli-
from briefly presented stimuli. The equivalent cability. There were many, mostly modestly
correlation with auditory inspection time was sized, studies reporting correlations between
−0.31. a biological variable and intelligence, but
Therefore, the current situation is that typically these were either not replicated or at-
apparently lower-level mental tasks—such as tempted replications tended to be too different
sensory discrimination, visual processing and to be characterized as such (Deary 2000, Jensen
reaction time—have fairly well-established sig- 1998). EEG and brain-evoked response studies
nificant and far-from-trivial correlations with suffered particularly in that regard. In looking
intelligence. Less is known about why these cor- forward to the research after 2000, my opinion
relations occur, and that is largely because of was that, “it is tempting to say that the hope lies
lack of understanding of the causes of individual mainly in the less-put-to-the-test approaches
differences in these so-called elementary cog- of functional brain scanning and molecular
nitive tasks. Although reaction time tasks are genetics” (Deary 2000, p. 312). This prognos-
quite widely applied in mental testing, current tication, made a decade ago, is both correct
research is in a state whereby there is less inter- and wrong. Probably it is correct to have
est in these sorts of tasks to explore the origins of identified brain imaging and genetics as the
intelligence differences. On the one hand, this two techniques that would be most applied to

460 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

human intelligence research and would deliver The broad conclusions are that the heritability
the most solid findings. However, it backed the of intelligence is now well established, with
wrong horse in both cases. Behavioral genetics some important moderation by age and other
and structural brain imaging have added more factors; that multivariate behavior genetic
than their respective molecular and functional studies have been informative about the causes
counterparts in our search for the causes of of correlations between intelligence and some
human intelligence. A recent review of the other variables with which intelligence corre-
biological foundations of intelligence—from lates; and that molecular genetic studies—with
the point of view of neuroscience—covers the exception of variation in APOE—have
behavioral and molecular genetic studies and yet to identify variations in specific genes
structural and functional brain imaging, and it that are firmly associated with intelligence
provides more detail than space allows in the differences.
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

present account (Deary et al. 2010a).

Behavior genetics. Individual studies from
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

the past decade may be used to exemplify im-

Genetics portant advances or consolidations in the un-
This overview of advances in the genetic con- derstanding of the environmental and genetic
tributions to human intelligence differences contributions to intelligence differences.
divides the research into behavioral and molec- Much of the additive genetic effect on in-
ular approaches. Behavioral studies use twin-, telligence is attributable to the general cogni-
adoption-, and family-based designs to obtain tive ability ( g) factor, as is much of the genetic
estimates of the proportion of the population influence on specific cognitive tests. A Dutch
variance in intelligence caused by genetic study administered the 11 subtests of the Wech-
(mostly additive) differences and by the shared sler Adult Intelligence Scale and Raven’s Ma-
and nonshared aspects of the environment. trices (a test of nonverbal reasoning) to 194
Molecular genetic studies use candidate gene twin pairs in their later teens (Rijsdijk et al.
or genomewide association techniques. Candi- 2002). The variation in full-scale IQ (almost
date gene studies are hypothesis driven. On the equivalent to g) attributable to additive genetic
basis of prior findings or on the known function factors was 82%. The additive genetic con-
of a gene’s protein, a gene that is polymorphic tributions to individual tests were examined
(has multiple alleles and might thereby express according to the psychometric three-level hi-
different phenotypes) is selected and a test is erarchy, as described by Carroll (1993). Addi-
made to discover whether people with different tive genetic contributions to the g factor, which
alleles tend to differ on intelligence generally capture the variance shared by all 12 tests, ac-
or on specific cognitive domains. Genomewide counted for a mean of 31% (range = 8 to 53) of
association studies (widely referred to using the variation in individual test scores. Additive
the acronym GWAS) are hypothesis free: One genetic contributions to cognitive domains—
tests the association between a phenotype (e.g., verbal comprehension, freedom from distrac-
an intelligence test score or g factor) and a tion, perceptual organization—accounted for
large number of genetic variants spread across a mean of 12% (range = 1 to 30). Additive
the human genome. Then one tries to find genetic contributions to test-specific variation
out which of the many nominally significant accounted for a mean of 14% (range = 0 to
associations are replicable and then what they 38). Therefore, these diverse mental tests corre-
might mean mechanistically (if they replicate). late phenotypically largely due to genetic causes
There have been at least five reviews of the ge- of g. In the VPR model of intelligence, based
netic studies of human intelligence in the past on data from the Minnesota Study of Twins
few years (Deary et al. 2006, 2009b; Lee et al. Reared Apart, the additive genetic contribu-
2010; Payton 2009; Plomin & Spinath 2004). tions were calculated for general intelligence

www.annualreviews.org • Intelligence 461

 PS63CH18-Deary ARI 31 October 2011 12:36

( g) and for cognitive domains at different levels When extended to adult samples, Dutch twin
of generality-specificity ( Johnson et al. 2007). family studies demonstrate that the percentages
For g, at the fourth stratum in their hierarchi- of variance accounted for by additive genetic
cal model, which contained the variance shared effects rise to over 80% for verbal IQ and al-
by all tests, additive genetic causes accounted most 70% for performance IQ (Posthuma et al.
for 77% of the variance. The third stratum, 2001). Lower estimates, however, were indi-
one step below g, had factors that accounted cated from the results of the Vietnam Era Twin
for variance shared by tests drawing on broad Registry, which estimated that genetic factors
verbal, perceptual, and image rotation abilities, cause 49% of variation in the Armed Forces
and the additive genetic contribution to these Qualification Test in young adulthood and 57%
was 78%, 77%, and 76%, respectively. The in late middle age, with the genetic effects also
second stratum had still-more-specific factors explaining almost all of the 0.74 correlation be-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

accounting for variance shared by tests draw- tween the two administrations of the test in this
ing on the following abilities, with the percent- longitudinal study (Lyons et al. 2009). Results
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

ages of variation accounted for by additive ge- from repeated cognitive testing of participants
netic causes given in parentheses: narrow verbal in the Swedish Adoption Twin Study of Aging
(79%), scholastic (69%), fluency (79%), num- indicate that genetic influences remain substan-
ber (72%), content memory (33%), perceptual tial into old age, with a lowering of the contri-
speed (67%), spatial (76%), and rotation (75%). bution (though it is still the majority) at about
There are two remarkable points here. The 80 years (Reynolds et al. 2005). A review of
first is the consistently high genetic contribu- twins studies with older people estimated the
tion to variance at the second to fourth strata. heritability of g to be about 80% at 65 years
Of course, it must be emphasized that much of and about 60% at age 82 (Lee et al. 2010).
the genetic influence on the second and third In old age, the genetic contribution is al-
strata derives from the genetic influence on g, most entirely to the level (intercept) of in-
because factors at these levels load highly on it. telligence rather than its change (slope) (Lee
The second is the relatively high environmental et al. 2010). Latent growth curve models of the
contribution to content memory. Swedish Adoption Twin Study of Aging showed
The heritability of intelligence is not the that genetic effects were largely to the inter-
same at different ages. A Dutch study with 209 cepts and not the slopes. Any small genetic ef-
pairs of twins examined intelligence at ages 5, fect on the slope tended to be on the smaller
7, and 10 years (RAKIT battery) and again quadratic rather than the much larger linear ef-
at 12 years (Wechsler Intelligence Scale for fect (Reynolds et al. 2005). Among the cogni-
Children-Revised) (Bartels et al. 2002). Across tive domains, the genetic effects on memory in-
these four ages, from 5 to 12, the percent- creased with age, but those on processing speed
ages of variation in intelligence accounted for and fluid ability decreased somewhat.
were as follows: additive genetic effects = 26, The heritability of intelligence might be
39, 54, 64; shared environment effects = 50, moderated by environmental factors. This idea
30, 25, 21 (the latter three values were non- was tested in 229 pairs of seven-year-old
significant); and unique environment effects twins from the National Collaborative Perina-
(includes error) = 24, 31, 21, 15. Similarly, the tal Project, in which over half the sample was
Twins Early Development Study in the United black and there was a high proportion of poor
Kingdom found that, for general intelligence, families (Turkheimer et al. 2003). To give just
the heritability was 23% in early childhood one interesting result, when socioeconomic sta-
(with shared environment accounting for 74% tus was split into high and low, the percentage
of the variance) and that this increased to 62% of variance in intelligence accounted for by ge-
by middle childhood (with shared environment netic factors was 71% and 10%, respectively.
only 33% by this stage) (Davis et al. 2009). The shared environment effects were 15% and

462 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

58%, respectively. Life events have also been volume is highly heritable (Thompson et al.
found to moderate genetic and environmental 2001), and it correlates modestly with intelli-
contributions to general intelligence variation gence (McDaniel 2005). A bivariate genetic-
in adults (Vinkhuyzen et al. 2011). environmental study of this correlation showed
Multivariate genetic-environmental studies that the correlations between intelligence and
have been useful in exploring the causes of the total volumes of gray matter and white mat-
the associations between intelligence and some ter were caused by genetic factors (Posthuma
of its correlates. To explain, behavior genetic et al. 2002).
methods may be used to describe the propor-
tions of environmental and genetic influences Molecular genetics. With almost-equal jus-
on single phenotypes, and they can also be used tification, this section of the overview could
to describe the proportions of environmental be very long or very short: very long, because
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

and genetic influences on the correlations be- dozens of candidate genes have been reported
tween two or more phenotypes. For example, an as being associated with intelligence; very short,
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

adolescent twin sample of over 500 twin pairs because almost none of them has been repli-
examined the correlations between psychome- cated. As a compromise, this section attempts to
tric intelligence and processing speed measures steer a course between the two extremes of giv-
including choice reaction time and inspection ing a list of type 1 statistical errors and bleakly
time (Luciano et al. 2004a). A general ge- stating that we, as yet, know nothing about the
netic factor influenced intelligence and process- genes that influence intelligence differences.
ing speed tests. More specific genetic factors Candidate gene studies of intelligence dif-
accounted for some test covariance and test- ferences have been disappointing. A review sur-
specific variance. Environmental effects were veyed a period of about 14 years in which there
mostly nonshared and test specific. A bivariate were more than 200 studies on approximately
environmental-genetic study using Australian 50 genes with polymorphisms that might be
and Dutch participants in a twin family de- related to intelligence (Payton 2009). It was
sign explored causal hypotheses concerning the concluded that, as yet, no individual genetic
correlation between inspection time and intel- variants are conclusively related to intelligence
ligence; in the literature there were compet- or its change with age in healthy individuals.
ing suggestions about which caused the other An exception is the gene for apolipoprotein E
(Luciano et al. 2005). The best model was one of (APOE), which is involved in cholesterol trans-
pleiotropy; that is, a common set of genes influ- port and neuronal repair. Possession of the e4
ences both intelligence and inspection time, but allele for this gene was the subject of a meta-
neither of these variables mediated the genetic analysis of 77 studies with nearly 41,000 healthy
influence on the other. For processing speed individuals (Wisdom et al. 2011). People with
and other cognitive models of intelligence this the e4 allele have, on average, lower general
is instructive: It suggests that these so-called el- cognitive function, with an effect size (d ) of
ementary cognitive tasks are aspects of g rather about −0.05. The effect is found in late-middle
than causes of it. Strong genetic correlations ex- and old age, and the effect appears to strengthen
ist between IQ scores from the Wechsler bat- with age. There may also be a very small effect
tery and the heavily genetically influenced gen- of a polymorphism in the gene for catechol-
eral executive function and the updating specific O-methyl transferase (Barnett et al. 2008). A
aspect of executive function (Friedman et al. meta-analysis of polymorphisms in the gene for
2008). There is mixed evidence about whether dystrobrevin-binding protein 1, which is asso-
the small but consistent correlation between ciated with risk of schizophrenia, showed ef-
intelligence and birth weight (in the normal fect sizes for two single-nucleotide polymor-
range; Shenkin et al. 2004) is caused by shared phisms around the 0.1 value (Zhang et al. 2010).
genetic factors (Luciano et al. 2004b). Brain It is also worth mentioning the Val66Met

www.annualreviews.org • Intelligence 463

 PS63CH18-Deary ARI 31 October 2011 12:36

polymorphism of the gene for brain-derived replicable discoveries to be made, the scale of
neurotrophic factor, which might have a small this research will have to be far larger (Cirulli
effect on intelligence, although there are un- et al. 2010, Davis et al. 2010).
certainties about the direction of association Through the use of a novel statistical
(Payton 2009). technique applied to genomewide data, the
A fascinating molecular genetic contribu- first purely biological evidence for the sub-
tion was an interaction between variants in the stantial heritability of fluid and crystallized
gene for fatty acid desaturase 2 and the effect of general intelligence has appeared (Davies et al.
breastfeeding on children’s intelligence (Caspi 2011). The study analyzed more than 500,000
et al. 2007). Though this was shown in two co- single-nucleotide polymorphisms in over 3,500
horts, another large cohort has not replicated older people from Scotland and England. The
the direction of association (Steer et al. 2010). authors used a new technique that employs
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

And, though maternal intelligence did not ap- information from all of the half-million-plus
pear to explain the effects in the original study, genetic variants to calculate “relatedness”
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

a meta-analysis of breastfeeding effects on chil- in these unrelated individuals. The method

dren’s intelligence did show that maternal in- afforded an estimate of the narrow-sense
telligence accounted for much of the appar- heritability of intelligence based purely on
ent protective effect of breastfeeding, leaving these biological data. The heritability estimate
it nonsignificant overall (Der et al. 2006). was 40% for crystallized intelligence and 51%
With regard to GWAS, the past decade has for fluid general intelligence. This means that
seen a remarkable maturing of genetic studies unknown causal genetic variants in linkage dis-
of intelligence. The first genomewide scans for equilibrium with the assessed single-nucleotide
intelligence used a few to many hundreds of polymorphisms account for much of the
microsatellite markers and hundreds of sibling genetic variation in intelligence. The results
pairs or families in linkage analyses (e.g., point to the possibility of many genes of small
Luciano et al. 2006, Posthuma et al. 2005). effect contributing to human intelligence vari-
Linkage refers to the fact that genetic loci that ation. Moreover, the authors were also able to
are relatively close to each other on the same use the purely biological information from the
chromosome tend to be inherited together. Scottish and English samples significantly to
Fundamentally, then, linkage analyses explore predict about 1% of the variance in intelligence
whether, among related individuals, certain scores in an independent Norwegian sample.
genetic marker loci tend to co-occur with phe- The best genetically oriented research on in-
notypic characteristics. These types of analyses, telligence now appreciates that individual gene
for example, indicated the possible importance effects are likely to be small or very small; exam-
of variants in the chromosome 6p region. Link- ines hundreds of thousands of single-nucleotide
age designs have, for the present, been pushed polymorphisms on every participant (and, in-
to the periphery of genetic research on intel- deed, uses the fact that single-nucleotide poly-
ligence with the coming of so many GWAS morphisms that occur quite close together
that test unrelated subjects. However, linkage show linkage disequilibrium—that is, there is
designs might well reappear in the search for a greater than zero chance of certain variants
rare variants that influence intelligence. The being inherited together—to impute these ef-
first genome-wide studies of intelligence to fectively to provide a few million); gathers in
use hundreds of thousands of single-nucleotide consortia that can bring together several or even
polymorphisms have appeared. Even with this many thousands of participants; often involves
density of genotyping, a few hundreds or thou- researchers with a primary interest in medi-
sands of subjects, and attempts at replication, it cal or other social variables but who happened
is becoming clear that the effects of individual to have collected intelligence data on their
genetic variants will be very small, and for participants; and has rigorous approaches to

464 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

significance and replication. This has arisen af- safety and relatively low cost have meant that
ter finding that no studies to date at the smaller almost all the brain imaging studies of intel-
scale—either candidate gene or genomewide— ligence have used MRI. Below, some studies
have found large or medium effects. The same exemplifying key empirical advances in the as-
maturing of studies has occurred with other sociation between brain structure and func-
quantitative traits, such as height. A consortium tions are described. First, though, attention is
of studies on height, which reported results on directed toward an adventurous and helpful
over 180,000 individuals, found significant ef- synthesis of brain imaging–intelligence work.
fects at more than 180 genetic loci, which to- The parieto-frontal integration theory of in-
gether explained 10% of the phenotypic vari- telligence (P-FIT) brought together 37 studies
ation (Lango Allen et al. 2010). At present, at that had employed neuroimaging techniques to
least three international consortia of studies are investigate differences in intelligence and rea-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

conducting genomewide association analyses of soning ( Jung & Haier 2007). They included
intelligence-related phenotypes. All have par- structural and functional MRI, diffusion ten-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

ticipant numbers well into five figures; none has sor MRI, magnetic resonance spectroscopy, and
reported findings to date. positron emission tomography studies, all of
If it is the case that genomewide searches which they thought could be used to sketch
for intelligence do not throw up any large or a coherent account of what it means to have
medium effects—or even a collection of smaller a brighter brain. Congruent brain imaging–
effects that seem to account for most of the ad- intelligence associations overlapped on how the
ditive genetic effect that is apparent from be- association cortices were linked by key white
havioral genetic studies—then the genetic con- matter pathways (particularly the arcuate fas-
tributions lie elsewhere. One option is in a very ciculus and superior longitudinal fasciculus). In
large number of very small effects, as suggested the P-FIT account, abstraction and elaboration
by Davies et al. (2011). A second is in rare vari- of incoming sensory information are conducted
ants (Penke et al. 2007), where the idea is that in the parietal cortex; the parietal interacts with
the accumulation of these across generations the frontal cortex, which tests hypotheses con-
causes differences in intelligence. Another is in cerning a problem; following a best solution,
what are called copy number variations: dele- the anterior cingulate constrains the selection
tions and duplications of stretches of DNA. Al- of responses and inhibits competing ones; and
though based on very few subjects and requiring the whole process depends for its efficiency on
replication, one study examined this possibility intact white matter connections between the re-
and found that people with more rare deletions gions. In addition to the P-FIT paper—and its
had lower Wechsler Adult Intelligence Scale associated discussion—readers are directed to
scores (Yeo et al. 2011). a recent special issue of the journal Intelligence
that was devoted to brain imaging, as intro-
duced by Haier (2009).
Brain Imaging
In the early years of brain imaging and Structural brain imaging. Only in the past
intelligence—the 1980s and 1990s—there were decade has there been enough evidence to con-
studies using positron and single-photon emis- clude with confidence something that had been
sion tomography. Both techniques were ex- mooted, debated, and, at times, ridiculed for
pensive and involved the administration of over a century: People with larger brains do
radioactive substances to subjects. These tech- tend to have higher intelligence test scores. In a
niques have been eclipsed by magnetic reso- meta-analysis of 37 samples examining whole-
nance imaging (MRI) in its various structural brain volume in healthy subjects (total N =
and functional forms. Although it is still not 1,530), the raw correlation was 0.29 (0.33 af-
inexpensive to scan each subject’s brain, its ter correction for range restriction; McDaniel

www.annualreviews.org • Intelligence 465

 PS63CH18-Deary ARI 31 October 2011 12:36

2005). A more recent review reckoned the correlations between g from the Wechsler
correlation between whole-brain size and gen- Adult Intelligence Scale and gray matter
eral intelligence is about 0.4 and that the cor- volumes in the frontal, temporal, occipital, and
relation between external head size and intelli- some sublobar (lentiform nucleus, thalamus,
gence (based on 59 samples with a total N = etc.) brain regions (Colom et al. 2006). More
63,405) is 0.20 (Rushton & Ankney 2009). sophisticated than these cross-sectional studies,
That being established, there has been a study of 307 children aged 7 to 19—who
interest in whether certain brain regions and had been imaged between one and three times
the pathways between them are associated with a two-year interval—investigated the
with intelligence differences—see the P-FIT association between general intelligence and
theory above ( Jung & Haier 2007)—and more the developmental trajectory of brain cortical
generally the extent to which intelligence is thickness (Shaw et al. 2006). An examination
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

localized or distributed through the brain. of their findings—where, for illustration,

One issue has been whether intelligence is intelligence was divided into superior, high,
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

more strongly associated—and where in the and average groups—revealed that the superior
brain—with gray or white matter. Certainly, intelligence group had a distinct trajectory
gray matter volume seems to be correlated sig- whereby their cortical thickness was the lowest
nificantly with intelligence. In 216 children and of the three groups at age 7, highest at about
adolescents, there were positive correlations age 12, and average again at age 19.
between general intelligence (from a short The above studies were performed on
Wechsler battery) and brain cortical thickness healthy individuals. Validation of the princi-
distributed through frontal, parietal, temporal, pal P-FIT ideas was found from an innovative
and occipital brain regions (Karama et al. 2009). study of 241 patients with brain lesions who
These were in agreement with, if partly more underwent brain imaging (with voxel-based
extensive than, the regions identified in the lesion-symptom mapping) and were tested on
P-FIT theory. Data from 65 men and women the Wechsler Adult Intelligence Scale battery
showed that cortical thickness (gray matter) (Gläscher et al. 2010). Conceptually, this is
was associated with intelligence, particularly what they did: They divided the brain into vox-
in the prefrontal and posterior temporal areas els and asked what the association was between
(Narr et al. 2007). The study had also found an having a lesion in a given voxel and the score ob-
intelligence–overall brain volume correlation tained on the g factor from the Wechsler bat-
of 0.36 (0.37 for overall gray matter; 0.26 tery. The significant associations with g were
for intracranial white matter volume). This with damage in a frontal-parietal network and
correlation between overall brain volume and the white matter tracts that connected them,
intelligence is almost identical to McDaniel’s with the superior longitudinal/arcuate fascicu-
(2005) estimate and that from a study of over lus again being prominent.
200 young adults, in which the correlation was The P-FIT ideas include both gray and
0.35 (Choi et al. 2008). Intelligence scores and white matter in the brain as contributing intelli-
voxel-based morphometry analyses of brain gence differences. The past decade has seen in-
MRI data from twins provided evidence for an creasing evidence of correlations between white
intelligence differences–associated network of matter–related brain-imaging variables and in-
frontal-occipital-parahippocampal gray matter telligence. People with more lesions in the
and connecting white matter of the superior white matter tend to have lower intelligence, al-
occipito-frontal fasciculus and corpus callo- though this has been found—with small effect
sum (Hulshoff Pol et al. 2006). Investigators sizes—mainly in older people, in whom these
using voxel-based morphometry analysis of lesions tend to accumulate and can be rated
MRIs of 48 adult human brains to separate using brain MRI images (Frisoni et al. 2007).
gray and white matter identified positive Eleven studies show, overall, some evidence

466 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

of an association between n-acetylaspartate—a (Raven’s Advanced Progressive Matrices) and

brain metabolite measured by magnetic res- the ability, correctly, to ignore “lure” stimuli
onance spectroscopy—and cognitive abilities, in a 3-back working memory task. Gray and
but the associations do not always go in the same colleagues (2003) concluded this was “the first
direction ( Jung et al. 2009). The development direct support for a major hypothesis about the
of diffusion tensor MRI has meant that indices neurobiological basis of gf ” (p. 319). An attempt
of brain white matter integrity can be produced to replicate this result beyond lure detection
in vivo. Several studies show that people with in an n-back task was not successful (Waiter
higher intelligence tend to have greater white et al. 2009). Both of these studies had interest-
matter integrity—typically assessed using a pa- ing findings based on well-aimed hypotheses.
rameter called fractional anisotropy—which ac- However, their ambitious titles must be read
cords with a distributed and connectionist view with appropriate skepticism: They appear to
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

of what it means to be higher in intelligence offer too broad a conclusion from small stud-
(Deary et al. 2010a provide an overview). Brain ies that are the beginnings rather than the end
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

white matter integrity is highly heritable, and of a long research effort to explain intelligence
the modest phenotypic correlation with intel- differences.
ligence appears to be caused by shared genetic A very valuable integrative review, which in-
factors (Chiang et al. 2009). cluded functional brain imaging and EEG tech-
niques, converges on an account of the intelli-
Functional brain imaging. Early in the past gent brain that is distributed (Neubauer & Fink
decade there were two high-profile articles that 2009). It also provides much diverse support
made many intelligence researchers sit up and for the view that the intelligent brain is more
take notice because of their striking titles: “A efficient.
neural basis for general intelligence” (Duncan
et al. 2000) and “Neural mechanisms of gen-
eral fluid intelligence” (Gray et al. 2003). In the Fluctuating Asymmetry
first study, a PET experiment on only 13 sub- On the basis of the idea that intelligence is an
jects devised high- and low-g-demanding ver- indicator of fitness, there has been consider-
sions of three tasks and looked for the brain able activity in examining whether it associated
regions that were consistently associated with with other aspects of bodily fitness. One aspect
greater activation in the high-g task versions. is health, and that is covered here under the
The answer was the lateral frontal cortex, and section on cognitive epidemiology; another is
the authors argued against a diffuse neural re- fluctuating asymmetry (Van Valen 1962): the
cruitment mechanism for g differences, but degree to which the same bodily parts on the
instead, “g reflects the function of a specific two sides of the body show an absolute devia-
neural system, including as one major part a tion from being identical in size. This is cov-
specific region of the lateral frontal cortex” ered here because it is an aspect of research
(Duncan et al. 2000, p. 459). It is clear from into the biology of intelligence that emerged
the P-FIT theory ( Jung & Haier 2007) and only in the previous decade or so. Except for
lesion-based studies (Gläscher et al. 2010) that the original report of two samples (Furlow et al.
Duncan did identify an important brain region 1997)—each with just over 100 participants—
for intelligence and that it is just one region in which intelligence correlated just over 0.2
in a network. In the second study (Gray et al. with symmetry, all studies have appeared in the
2003), the straightforward result was that neu- twenty-first century. A meta-analysis of 14 sam-
ral activity (inferred from the blood oxygen– ples (published and unpublished) with a total
level dependent signal in functional MRI) in N of 1,871 estimated that the correlation be-
the lateral prefrontal cortex almost completely tween intelligence and fluctuating asymmetry
attenuated the association between intelligence was −0.12 to −0.20 (Banks et al. 2010). That

www.annualreviews.org • Intelligence 467

 PS63CH18-Deary ARI 31 October 2011 12:36

is, people with higher intelligence test scores total number of studies, total number of sub-
tend to be more symmetrical. To explain these jects) as follows: education = 0.46 (0.36 to
findings, Banks et al. appealed to an account 0.75, 59, 84,828); occupation = 0.37 (0.28 to
which suggests that both intelligence and sym- 0.57, 45, 72,290); and income = 0.21 (−0.01
metry are markers of a general fitness latent trait to 0.40, 31, 58,758) (Strenze 2007). With only
that is associated with survival and reproduc- the objectively defined better studies included,
tive success (Miller 2000, Keller & Miller 2007). with sample-size weighting, and with correc-
However, a first attempt to examine the genetic tion for unreliability and dichotomization, the
correlation between intelligence and symmetry effect sizes were estimated as, respectively, 0.56,
found neither a genetic nor a phenotypic cor- 0.45, and 0.23. In this meta-analysis, education
relation ( Johnson et al. 2008b). was assessed using educational level. If objec-
tive results are used from national examina-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

tions, the correlation between intelligence and

PREDICTIVE VALIDITY education is considerably higher. For example,
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

OF INTELLIGENCE the prospective correlation between the gen-

Intelligence predicts important things in life. eral intelligence latent trait from the Cogni-
The predictive validity of intelligence for ed- tive Abilities Test at age 11 years and a general
ucation, occupational success, and social mo- educational latent trait (based on English na-
bility was well documented prior to the past tional General Certificate of Secondary Educa-
decade. That said, there was some new re- tion scores at age 16 years; N = 13,248) was
search in these areas. Some progress was also 0.81 (Deary et al. 2007a). By way of balance—
made in assessing the association between intel- albeit in a study with two orders of mag-
ligence and related personal-social constructs. nitude fewer subjects—it is noted that self-
A useful meta-analysis showed that the correla- discipline (as rated by the person or by others)
tion between intelligence and ego-development can have stronger associations with educational
stage is between 0.20 and 0.34, making them outcomes than intelligence has (Duckworth &
related but not identical concepts (Cohn & Seligman 2005). There is also evidence for some
Westenberg 2004). Long-term follow-up stud- people’s being more motivated than others in
ies of the British cohorts born in 1958 and 1970 certain intelligence testing situations, which
showed strong associations between higher could inflate intelligence–life outcomes corre-
childhood intelligence and more socially lib- lations (Duckworth et al. 2011). In another in-
eral attitudes at age about 30 years (Deary et al. vestigation of real-life outcomes of intelligence,
2008, Schoon et al. 2010). The most novel con- a large study of trainee truckers showed that
tribution of intelligence as a predictor was the intelligence was associated with “preferences
emergence and growth of the new research field and choices in ways that favor economic suc-
of cognitive epidemiology; the first years of the cess” (Burks et al. 2009). Those with higher
twenty-first century established intelligence as intelligence were more patient, were better at
a predictor of health, illness, and death. taking calculated risks, were better at predict-
ing how other people would act and how they
should act as a result, and persevered longer
Education, Occupation, in a job when there was a financial penalty for
and Social Mobility leaving.
Although they have been studied for many Much remains to be discovered about so-
years, it is useful to see a meta-analytic ap- cial mobility. The United Kingdom, because
proach to the associations of intelligence with it has various prospective cohorts that have
some of its well-known correlates in the field been studied from youth and are now at various
of socioeconomic success. Intelligence had av- stages of adulthood and old age, has been es-
erage correlations (95% confidence intervals, pecially informative, but of course these results

468 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

do not necessarily generalize to other countries of its determinants to any extent. That has
or cultures. However, even using the same co- changed. Two of the prominent social corre-
hort’s data, different researchers have come to lates of intelligence—education and social class
different conclusions. For example, consider (parental and own in adulthood)—were already
some analyses from the past decade on the known to be associated with health inequali-
U.K.’s National Child Development Study (the ties; therefore, it is not surprising to see in-
1958 British Birth Cohort), which gathered telligence added to the list. The first journal
prospective data on all children born in Great report of an intelligence-death association was
Britain in one week in March 1958. Some re- with Australian male Vietnam veterans; those
searchers emphasized that, with respect to so- whose mental test scores were lower on en-
cial class destinations at age 33, there were still try to the armed services were more likely to
substantial effects of parental social class after have died (principally from external causes) by
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

adjusting for intelligence and academic effort midlife (O’Toole & Stankov 1992). Subsequent
(Breen & Goldthorpe 2002). With the same findings took place in the next decade. A follow-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

data, others emphasized that whereas parental up of over 2,000 subjects of the Scottish Men-
social class accounted for about 25% of people’s tal Survey of 1932 (of people born in 1921)
own social class at age 33, intelligence, motiva- found that a standard deviation disadvantage in
tion, and qualifications accounted for over 60% childhood (age 11 years) intelligence was asso-
(Saunders 2002). Others have shown that so- ciated with 21% (95% confidence interval 25%
cial mobility from all the social classes is driven to 16%) lower survival up to age 76 (Whalley &
about equally by intelligence from childhood Deary 2001). This was a new and healthy devel-
(Nettle 2003). General findings—using struc- opment for intelligence research: its being in-
tural equation modeling—from analyses of var- cluded in large-scale epidemiological health re-
ious British cohorts (born in 1921, 1936, the search with population-representative samples.
1950s, and 1970) are that education tends to A few years later the term “cognitive epidemi-
mediate the influence of childhood intelligence ology” was coined (Deary & Der 2005) to de-
on adult socioeconomic status; the effect of in- scribe this field of research. A glossary of the
telligence on education is stronger (insofar as field is available (Deary & Batty 2007), and a
they can be compared numerically) than that of special issue of the journal Intelligence (2009)
parental social class; and that childhood behav- appeared, with 13 new empirical studies on the
ioral disturbance is correlated significantly with topic.
intelligence and contributes, at most, only small The association between lower intelligence
amounts of additional (beyond intelligence and test scores in childhood or youth and dying
parental social class) variance to education or earlier has been replicated in many studies,
adult social class (Deary et al. 2005a; Johnson typically involving the follow-up of thousands
et al. 2010; von Stumm et al. 2009, 2010). Most of subjects for up to several decades. The
of these latter analyses were completed in men largest single study to date included about
because of the difficulty in assessing socioeco- one million men (Batty et al. 2009). This was
nomic status among women at historical times conducted in Sweden and was possible because
when women either tended not to be in em- almost all Swedish men are conscripted into
ployment to the extent that they are now or military or civil service in young adulthood.
tended to be less likely to attain employment in For decades, young Swedish males have taken
accordance with their abilities. the same set of mental tests. Thereafter, the
unique code number that is assigned to Swedish
citizens was used prospectively to link these
Health, Illness, and Death cognitive test data with data held in education,
Prior to the past decade, health was not health, and other public databases. A system-
an outcome that counted intelligence as one atic review and meta-analysis has identified 16

www.annualreviews.org • Intelligence 469

 PS63CH18-Deary ARI 31 October 2011 12:36

independent studies of the intelligence- hypothesis has some merit but does not fully
mortality association and found that one account for the effects (Batty et al. 2006, Singh-
standard deviation advantage in intelligence Manoux et al. 2005). A promising route toward
was associated with 24% lower risk of death understanding has been the finding that child-
(95% CI = 23% to 25%) over a follow-up hood intelligence is associated with many health
range of 17 to 69 years (Calvin et al. 2011). behaviors during the life course: smoking, alco-
This paper and an extensive narrative review hol intake, physical activity, and dietary choices
with discussion (Deary et al. 2010b) debate (Batty et al. 2007, Deary et al. 2010b, Weiser
possible causes of the association and the range et al. 2009); risk factors for cardiovascular dis-
of causes of death with which intelligence is ease, including the metabolic syndrome as well
significantly associated. The latter include as its elements such as poor glucose regulation,
deaths from cardiovascular disease, suicide, higher blood pressure, high waist-hip ratio, and
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

homicide, and accidents, but not cancer. disadvantageous lipid profile (Batty et al. 2008,
Prospectively, physical and mental health Power et al. 2010); and a disadvantageous diur-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

outcomes are both associated with childhood or nal cortisol profile in middle age (Power et al.
early adult intelligence, especially cardiovascu- 2008).
lar disease among the physical illnesses (Deary There are other suggestions to explain the
et al. 2010b). This applies to most categories of intelligence-mortality association. For exam-
psychiatric disorders, and a standard deviation ple, it has been suggested that education and
disadvantage in intelligence at about age 20 has adult social class mediate and explain the as-
been associated with, for example, about 50% sociation, perhaps in association with stress
greater risk of hospitalization for schizophre- (Sapolsky 2005). Substantial mediation has
nia, mood disorder, and alcohol-related disor- been found statistically, but it is moot whether
ders (Gale et al. 2010) and for personality dis- these are explanatory factors or partial surro-
orders (Moran et al. 2009). Lower childhood gates for intelligence (Calvin et al. 2011). The
intelligence is also associated in early to middle idea that intelligence—even in childhood—
adulthood with more self-reported psychologi- relates to later health because it is an index
cal distress (Gale et al. 2009) and with a greater of general bodily “system integrity” achieved
risk of vascular dementia (McGurn et al. 2008). some validation when it was found that reac-
Although demonstrating, replicating, and tion time variance largely accounted for the as-
partially refining the association between in- sociation (Deary & Der 2005). The idea needs
telligence and illness and mortality has been a further development: theoretically, to elaborate
substantial achievement in the past ten years, more fully the notion of system integrity; and
the major interest now lies in understanding the empirically, to identify more marker variables
association. Finding its mechanisms will be im- of the construct and the testing of their associ-
portant for applying these discoveries in public ation with each other, intelligence, and mortal-
health. The attempt to understand has seen dif- ity, illness, and health.
ferential psychologists work more closely with
epidemiologists than before and has also seen
the two professions challenge each other. One INTELLIGENCE AND AGING
challenge came from what is arguably cogni- This is a topic that is growing in importance
tive epidemiology’s most comprehensive the- in intelligence research, and it also has signifi-
oretical statement, in which it was suggested cant public policy relevance (Beddington et al.
that the associations between health and educa- 2008). People are living longer, the propor-
tion and socioeconomic status might be caused tion of older people in the population is grow-
by intelligence (Gottfredson 2004). Direct tests ing, and losing cognitive ability is an espe-
of this idea with large epidemiological sam- cially feared aspect of growing older. Research
ples from the United Kingdom suggest that the includes the following questions: How do

470 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

aspects of intelligence change with age? Do all common factor representing people’s aging
domains of cognitive ability decline together? slopes on seven different cognitive domains ac-
How much stability and variation is there in in- counted for 62% of the variance. A study that is
telligence across the life course? Is age kinder to congruent with these, and extended their find-
the initially more able? Are there discoverable ings, was by Tucker-Drob (2011). He reported
(and ideally modifiable) determinants of indi- analyses on over 1,200 people aged from 18 to
vidual differences in the aging of intelligence? 95 years and who had been tested over a period
And can the decline in age-sensitive aspects in- of up to seven years on a dozen mental tests.
telligence be ameliorated? My opinion is that Again, the analyses employed the three-level hi-
cognitive aging should be an integrated part of erarchical model of cognitive differences. The
broader intelligence research, not a topic on its domains of function tested were abstract rea-
own: Intelligence as a whole should be seen as a soning, spatial visualization, episodic mem-
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

life-course topic, with developmental psychol- ory, and processing speed. A general factor—
ogy, individual differences, and aging aspects common to all the domains—accounted for an
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

contributing to a rounded account of how and average of 39% of the differences in individual
why intelligence differences develop, are main- variables, 33% was accounted for at the domain
tained, and decline (e.g., Foresight Mental Cap- level, and a mean of 28% was test specific. It is
ital Wellbeing Project 2008). Here, some ad- notable that the general effect on ‘aging’ was
vances over the past decade are highlighted. found even in younger and middle-age groups
Aging research on intelligence has been as well as in older people.
modeled using the hierarchical account of in- The past decade has seen the longest follow-
telligence differences (Carroll 1993). This is the up studies of intelligence differences, with
hierarchy that was described at the start of this follow-up studies of the Scottish Mental Sur-
article, whereby cognitive ability variance may veys of 1932 and 1947. When the same intelli-
be portioned into variance shared by all tests gence test is administered at age 11 years and
( g), variance shared by tests assessing the same again to individuals when they are in their late
cognitive domain, and variance specific to each seventies, the correlations are between 0.6 and
test. A comprehensive example is a combined 0.7 (Deary et al. 2000, 2004b) and are still above
analysis of 33 cross-sectional studies, involving 0.5 when the individuals are in their late eighties
16 cognitive tasks, with a total N of about 7,000 (Gow et al. 2011). Obviously, these correlations
(Salthouse 2004). The covariance of the 16 tasks imply that at least one-quarter to one-half of
formed a hierarchy with g at the pinnacle and the variance in intelligence is stable across most
group factors of reasoning, spatial ability, mem- of the human life course. The obverse is that
ory, processing speed, and vocabulary. Apart there is also considerable change in the rank
from vocabulary tests—which tended to peak order of intelligence across the life course, and
at about age 60—the other types of test showed there is a lively and varied set of research direc-
mean declines from young or middle adult- tions seeking the determinants—psychosocial
hood. It is notable, too, that Salthouse modeled and biomedical—of aging-associated cognitive
processing speed as one of the domains of in- change. Overviews of this research are pro-
telligence that ages, alongside others at the sec- vided by Deary et al. (2009c), Foresight Men-
ond level, instead of its being the cause of other tal Capital and Wellbeing Project (2008), and
domains’ aging effects, as he has done previ- Plassman et al. (2010). Plassman and colleagues
ously. The largest effect of age was on general systematically reviewed observational studies
intelligence, with additional, specific smaller ef- and randomized controlled trials and covered
fects on memory and processing speed. In a genetics, environmental toxins, medical factors,
longitudinal study, Wilson et al. (2002) also social and behavioral factors, and nutrition.
found that when one aspect of intelligence de- Support for most factors was limited, although
clines, the other aspects tend to go also. A it was better for the risk factors of smoking, the

www.annualreviews.org • Intelligence 471

 PS63CH18-Deary ARI 31 October 2011 12:36

APOE e4 allele, and some medical conditions. twins and that any small deficit probably dis-
Among the determinants of age-related cog- appears by adolescence (Webbink et al. 2008).
nitive change, there is still controversy about These large, new analyses of recently born chil-
whether people with relatively high childhood dren probably override the recent opinion that
intelligence have more gentle cognitive decline although there was heterogeneity in studies
in middle and old age, with some studies sug- of singleton–twin intelligence differences, this
gesting they do (Richards et al. 2004) and some might not be explained by a date-of-study ef-
suggesting that there is no such association fect (Voracek & Haubner 2008). This disap-
(Gow et al. 2011). pearance of the “cognitive cost” of being a twin
With regard to ameliorating the aging of is fascinating. Some have put this down to bet-
intelligence among people without cognitive ter perinatal care, but this has not been demon-
pathology, there is still discussion about how strated clearly.
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

to separate normal and pathological cogni- Sex differences in intelligence remains a hot
tive aging, though with a suggestion that the topic, and the past decade saw much debate as
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

former might be a specific target for thera- well as substantial new analyses. Four issues ap-
peutics (Shineman et al. 2010). There is still pear to be important in assessing this question.
uncertainty, too, about whether retaining en- First is the quality of the samples: Some have
gagement in physical, social, and intellectual much better population representativeness than
activity helps to preserve what intelligence we others. Second is the age of the samples: Some
have (Bielak, 2010), although some are positive suggest that mean differences appear only after
about this (Hertzog et al. 2009). puberty. Third is the type of mental ability: The
issue addressed here will be general intelligence
rather than abilities such as spatial or verbal
CONTROVERSY OR CONSENSUS? that, for whatever causes, tend to be accepted as
Two long-running controversies in intelligence having more replicable sex differences. Fourth
research attracted much empirical attention in is the research issue at hand: There are ques-
the past decade: whether—and if so, why— tions about whether the mean and/or the vari-
there are twin-singleton or male-female differ- ance of intelligence is different between males
ences in intelligence. Both are discussed below. and females. These will all be kept in mind
There was also a major re-examination of the is- and referred to as some recent research re-
sue of race differences in intelligence, and read- sults on sex differences in intelligence are
ers are referred to the target article and its sub- discussed.
sequent discussion (Rushton & Jensen 2005). In whole populations of Scottish 11-year-
Two analyses of whole-population or olds from the 1930s and 1940s there was more
population-representative sets of twins found variance among males in intelligence by com-
that, for children aged 11 years in Scotland parison with females, but little difference in
in the 1930s, 1940s, and 1960s, twins had mean intelligence; proportionately, more males
mean IQ scores of about one-third of a stan- were at both ends of the intelligence distribu-
dard deviation lower than singletons (Deary tion and fewer were in the middle (Deary et al.
et al. 2005b, Ronalds et al. 2005). However, 2003, Johnson et al. 2008b). This pattern was
a population-representative study of 11-year- replicated in a population-representative sam-
olds tested on the Cognitive Abilities Test in ple of almost one-third of a million boys and
the United Kingdom in 2004 showed no sig- girls tested on the Cognitive Abilities Test in
nificant difference: Twins were only about 1% the United Kingdom in the early 2000s: There
of a standard deviation lower than singletons on were trivial differences in mean general intel-
general intelligence (Calvin et al. 2009). Studies ligence and proportionately more boys than
of more recently born Dutch twins also suggest girls at the extremes (Strand et al. 2006). This
a much lower—or no—cognitive deficit among pattern was replicated in the same test in a

472 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

population-representative sample of 300,000+ biased toward higher male means when only
students in grades 3 to 11 in the United States, those subjects who took part in subsequent
with the authors concluding that, “The re- waves were tested. This was extended further—
sults showed an astonishing consistency in sex to include a mathematical model of the re-
differences across countries, grades, cohorts, cruitment process and to analyses of both the
and test forms” (Lohman & Lakin 2009). 1958 and 1970 British birth cohorts—and it
Whereas these epidemiological-quality data was again found that sample restriction by these
in children make the state of sex differences means can bring about apparently higher mean
clear, the picture is less clear among adults. intelligence in males even when none is present
On the basis of meta-analyses of studies using in the original samples (Madhyastha et al. 2009).
Raven’s Progressive Matrices—a widely used
nonverbal intelligence test with items com-
CONCLUDING REMARKS
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

posed using abstract line drawings and demand-

ing inductive reasoning—it was argued that A desultory reading of this necessarily selec-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

men have a higher mean level of general intel- tive overview of intelligence research in the first
ligence than do women (Irwing & Lynn 2005, decade of the twentieth century—about one
Lynn & Irwing 2004). It was argued against this hundred years after the first scientific research
that the better-quality studies tended to show on intelligence—will lead some readers to echo
no differences and that many of the samples Jean-Baptiste Alphonse Karr: “Plus ça change,
might be biased toward less selectivity among plus c’est la même chose.” They would be wrong,
women, which would reduce their mean scores for two reasons.
as a group (Blinkhorn 2005). To test this idea, First, yes, some issues do seem to have con-
a novel design was applied to data from the tinued for a very long time: the psychometric
U.S. National Longitudinal Survey of Youth structure of intelligence, sex differences in in-
1979 (Deary et al. 2007b). When brother-sister telligence, environment and genetic contribu-
pairs (N = 1,292 pairs) were compared on tions to intelligence, the aging of intelligence,
a general cognitive ability ( g) factor derived and whether intelligence drives educational at-
from the Armed Services Vocational Aptitude tainment and social mobility. In all of these,
Battery and the shorter Armed Forces Qual- the basics of what we know now were avail-
ification Test, the male mean advantage was able empirically decades ago. However, there
trivial (less than 7% of a standard deviation), was often so much opposition to the findings—
but males once again had substantially greater creating uncertainty—that more, newer, and
variance. better research has now far more firmly founded
A suggestion was made that an apparent the conclusions that can be offered in these ar-
male advantage in tests of general intelligence eas. In addition, research in most of these ar-
could in part be due to the combination of the eas has added important details to the basic
following factors: Males have greater variance findings that were available early on. In part,
in intelligence, and recruitment into studies and these have come from better samples, better
into subsequent follow-ups of existing studies research designs, and more sophisticated and
is restricted by intelligence and sex, with more appropriate statistical modeling. On that last
women and people of higher intelligence tend- point, it remains the case that fully to engage
ing to take part (Dykiert et al. 2009). The au- in the discussions about intelligence research it
thors argued that the combination of these fac- is necessary to be numerate, and to quite an ad-
tors could produce an apparently higher male vanced level in specialist areas such as aging and
intelligence mean even when none existed in the environment-genetics.
population. This was tested in the British Co- Second, there are some genuinely new de-
hort Study 1970 and found to be true: Sex dif- velopments. The importance of intelligence for
ferences in intelligence at age 10 became more health, illness, and mortality is new. It adds a

www.annualreviews.org • Intelligence 473

 PS63CH18-Deary ARI 31 October 2011 12:36

great deal to the predictive validity of intelli- correlations and fresh evidence of apparent
gence. The brain associations of intelligence are causal associations with intelligence. Neverthe-
new: They might not be large in effect size, but less, this area is not as active as it was. A moun-
they offer foundational findings for a develop- taineering analogy might help. For those seek-
ing biological account of what it means to have ing a causal account of intelligence differences,
higher levels of intelligence. The molecular genetic and brain imaging approaches are far
genetics knowledge about intelligence—albeit more reachable handholds and footholds than
largely knowing what the molecular genetics of they were a decade ago. Perhaps researchers
intelligence is not like—is new. Because longi- are wondering whether they can simply bypass
tudinal cohorts—first studied in childhood— the intermediate cognitive construct footholds,
are still being studied at older ages, we now some of which might be illusory (based on un-
have a far better knowledge about how intel- validated constructs) and some of which might
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

ligence plays out into social as well as health take the researcher laterally rather than verti-
outcomes in middle and older ages. In health, cally (because they end up redescribing intelli-
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

genetics, biology, and social science we are gence in terms at the same explanatory level).
seeing and will see the incorporation of mea- If reorienting is the correct word—or
sured intelligence as a phenotype in better sam- perhaps it is synthesis or consolidation—then a
ples, of more epidemiological quality: larger, reorienting of intelligence research is possible
more population-representative, followed up because of the research in the past decade or
for longer, and better characterized in terms two, in terms of a life-course model. The fairly
of other phenotypes and genetic information. newly appreciated very long-term stability of
The past decade has seen the first studies of the intelligence, the influences of birth weight and
stability and change in intelligence that have perinatal growth, and the associations among
lasted more than 70 years, health-related stud- intelligence, sensation, and health all point to-
ies of over one million men, brain imaging stud- ward the need for intelligence researchers to be
ies with numbers in three figures (instead of integrated with a wide range of cognate scien-
two or one), and the first genetic studies using tists interested in general health and well-being
hundreds of thousands of genetic markers. The across the human life course. Intelligence is
forthcoming consortia on the genetics of intel- part of health. Whether it is called intelligence,
ligence will have sample sizes in at least five or cognitive capital, or cognitive reserve, and
figures, and large studies that will include in- whether we explain these links with ideas of
telligence and brain imaging and genetics will general bodily system integrity or common
increasingly appear. In social matters we shall cause (or a set of common causes), the develop-
know more too: As cohorts such as the United ment, adult operation, and eventual trajectory
Kingdom’s 1946, 1958, and 1970 cohorts grow of decline in people’s intelligence will be a
older, we shall know more about the lifelong combination of shared influences with the rest
social mobility of these individuals (and their of the body and—who knows how much?—
children) and the part played by intelligence in influences that are specific to the brain. An
concert with other factors. attempt at displaying the integrated science
Real progress in the cognitive correlates of that lifelong intelligence research should be is
intelligence has been made over the past decade: shown in a U.K. government report (Foresight
It is clear that there is a sensory discrimination– Mental Capital and Wellbeing Project 2008,
intelligence correlation; the inspection time– appendix B, p. 53).
intelligence correlation has been firmed up in And before one hears the old saws that there
larger samples and in adults and children; a is more to life than being clever—sure there is,
much larger reaction time–intelligence corre- like being happy, healthy, and free—and that
lation has emerged than had been appreci- there is more to achieving one’s desired posi-
ated heretofore; and working memory has large tion than being clever—that’s trivially true, as

474 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

studies looking at personality traits and effort thing that people value and that has a big influ-
and motivation, for example, show—we should ence on people’s lives. This brings with it the
remember that research on the nature, causes, responsibility to be broad-minded and intelli-
and consequences of intelligence is about some- gent in researching intelligence.

SUMMARY POINTS
1. Intelligence differences continue to be a focus for lively research in psychology and also
of considerable interest to nonspecialist psychologists, academics in other fields, and the
public.
2. The past decade produced many books on intelligence, from introductory accounts to
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

specialist discussions of specific issues. There are also historical accounts and books
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

challenging the measurement and study of intelligence differences.

3. There is new research on the psychometric structure of intelligence. The g factor
from different test batteries ranks people in the same way. There is still debate about
the number of levels at which the variations in intelligence is best described. There
is still little empirical support for an account of intelligence differences that does not
include g.
4. There has been progress in establishing that sensory discrimination, inspection time, and
reaction time are all associated with intelligence and achieving estimates of the population
effect sizes. However, they now attract less attention as possible ways to understand
intelligence differences, although sensory discrimination does attract attention as part of
the common cause account of aging and intelligence.
5. The biology of intelligence is the subject of much research. Behavior genetics research
continues to refine what we know about environmental and genetic contributions to
intelligence, such as moderating effects of age and social circumstances, and the shared
genetic influences of intelligence with, for example, brain size, processing speed, and
birth weight. Molecular genetic research on intelligence has had a dry time with candidate
gene studies and is now poised to take on sufficiently powered genomewide association
studies. Brain imaging studies of intelligence are providing more replicated findings that
are cohering around an account of a defined but distributed network in the brain that
works more efficiently in people with higher intelligence scores.
6. New work on education and social mobility and social position as the outcomes of
intelligence differences has plotted people’s life courses from impressive longitudinal
studies. Health outcomes are a new and burgeoning outcome for intelligence differ-
ences, and it is only in the past decade that the new field of cognitive epidemiology has
emerged.
7. Aging is another expanding focus for intelligence research, with new findings. Also, this
field increasingly takes a life-course view and is becoming more integrated with the study
of intelligence differences in younger adulthood and in child development.
8. Controversial issues continue to be studied in intelligence. One such issue is the changing
twin–singleton intelligence difference. Also, sex differences in intelligence continue to
attract new research, with studies of both mean and variance differences.

www.annualreviews.org • Intelligence 475

 PS63CH18-Deary ARI 31 October 2011 12:36

FUTURE ISSUES
1. Molecular genetics studies of intelligence are required that are sufficiently powered to
discover the likely very small effects of single-nucleotide polymorphisms in genomewide
studies. Thereafter, there will almost certainly be a continued search for other loci of
intelligence differences that will require gene sequencing studies, the analysis of copy
number variations, gene methylation studies, and transcriptomics, proteomics, and other
related studies.
2. Structural and functional brain imaging studies of intelligence are increasing in power
and will continue to do so, and will also become more genetically informed.
3. The incorporation of intelligence tests into more cohort studies will enhance the study
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

of social mobility, health, and other life outcomes across the human life course.
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

4. Intelligence research is increasingly becoming, and should continue to become, an inte-

grated part of understanding the causes and consequences of human development and
human lifelong health and well-being.

DISCLOSURE STATEMENT
The author is not aware of any affiliations, memberships, funding, or financial holdings that might
be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS
The work was undertaken within the University of Edinburgh Center for Cognitive Ageing
and Cognitive Epidemiology, part of the cross-council Lifelong Health and Wellbeing Initiative
(G0700704/84698). Funding from the Biotechnology and Biological Sciences Research Council
(BBSRC), Engineering and Physical Sciences Research Council (EPSRC), Economic and Social
Research Council (ESRC), and Medical Research Council (MRC) is gratefully acknowledged.

LITERATURE CITED
Acton GS, Schroeder DH. 2001. Sensory discrimination as related to general intelligence. Intelligence 29:263–
71
Banks GC, Batchelor JH, McDaniel MA. 2010. Smarter people are (a bit) more symmetrical: a meta-analysis
of the relationship between intelligence and fluctuating asymmetry. Intelligence 38:393–401
Barnett JH, Scoriels L, Munafo MR. 2008. Meta-analysis of the cognitive effects of the catechol-O-transferase
gene Val158Met polymorphism. Biol. Psychiatry 64:137–44
Bartels M, Rietveld MJH, Van Baal GCM, Boomsma DI. 2002. Genetic and environmental influences on the
development of intelligence. Behav. Genet. 32:237–49
Bartholomew DJ. 2004. Measuring Intelligence: Facts and Fallacies. Cambridge, UK: Cambridge Univ. Press
Bartholomew DJ, Deary IJ, Lawn M. 2009. A new lease of life for Thomson’s bonds model of intelligence.
Psychol. Rev. 116:567–79
Batty DJ, Deary IJ, Schoon I, Gale CR. 2007. Mental ability across childhood in relation to risk factors for
premature mortality in adult life: the 1970 British Cohort Study. J. Epidemiol. Comm. Health 61:997–1003
Batty GD, Der G, Macintyre S, Deary IJ. 2006. Does IQ explain socioeconomic inequalities in health? Evidence
from a population based cohort study in the west of Scotland. Br. Med. J. 332:580–84
Batty GD, Gale CR, Mortensen LH, Langenberg C, Shipley MJ, et al. 2008. Pre-morbid intelligence, the
metabolic syndrome and mortality: the Vietnam Experience Study. Diabetologia 51:436–43

476 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

Batty GD, Wennerstad KM, Davey Smith G, Gunnell D, Deary IJ, et al. 2009. IQ in late adolescence/early
adulthood and mortality by middle age: cohort study of one million Swedish men. Epidemiology 20:100–9
Beddington J, Cooper CL, Field J, Goswami U, Huppert FA, et al. 2008. The mental wealth of nations. Nature
455:1057–60
Bielak AAM. 2010. How can we not “lose it” if we still don’t understand how to “use it”? Gerontology 56:507–19
Binet A. 1905. New methods for the diagnosis of the intellectual level of subnormals. L’Annee Psychol. 12:191–
244. Transl. ES Kite, 1916, in The Development of Intelligence in Children. Vineland, NJ: Publ. Training
School Vineland
Blinkhorn S. 2005. Intelligence: a gender bender. Nature 438:31–32
Brand CR. 1996. The g Factor. Chichester, UK: Wiley. (This book was withdrawn by the publisher shortly
after publication.)
Breen R, Goldthorpe JH. 2002. Merit, mobility and method: another reply to Saunders. Br. J. Sociol. 53:575–82
Burks SV, Carpenter JP, Goette L, Rustichini A. 2009. Cognitive skills affect economic preferences, strategic
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

behavior, and job attachment. Proc. Natl. Acad. Sci. USA 106:7745–50
Calvin C, Deary IJ, Fenton C, Roberts BA, Der G, et al. 2011. Intelligence in youth and all-cause mortality:
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

systematic review with meta-analysis. Int. J. Epidemiol. In press

Calvin C, Fernandes C, Smith P, Visscher PM, Deary IJ. 2009. Is there still a cognitive cost of being a twin
in the UK? Intelligence 37:243–48
Carroll JB. 1993. Human Cognitive Abilities: A Survey of Factor Analytic Studies. Cambridge, UK: Cambridge
Univ. Press
Carson J. 2007. The Measure of Merit: Talents, Intelligence, and Inequality in the French and American Republics,
1750–1940. Princeton, NJ: Princeton Univ. Press
Caspi A, Williams B, Kim-Cohen J, Craig IW, Milne BJ, et al. 2007. Moderation of breastfeeding effects on
the IQ by genetic variation in fatty acid metabolism. Proc. Natl. Acad. Sci. USA 104:18860–65
Chiang M-C, Barysheva M, Shattuck DW, Lee AD, Madsen SK. 2009. Genetics of brain fiber architecture
and intellectual performance. J. Neurosci. 29:2212–24
Choi YY, Shamosh NA, DeYoung CG, Lee MJ, Lee J-M, et al. 2008. Multiple bases of human intelligence
revealed by cortical thickness and neural activation. J. Neurosci. 28:10323–29
Cianciolo AT, Sternberg RJ. 2004. Intelligence: A Brief History. Malden, MA: Blackwell
Cirulli ET, Kasperaviciute D, Attix DK, Need AC, Ge D, et al. 2010. Common genetic variation and perfor-
mance on standardized cognitive tests. Eur. J. Hum. Genet. 18:815–20
Cohn LD, Westenberg PM. 2004. Intelligence and maturity: meta-analytic evidence for the incremental and
discriminant validity of Loevinger’s measure of ego development. J. Personal. Soc. Psychol. 86:760–72
Colom R, Jung RE, Haier RJ. 2006. Distributed brain sites for the g-factor of intelligence. NeuroImage 31:1359–
65
Cronbach LJ. 1957. The two disciplines of scientific psychology. Am. Psychol. 12:671–84
Davies G, Tenesa A, Payton A, Yang J, Harris SE, et al. 2011. Genome-wide association studies establish that
human intelligence is highly heritable and polygenic. Mol. Psychiatry. doi:10.1038/mp.2011.85. In press
Davis OS, Butcher LM, Docherty SJ, Meaburn EL, Curtis CJ, et al. 2010. A three-stage genome-wide asso-
ciation study of general cognitive ability: hunting the small effects. Behav. Genet. 40:759–67
Davis OS, Haworth CM, Plomin R. 2009. Dramatic increase in heritability of cognitive development from
early to middle childhood: an 8-year longitudinal study of 8700 pairs of twins. Psychol. Sci. 20:1301–8
Deary IJ. 2000. Looking Down on Human Intelligence: From Psychometrics to the Brain. Oxford, UK: Oxford Univ.
Press
Deary IJ. 2001. Intelligence: A Very Short Introduction. Oxford, UK: Oxford Univ. Press
Deary IJ, Batty GD. 2007. Cognitive epidemiology: a glossary. J. Epidemiol. Comm. Health 61:378–84
Deary IJ, Batty GD, Gale CR. 2008. Bright children become enlightened adults. Psychol. Sci. 19:1–6
Deary IJ, Bell PJ, Bell AJ, Campbell ML, Fazal ND. 2004a. Sensory discrimination and intelligence: testing
Spearman’s other hypothesis. Am. J. Psychol. 117:1–19
Deary IJ, Corley J, Gow AJ, Harris SE, Houlihan LM, et al. 2009a. Age-associated cognitive decline. Br. Med.
Bull. 92:135–52
Deary IJ, Der G. 2005. Reaction time explains IQ’s association with death. Psychol. Sci. 16:64–69

www.annualreviews.org • Intelligence 477

 PS63CH18-Deary ARI 31 October 2011 12:36

Deary IJ, Der G, Ford G. 2001. Reaction times and intelligence differences: a population-based cohort study.
Intelligence 29:389–99
Deary IJ, Irwing P, Der G, Bates TC. 2007b. Brother-sister differences in the g factor in intelligence: analysis
of full, opposite-sex siblings from the NLSY1979. Intelligence 35:451–56
Deary IJ, Johnson W, Houlihan LM. 2009b. Genetic foundations of human intelligence. Hum. Genet. 126:215–
32
Deary IJ, Pattie A, Wilson V, Whalley LJ. 2005b. The cognitive cost of being a twin: two whole-population
surveys. Twin Res. Hum. Genet. 8:376–83
Deary IJ, Penke L, Johnson W. 2010a. The neuroscience of human intelligence differences. Nat. Rev. Neurosci.
11:201–11
Deary IJ, Spinath FM, Bates TC. 2006. Genetics of intelligence. Eur. J. Hum. Genet. 14:690–700
Deary IJ, Strand S, Smith P, Fernandes C. 2007a. Intelligence and educational achievement. Intelligence 35:13–
21
Deary IJ, Taylor MD, Hart CL, Wilson V, Davey Smith G, et al. 2005a. Intergenerational social mobility and
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

mid-life status attainment: influences of childhood intelligence, childhood social factors, and education.
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Intelligence 33:455–72
Deary IJ, Thorpe G, Wilson V, Starr JM, Whalley LJ. 2003. Population sex differences in IQ at age 11: the
Scottish Mental Survey 1932. Intelligence 31:533–42
Deary IJ, Weiss A, Batty GD. 2010b. Intelligence and personality as predictors of illness and death: how
researchers in differential psychology and chronic disease epidemiology are collaborating to understand
and address health inequalities. Psychol. Sci. Publ. Interest 11:53–79
Deary IJ, Whalley LJ, Lemmon H, Crawford JR, Starr JM. 2000. The stability of individual differences in
mental ability from childhood to old age: follow-up of the 1932 Scottish Mental Survey. Intelligence
28:49–55
Deary IJ, Whalley LJ, Starr JM. 2009c. A Lifetime of Intelligence: Follow-up Studies of the Scottish Mental Surveys
of 1932 and 1947. Washington, DC: Am. Psychol. Assoc.
Deary IJ, Whiteman MC, Starr JM, Whalley LJ, Fox HC. 2004b. The impact of childhood intelligence in
later life: following up the Scottish Mental Surveys of 1932 and 1947. J. Personal. Soc. Psychol. 86:130–47
Der G, Batty GD, Deary IJ. 2006. Effect of breast feeding on intelligence in children: prospective study,
sibling pairs analysis, and meta-analysis. Br. Med. J. 333:945–48
Duckworth AL, Quinn PD, Lynam DR, Loeber R, Stouthamer-Loeber M. 2011. Role of test motivation in
intelligence testing. Proc. Natl. Acad. Sci. USA 108:7716–20
Duckworth AL, Seligman MEP. 2005. Self-discipline outdoes IQ in predicting academic performance of
adolescents. Psychol. Sci. 16:939–44
Duncan J. 2010. How Intelligence Happens. New Haven, CT: Yale Univ. Press
Duncan J, Seitz RJ, Kolodny J, Bor D, Herzog H, et al. 2000. A neural basis for general intelligence. Science
289:457–60
Dykiert D, Gale CR, Deary IJ. 2009. Are apparent sex differences in mean IQ scores created in part by sample
restriction and increased male variance? Intelligence 37:42–47
Flynn J. 2007. What Is Intelligence? Cambridge, UK: Cambridge Univ. Press
Foresight Mental Capital and Wellbeing Project. 2008. Final Project Report—Executive Summary. London:
Gov. Off. Sci.
Friedman NP, Miyake A, Young SE, DeFries JC, Corley RP, et al. 2008. Individual differences in executive
functions are almost entirely genetic in origin. J. Exp. Psychol.: Gen. 137:201–25
Frisoni GB, Galluzzi S, Pantoni L, Filippi M. 2007. The effect of white matter lesions on cognition in the
elderly: small but detectable. Nat. Clin. Pract. Neurol. 3:620–27
Furlow B, Armijo-Prewitt T, Gangestad SW, Thornhill R. 1997. Fluctuating asymmetry and psychometric
intelligence. Proc. R. Soc. Lond. B 264:823–29
Gale CR, Batty GD, Tynelius P, Deary IJ, Rasmussen F. 2010. Intelligence in early adulthood and subsequent
hospitalization for mental disorders. Epidemiology 21:70–77
Gale CR, Hatch SL, Batty GD, Deary IJ. 2009. Intelligence in childhood and risk of psychological distress
in adulthood: the National Child Development Survey and the 1970 British Cohort Study. Intelligence
37:592–99

478 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

Gardner H. 1983. Frames of Mind. New York: Basic Books

Gardner H. 1993. Multiple Intelligences. New York: Basic Books
Garlick D. 2010. Intelligence and the Brain: Solving the Mystery of Why People Differ in IQ and How a Child Can
Be a Genius. Burbank, CA: Aesop
Geary DC. 2005. The Origin of Mind: Evolution of Brain, Cognition, and General Intelligence. Washington, DC:
Am. Psychol. Assoc.
Gläscher J, Rudrauf D, Colom R, Paul LK, Tranel D, et al. 2010. Distributed neural system for general
intelligence revealed by lesion mapping. Proc. Natl. Acad. Sci. USA 107:4705–9
Gottfredson LS. 2004. Intelligence: is it the epidemiologists’ elusive “fundamental cause” of social class in-
equalities in health? J. Personal. Soc. Psychol. 86:174–99
Gow AJ, Johnson W, Pattie A, Brett CE, Roberts B, et al. 2011. Stability and change in intelligence from age
11 to ages 70, 79, and 87: the Lothian Birth Cohorts of 1921 and 1936. Psychol. Aging 26:232–40
Gray JR, Chabris CF, Braver TS. 2003. Neural mechanisms of general fluid intelligence. Nat. Neurosci. 6:316–
22
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

Haier RJ. 2009. Neuro-intelligence, neuro-metrics and the next phase of brain imaging studies. Intelligence
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

37:121–23
Herrnstein RJ, Murray C. 1994. The Bell Curve: Intelligence and Class Structure in American Life. New York:
Free Press
Hertzog C, Kramer AF, Wilson RS, Lindenberger U. 2009. Enrichment effects on adult cognitive develop-
ment. Psychol Sci. Publ. Int. 9:1–65
Hulshoff Pol H, Schnack HG, Posthuma D, Mandl RCW, Barré WF, et al. 2006. Genetic contributions to
human brain morphology and intelligence. J. Neurosci. 26:10235–42
Hunt E. 2011. Human Intelligence. Cambridge, UK: Cambridge Univ. Press
Intelligence. 2009. Special issue. Intelligence, health and death: the emerging field of cognitive epidemiology.
Intelligence 37:6(whole issue)
Irwing P, Lynn R. 2005. Sex differences in means and variability on the progressive matrices in university
students: a meta-analysis. Br. J. Psychol. 96:505–24
Jensen AR. 1998. The g Factor. New York: Praeger
Johnson W, Bouchard TJ. 2005a. The structure of human intelligence: It is verbal, perceptual, and image
rotation (VPR), not fluid and crystallized. Intelligence 33:393–416
Johnson W, Bouchard TJ. 2005b. Constructive replication of the visual [sic]-perceptual-image rotation model
in Thurstone’s (1941) battery of 60 tests of mental ability. Intelligence 33:417–30
Johnson W, Bouchard TJ, Krueger RF, McGue M, Gottesman II. 2004. Just one g: consistent results from
three test batteries. Intelligence 32:95–107
Johnson W, Bouchard TJ, McGue M, Segal NL, Tellegen A, et al. 2007. Genetic and environmental influ-
ences on the Verbal-Perceptual-Image Rotation (VPR) model of the structure of mental abilities in the
Minnesota Study of Twins Reared Apart. Intelligence 35:542–62
Johnson W, Carothers A, Deary IJ. 2008a. Sex differences in variability in general intelligence: a new look at
the old question. Perspect. Psychol. Sci. 3:518–31
Johnson W, Gow AJ, Corley J, Starr JM, Deary IJ. 2010. Location in cognitive and residential space at age 70
reflects a lifelong trait over parental and environmental circumstances: the Lothian Birth Cohort 1936.
Intelligence 38:402–11
Johnson W, Segal NL, Bouchard TJ. 2008b. Fluctuating asymmetry and general intelligence: no genetic or
phenotypic association. Intelligence 36:279–88
Johnson W, te Nijenhuis J, Bouchard TJ. 2008c. Still just 1 g: consistent results from five test batteries.
Intelligence 36:81–95
Jung RE, Gasparovic C, Chavez RS, Caprihian A, Barrow R, et al. 2009. Imaging intelligence with magnetic
resonance spectroscopy. Intelligence 37:192–98
Jung RE, Haier RJ. 2007. The Parieto-Frontal Integration Theory (P-FIT) of intelligence: converging neu-
roimaging evidence. Behav. Brain Sci. 30:135–54, discussion 154–87
Karama S, Ad-Dab’bagh Y, Haier RJ, Deary IJ, Lyttelton OC, et al. 2009. Positive association between cog-
nitive ability and cortical thickness in a representative US sample of healthy 6 to 18 year-olds. Intelligence
37:145–55

www.annualreviews.org • Intelligence 479

 PS63CH18-Deary ARI 31 October 2011 12:36

Keller MC, Miller G. 2007. Resolving the paradox of common, harmful, heritable mental disorders: Which
evolutionary genetic models work best? Behav. Brain Sci. 29:385–404
Lango Allen H, Estrada K, Lettre G, Berndt SI, Weedon MN, et al. 2010. Hundreds of variants clustered in
genomic loci and biological pathways affect human height. Nature 467:832–38
Lee T, Henry JD, Trollor JN, Sachdev PS. 2010. Genetic influences on cognitive functions in the elderly: a
selective review of twin studies. Brain Res. Rev. 64:1–13
Lohman DF. 1994. Component scores as residual variation (or why the intercept correlates best). Intelligence
19:1–11
Lohman DF, Lakin JM. 2009. Consistencies in sex differences on the Cognitive Abilities Test across countries,
grades, test forms, and cohorts. Br. J. Educ. Psychol. 79:389–407
Luciano M, Posthuma D, Wright M, de Geus EJC, Smith GA, et al. 2005. Perceptual speed does not cause
intelligence, and intelligence does not cause perceptual speed. Biol. Psychol. 70:1–8
Luciano M, Wright MJ, Duffy DL, Wainwright MA, Zhu G, et al. 2006. Genome-wide scan of IQ finds
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

significant linkage to a quantitative trait locus on 2q. Behav. Genet. 36:45–55

Luciano M, Wright MJ, Geffen GM, Geffen LB, Smith GA, et al. 2004a. A multivariate genetic analysis of
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

cognitive abilities in an adolescent twin sample. Aust. J. Psychol. 56:79–88

Luciano M, Wright M, Martin N. 2004b. Exploring the etiology of the association between birthweight and
IQ in an adolescent twin sample. Twin Res. 7:62–71
Lynn R, Irwing P. 2004. Sex differences on the progressive matrices: a meta-analysis. Intelligence 32:481–98
Lynn R, Vanhanen T. 2002. IQ and the Wealth of Nations. Westport, CT: Praeger
Lyons MJ, York TP, Franz CE, Grant MD, Eaves LJ, et al. 2009. Genes determine stability and the environ-
ment determines change in cognitive ability during 35 years of adulthood. Psychol. Sci. 20:1146–52
Mackintosh NJ. 1996. Science struck dumb. Nature 381:33
Madhyastha TM, Hunt E, Deary IJ, Gale CR, Dykiert D. 2009. Recruitment modeling applied to longutidinal
studies of group differences in intelligence. Intelligence 37:422–27
McDaniel MA. 2005. Big-brained people are smarter: a meta-analysis of the relationship between in vivo brain
volume and intelligence. Intelligence 33:337–46
McGurn B, Deary IJ, Starr JM. 2008. Childhood cognitive ability and risk of late-onset Alzheimer and vascular
dementia. Neurology 71:1051–56
Miller GF. 2000. The Mating Mind. New York: Doubleday
Moran P, Klinteberg BA, Batty GD, Vagero D. 2009. Childhood intelligence predicts hospitalization with
personality disorder in adulthood: evidence from a population-based study in Sweden. J. Personal. Disord.
23:535–40
Murdoch S. 2007. IQ: The Smart History of a Failed Idea. Hoboken, NJ: Wiley
Narr KL, Woods RP, Thompson PM, Szeszko P, Robinson D, et al. 2007. Relationships between IQ and
regional cortical gray matter thickness in healthy adults. Cereb. Cortex 17:2163–71
Nettle D. 2003. Intelligence and class mobility in the British population. Br. J. Psychol. 94:551–61
Neubauer AC, Fink A. 2009. Intelligence and neural efficiency. Neurosci. Biobehav. Rev. 33:1004–23
Nisbett R. 2009. Intelligence and How to Get It. New York: Norton
O’Toole BI, Stankov L. 1992. Ultimate validity of psychological tests. Personal. Individ. Differ. 13:699–716
Payton A. 2009. The impact of genetic research on our understanding of normal cognitive aging: 1995 to
2009. Neuropsychol. Rev. 19:451–77
Penke L, Denissen JJA, Miller GF. 2007. The evolutionary genetics of personality. Eur. J. Personal. 21:549–87
Plassman BL, Williams JW, Burke JR, Holsinger T, Benjamin S. 2010. Systematic review: factors associated
with risk for and possible prevention of cognitive decline in later life. Ann. Int. Med. 153:182–93
Plomin R, Spinath F. 2004. Intelligence: genetics, genes and genomics. J. Personal. Soc. Psychol. 86:112–29
Posthuma D, de Geus EJC, Barré WF, Hulshoff Pol HE, Kahn RS, et al. 2002. The association between brain
volume and intelligence is of genetic origin. Nat. Neurosci. 5:83–84
Posthuma D, de Geus EJC, Boomsma DI. 2001. Perceptual speed and IQ are associated through common
genetic factors. Behav. Genet. 31:593–602
Posthuma D, Luciano M, de Geus EJC, Wright MJ, Slagboom PE, et al. 2005. A genomewide scan for
intelligence identifies quantitative trait loci on 2q and 6p. Am. J. Hum. Genet. 77:318–26

480 Deary
 PS63CH18-Deary ARI 31 October 2011 12:36

Power C, Jefferis BJ, Manor O. 2010. Childhood cognition and risk factors for cardiovascular disease in
midadulthood: the 1958 British Birth Cohort Study. Am. J. Public Health 100:129–36
Power C, Li L, Hertzman C. 2008. Cognitive development and cortisol patterns in mid-life: findings from a
British birth cohort. Psychoneuroendocrinology 33:530–39
Reynolds CA, Finkel D, McArdle JJ, Gatz M, Berg S, et al. 2005. Quantitative genetic analysis of latent growth
curve models of cognitive abilities in adulthood. Dev. Psychol. 41:3–16
Richards M, Shipley B, Fuhrer R, Wadsworth MEJ. 2004. Cognitive ability in childhood and cognitive decline
in midlife: longitudinal birth cohort study. Br. Med. J. 328:552–56
Rijsdijk FV, Vernon PA, Boomsma DI. 2002. Application of hierarchical genetic models to Raven and WAIS
subtests: a Dutch twin study. Behav. Genet. 32:199–210
Ronalds GA, De Stavola BL, Leon DA. 2005. The cognitive cost of being a twin: evidence from comparisons
within families in the Aberdeen children of the 1950s cohort. Br. Med. J. 331:1306
Rushton JP, Ankney CD. 2009. Whole brain size and general mental ability: a review. Int. J. Neurosci. 119:691–
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

731
Rushton JP, Jensen AR. 2005. Thirty years of research on race differences in cognitive ability. Psychol. Publ.
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Policy Law 11:235–94

Salthouse TA. 2004. Localizing age-related individual differences in a hierarchical structure. Intelligence
32:541–61
Salthouse TA. 2010. Major Issues in Cognitive Aging. Oxford, UK: Oxford Univ. Press
Sapolsky RM. 2005. The influence of social hierarchy on primate health. Science 308:648–52
Saunders P. 2002. Reflections on the meritocracy debate in Britain: a response to Richard Breen and John
Goldthorpe. Br. J. Sociol. 53:559–74
Saunders P. 2010. Social Mobility Myths. London: Civitas
Schaie KW. 2005. Developmental Influences on Adult Intelligence: The Seattle Longitudinal Study. Oxford, UK:
Oxford Univ. Press
Schoon I, Cheng H, Gale CR, Batty GD, Deary IJ. 2010. Social status, cognitive ability, and educational
attainment as predictors of liberal social attitudes and political trust. Intelligence 38:144–50
Shaw P, Greenstein D, Lerch J, Clasen L, Lenroot R, et al. 2006. Intellectual ability and cortical development
in children and adolescents. Nature 440:676–79
Shenk D. 2010. The Genius in All of Us: Why Everything You’ve Been Told About Genetics, Talent and Intelligence
is Wrong. London: Icon
Shenkin SD, Starr JM, Deary IJ. 2004. Birth weight and cognitive ability in childhood: a systematic review.
Psychol. Bull. 130:989–1013
Sheppard LD. 2008. Intelligence and information processing: a review of 50 years of research. Personal. Individ.
Differ. 44:533–49
Shineman DW, Salthouse TA, Launer LJ, Hof PR, Bartzokis G, et al. 2010. Therapeutics for cognitive aging.
Ann. N. Y. Acad. Sci. 1191:E1–10
Singh-Manoux A, Ferrie JE, Lynch JW, Marmot M. 2005. The role of cognitive ability (intelligence) in
explaining the association between socioeconomic position and health: evidence from the Whitehall II
prospective cohort study. Am. J. Epidemiol. 161:831–39
Spearman C. 1904. “General intelligence” objectively determined and measured. Am. J. Psychol. 15:201–93
Stanovich KE. 2009. What Intelligence Tests Miss: The Psychology of Rational Thought. New Haven, CT: Yale
Univ. Press
Steer CD, Davey Smith G, Emmett PM, Hibbeln JR, Golding J. 2010. FADS2 polymorphisms modify the
effect of breastfeeding on child IQ. PLoS One 5:e11570
Sternberg RJ, Grigorenko EL. 2002. The General Factor of Intelligence: How General Is It? Mahwah, NJ: Erlbaum
Strand S, Deary IJ, Smith P. 2006. Sex differences in cognitive abilities test scores: a UK national picture. Br.
J. Educ. Psychol. 76:463–80
Strenze T. 2007. Intelligence and socioeconomic success: a meta-analytic review of longitudinal research.
Intelligence 35:401–26
Thompson PM, Cannon TD, Narr KL, van Erp T, Poutanen VP, et al. 2001. Genetic influences on brain
structure. Nat. Neurosci. 4:1253–58

www.annualreviews.org • Intelligence 481

 PS63CH18-Deary ARI 31 October 2011 12:36

Tucker-Drob EM. 2011. Global and domain-specific changes in cognition throughout adulthood. Dev. Psychol.
47:331–43
Turkheimer E, Haley A, Waldron M, D’Onfrio BM, Gottesman II. 2003. Socioeconomic status modified
heritability of IQ in young children. Psychol. Sci. 14:623–28
van der Maas HLJ, Dolan CV, Grasman RP, Wicherts JM, Huizenga HM, Raijmakers ME. 2006. A dynamical
model of general intelligence: the positive manifold of intelligence by mutualism. Psychol. Rev. 13:842–60
Van Valen L. 1962. A study of fluctuating asymmetry. Evolution 16:125–42
Vernon PE. 1956. The Measurement of Abilities. London: Univ. London Press
Vinkhuyzen AAE, van der Sluis S, Posthuma D. 2011. Life events moderate variation in cognitive ability ( g)
in adults. Mol. Psychiatry 16:4–6
Visser BA, Ashton MC, Vernon PA. 2006. Beyond g: putting multiple intelligences to the test. Intelligence 34:
487–502
von Stumm S, Gale CR, Batty GD, Deary IJ. 2009. Childhood intelligence, behaviour, and locus of control as
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

determinants of intergenerational social mobility: the British Cohort Study 1970. Intelligence 37:329–40
von Stumm S, Macintyre S, Batty GD, Clark H, Deary IJ. 2010. Intelligence, social class of origin, childhood
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

behavior disturbance and education as predictors of status attainment in midlife in men: the Aberdeen
Children of the 1950s study. Intelligence 38:202–11
Voracek M, Haubner T. 2008. Twin-singleton differences in intelligence: a meta-analysis. Psychol. Rep.
102:951–62
Waiter GD, Deary IJ, Staff RT, Murray AD, Fox HC, et al. 2009. Exploring possible neural mechanisms of
intelligence differences using processing speed and working memory tasks: an fMRI study. Intelligence
37:199–206
Webbink D, Posthuma D, Boomsma DI, de Geus EJC, Visscher PM. 2008. Do twins have lower cognitive
ability than singletons? Intelligence 36:539–47
Weiser M, Zarka S, Wrbeloff N, Kravitz E, Lubin G. 2009. Cognitive test scores in male adolescent cigarette
smokers compared to non-smokers: a population-based study. Addiction 105:358–63
Whalley LJ, Deary IJ. 2001. Longitudinal cohort study of childhood IQ and survival up to age 76. Br. Med. J.
322:819–22
White J. 2006. Intelligence, Destiny and Education: The Ideological Roots of Intelligence Testing. London: Routledge
Wilhelm O, Engle RW, eds. 2005. Handbook of Measuring and Understanding Intelligence. London: Sage
Wilson RS, Beckett LA, Barnes LL, Schneider JA, Bach J, et al. 2002. Individual differences in rates of change
in cognitive abilities of older persons. Psychol. Aging 17:179–93
Wisdom NM, Callahan JL, Hawkins KA. 2011. The effects of apolipoprotein E on non-impaired cognitive
function: a meta-analysis. Neurobiol. Aging 32:63–74
Woolley AW, Chabris CF, Pentland A, Hashmi N, Malone TW. 2010. Evidence for a collective intelligence
factor in the performance of human groups. Science 330:686–88
Wright Gillham N. 2001. A Life of Sir Francis Galton: From African Exploration to the Birth of Eugenics. Oxford,
UK: Oxford Univ. Press
Yeo RA, Gangestad SW, Liu J, Calhoun VD, Hutchison KE. 2011. Rare copy number deletions predict
individual variation in intelligence. PLoS One 6:e16339
Zhang J-P, Burdick KE, Lencz T, Malhotra AK. 2010. Meta-analysis of genetic variation in DTNBP1 and
general cognitive ability. Biol. Psychiatry 68:1126–33

482 Deary
 PS63-FrontMatter ARI 10 November 2011 9:52

Annual Review of
Psychology

Volume 63, 2012 Contents

Prefatory
Working Memory: Theories, Models, and Controversies
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

Alan Baddeley p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 1
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Developmental Psychobiology
Learning to See Words
Brian A. Wandell, Andreas M. Rauschecker, and Jason D. Yeatman p p p p p p p p p p p p p p p p p p p p p31
Memory
Remembering in Conversations: The Social Sharing
and Reshaping of Memories
William Hirst and Gerald Echterhoff p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p55
Judgment and Decision Making
Experimental Philosophy
Joshua Knobe, Wesley Buckwalter, Shaun Nichols, Philip Robbins,
Hagop Sarkissian, and Tamler Sommers p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p81
Brain Imaging/Cognitive Neuroscience
Distributed Representations in Memory: Insights from Functional
Brain Imaging
Jesse Rissman and Anthony D. Wagner p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 101
Neuroscience of Learning
Fear Extinction as a Model for Translational Neuroscience:
Ten Years of Progress
Mohammed R. Milad and Gregory J. Quirk p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 129
Comparative Psychology
The Evolutionary Origins of Friendship
Robert M. Seyfarth and Dorothy L. Cheney p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 153
Emotional, Social, and Personality Development
Religion, Morality, Evolution
Paul Bloom p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 179

vi
 PS63-FrontMatter ARI 10 November 2011 9:52

Adulthood and Aging

Consequences of Age-Related Cognitive Declines
Timothy Salthouse p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 201
Development in Societal Context
Child Development in the Context of Disaster, War, and Terrorism:
Pathways of Risk and Resilience
Ann S. Masten and Angela J. Narayan p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 227
Social Development, Social Personality, Social Motivation, Social Emotion
Social Functionality of Human Emotion
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.

Paula M. Niedenthal and Markus Brauer p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 259

Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Social Neuroscience
Mechanisms of Social Cognition
Chris D. Frith and Uta Frith p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 287
Personality Processes
Personality Processes: Mechanisms by Which Personality Traits
“Get Outside the Skin”
Sarah E. Hampson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 315
Work Attitudes
Job Attitudes
Timothy A. Judge and John D. Kammeyer-Mueller p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 341
The Individual Experience of Unemployment
Connie R. Wanberg p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 369
Job/Work Analysis
The Rise and Fall of Job Analysis and the Future of Work Analysis
Juan I. Sanchez and Edward L. Levine p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 397
Education of Special Populations
Rapid Automatized Naming (RAN) and Reading Fluency:
Implications for Understanding and Treatment of Reading Disabilities
Elizabeth S. Norton and Maryanne Wolf p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 427
Human Abilities
Intelligence
Ian J. Deary p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 453
Research Methodology
Decoding Patterns of Human Brain Activity
Frank Tong and Michael S. Pratte p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 483

Contents vii
 PS63-FrontMatter ARI 10 November 2011 9:52

Human Intracranial Recordings and Cognitive Neuroscience

Roy Mukamel and Itzhak Fried p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 511
Sources of Method Bias in Social Science Research
and Recommendations on How to Control It
Philip M. Podsakoff, Scott B. MacKenzie, and Nathan P. Podsakoff p p p p p p p p p p p p p p p p p p p p 539
Neuroscience Methods
Neuroethics: The Ethical, Legal, and Societal Impact of Neuroscience
Martha J. Farah p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 571

Indexes
Access provided by Universidad Nacional de Colombia on 08/01/21. For personal use only.
Annu. Rev. Psychol. 2012.63:453-482. Downloaded from www.annualreviews.org

Cumulative Index of Contributing Authors, Volumes 53–63 p p p p p p p p p p p p p p p p p p p p p p p p p p p 593

Cumulative Index of Chapter Titles, Volumes 53–63 p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 598

Errata

An online log of corrections to Annual Review of Psychology articles may be found at

http://psych.AnnualReviews.org/errata.shtml

viii Contents