IQ Scores as a Measure of Inherited Ability

279
Evaluating Intelligence Tests
The results of their research suggest that intelligence test scores are most valid for
assessing aspects of intelligence that are related to schoolwork, such as abstract reasoning and understanding verbal material. Their validity—as measured by correlating IQ scores with high school grades—is reasonably good, about .50 (Brody &
Erlichman, 1998). Scores on tests that focus more specifically on reasoning skills show
even higher correlations with school performance (Kuncel, Hezlett, & Ones, 2004;
Lohman & Hagen, 2001).
There is also evidence that employees who score high on verbal and mathematical
reasoning tests tend to perform better at work than those who earn lower scores
(Borman, Hanson, & Hedge, 1997; Johnson & Neal, 1998; Pulakos et al., 2002), especially if their jobs require complex reasoning and judgment skills (Gottfredson, 1997).
Later we describe a study that kept track of people for seventy years and found that those
who had high IQ scores as children tended to be well above average in terms of academic
and financial success in adulthood (Cronbach, 1996; Oden, 1968; Terman & Oden, 1947).
IQ scores also appear to be highly correlated with performance on routine tasks such as
reading medicine labels and using a telephone book (Gottfredson, 1997, 2004).
So, by the standard measures for judging psychological tests, scores on intelligence
tests have good reliability and good validity for predicting success in school and in
many life situations and occupations. However, an IQ score is not a perfect measure of
how “smart” a person is. Because intelligence tests do not measure the full array of cognitive abilities, a particular test score tells only part of the story, and even that part may
be distorted. Many factors other than cognitive ability, including reactions to the testing
situation, can influence test performance. For example, children who are suspicious of
strangers and adults who fear making mistakes may become anxious and fail even to
try answering certain questions, thus artificially lowering their IQ scores (Fagan, 2000).
Claude Steele and his colleagues have suggested that some people’s test-related anxiety
stems from a phenomenon known as stereotype threat (Steele & Aronson, 2000).
According to Steele, concern over negative stereotypes about the cognitive abilities of
the group to which they belong can impair the performance of some women—and
some members of ethnic minorities—such that the test scores they earn underestimate
those abilities (Blascovich et al., 2001; Cadinu et al., 2005; Inzlicht & Ben-Zeev, 2000).
The extent to which stereotype threat actually affects performance on cognitive abilities
tests is uncertain, however (Cullen, Hardison, & Sackett, 2004; Sackett, Hardison, &
Cullen, 2004).
IQ Scores as a Measure of Inherited Ability
Alfred Binet believed that intelligence could be improved with training and practice at
mental tasks. Lewis Terman saw it as an inherited characteristic. Both were partly right.
Years of research have led psychologists to conclude that intelligence is developed ability. This means that intelligence is influenced partly by genetics but also by educational,
cultural, and other environmental factors and experiences that shape the very knowledge, reasoning, and other skills that intelligence tests measure (Atran, Medin, & Ross,
2005; Garlick, 2003; Plomin & Spinath, 2004).
To explore the influence of genetics on individual differences in IQ scores, psychologists have compared the correlations in scores between people who share varying
degrees of similarity in genetic makeup and environment. For example, they have examined the IQ scores of identical twins (pairs with exactly the same genes) who were separated when very young and raised in different environments. They have also examined the scores of identical twins raised together. (Research designs used in these
behavioral genetics studies are described in the introductory chapter.)
These studies find, first, that genetic factors are strongly related to IQ scores. When
identical twins who were separated at birth and adopted by different families are tested
many years later, the correlation between their scores is usually high and positive, at
least .60 (Bouchard, 1999). If one twin receives a high IQ score, the other probably
will, too; if one is low, the other is likely to be low as well. However, studies of IQ correlations also highlight the importance of the environment (Scarr, 1998). Consider any
280
7.11
Correlations of IQ Scores
The correlation in IQ between pairs
increases with increasing similarity in
heredity or environment.
+1.00
Correlation of IQ scores
FIGURE
Chapter 7 Thought, Language, and Intelligence
+0.90
+0.80
+0.70
+0.60
+0.50
+0.40
+0.30
+0.20
+0.10
Identical twins reared together
Siblings reared apart
Identical twins reared apart
Unrelated children reared together
Nonidentical twins reared together
Unrelated children reared apart
Siblings reared together
Source: Bouchard & McGue (1981).
two people—twins, nontwin siblings, or unrelated children—brought up together in a
foster home. No matter what the degree of genetic similarity in these pairs, the correlation between their IQ scores is higher if they share the same home than if they are
raised in different homes, as Figure 7.11 shows (Scarr & Carter-Saltzman, 1982).
The role of environmental influences is also seen in studies that compare children’s
IQ scores before and after environmental changes such as adoption (van IJzendoom &
Juffer, 2005). Generally, when children from relatively impoverished backgrounds were
adopted into homes offering a more enriching intellectual environment—including
interesting materials and experiences, as well as a supportive, responsive adult—they
showed modest increases in their IQ scores (Weinberg, Scarr, & Waldman, 1992).
A study of French children who were adopted soon after birth demonstrates the
importance of both genetic and environmental influences. These children were tested
after years of living in their adopted homes. Those whose biological parents were from
higher socioeconomic groups (where higher IQs are more common) had higher IQ
scores than those whose biological parents came from lower socioeconomic groups,
regardless of the socioeconomic status of the adopted homes (Capron & Duyme, 1989,
1996). These findings are supported by data from the Colorado Adoption Project
(Cardon & Fulker, 1993; Cardon et al., 1992), and they suggest that a genetic component of the children’s cognitive abilities continued to exert an influence in the adoptive
environment. At the same time, when children from low socioeconomic backgrounds
were adopted by parents who provided academically enriched environments, their IQ
scores rose by twelve to fifteen points (Capron & Duyme, 1989).
Programs designed to enhance young children’s school readiness and academic ability have also been associated with improved scores on tests of intelligence (Neisser et al.,
1996; Ripple et al., 1999). These early-intervention programs may be partly responsible for the steady increase in average IQ scores seen throughout the world over the past
six decades (Flynn, 1999; Neisser, 1998).
Some researchers have concluded that the influences of genetic and environmental factors on intelligence appear to be about equal. Others see a somewhat larger role
for genetic factors (Herrnstein & Murray, 1994; Loehlin, 1989; Petrill et al., 1998;
Plomin, 1994). One research team has even suggested that specific genes are associated with extremely high IQs (Chorney et al., 1998). Still, it is important to
understand that any estimates of the relative contributions of heredity and environment apply only to groups, not to individuals. It would be inaccurate to say that
50 percent of your IQ score is inherited and 50 percent learned. It is more accurate
to say that about half of the variability in the IQ scores of a group of people can be