THINKING CRITICALLY Are Intelligence Tests Unfairly Biased Against Certain Groups

283
Evaluating Intelligence Tests
In short, it appears that some important nongenetic factors serve to decrease the
average scores of African American and Hispanic American children. Whatever heredity might contribute to children’s performance, it may be possible for them to improve,
given the removal of negative environmental conditions. (“In Review: Influences on IQ
Scores” summarizes our discussion of environmental and genetic factors affecting performance on intelligence tests.)
F e a t u r e Ty p e : i n . r e v i e w
T H I N K I N G C R I T I C A L LY
S
ummarizing a person’s cognitive abilities
with an IQ score runs the risk of overAre Intelligence Tests
simplifying reality and making errors,
Unfairly Biased Against
but intelligence tests can also prevent errors.
Certain Groups?
If boredom makes a child appear mentally
slow, or even retarded, a properly conducted
test is likely to reveal the child’s potential. And, as Binet had hoped, intelligence tests
have been enormously helpful in identifying children who need special educational
attention. These tests can minimize the chances of assigning children to special classes
that they don’t need or to advanced work that they cannot yet handle.
Still, there is great concern over the fact that members of ethnic minorities and other
environmentally disadvantaged groups have not had an equal chance to develop the
knowledge and skills that are required to achieve high IQ scores.
■ What am I being asked to believe or accept?
Some critics claim that, indeed, standard intelligence tests are not fair. They argue that a
disproportionately large number of people in some ethnic minority groups score low on
intelligence tests for reasons that are unrelated to cognitive ability, job potential, or other
criteria that the tests are supposed to predict (Helms, 1992, 1997; Kwate, 2001; Neisser
et al., 1996). They say that using intelligence tests to make decisions about people—such
as assigning them to particular jobs or special classes—causes members of certain groups
to be unfairly deprived of equal employment or educational opportunities.
■ Is there evidence available to support the claim?
Research reveals several possible sources of bias in intelligence test scores. First, noncognitive factors can influence a person’s performance on IQ tests and may put certain groups
at a disadvantage. We have seen, for example, that children from some ethnic and socioeconomic groups may be less motivated than other children to perform well on standardized tests. They may also be less comfortable in the testing situation and less likely
to trust adult testers (Steele, 1997). So the differences in test scores may partly reflect
motivational or emotional differences among various groups, not intellectual ones.
Second, many intelligence test items are still drawn from the vocabulary and experiences of the dominant middle-class culture in the United States. As a result, these tests
often measure achievement in acquiring knowledge valued by that culture. Not all
cultures value the same things, however (Nisbett, 2003; Serpell, 1994; Sternberg &
Grigorenko, 2004b). For example, a study of Cree Indians in northern Canada revealed
that words and phrases meaning “competent” included good sense of direction. At the
“incompetent” end of the scale was the phrase lives like a white person (Berry & Bennett,
1992). A European American might not perform well on a Cree intelligence test based
on these criteria. In fact, as illustrated in Figure 7.13, poor performance on a culturespecific test is probably due more to unfamiliarity with culture-based concepts than to
lack of cognitive ability.
Third, some tests may reward those who interpret questions as expected by the test
designer. Conventional intelligence tests have clearly defined “right” and “wrong”
answers. Yet a person may interpret test questions in a manner that is “intelligent” or
284
7.13
An Intelligence Test?
How did you do on this “intelligence test”? If, like most peoby
ple, you are unfamiliar with
the material being tested by these rather
obscure questions, your score was probably low. Would it be fair to say, then, that
you are not very intelligent?
doing
2
learn
Take a minute to answer each of these questions, and check your answers
against the key below.
1. What fictional detective was created by Leslie Charteris?
2. What planet travels around the sun every 248 years?
3. What vegetable yields the most pounds of produce per acre?
4. What was the infamous pseudonym of broadcaster Iva Toguri d’Aquino?
5. What kind of animal is Dr. Dolittle’s Pushmi-Pullyu?
Answers: (1) Simon Templar (2) Pluto (3) Cabbage (4) Tokyo Rose (5) A two-headed llama
FIGURE
Chapter 7 Thought, Language, and Intelligence
“correct,” but that produces a “wrong” answer. For example, when one child was asked,
“In what way are a pen and a pencil alike?” he replied, “They can both poke holes in
paper.” The fact that you don’t give the answer that the test designer was looking for
does not mean that you can’t. When Liberian rice farmers were asked to sort objects,
they tended to put a knife in the same group as vegetables. This was the clever way to
do it, they said, because the knife is used to cut vegetables. When asked to sort the
objects as a “stupid” person would, the farmers grouped the cutting tools together, the
vegetables together, and so on, as most North Americans would (Segall et al., 1990).
■ Can that evidence be interpreted another way?
The same evidence might be interpreted as showing that although traditional intelligence tests do not provide a pure measure of inherited cognitive ability, they do provide a fair test of whether a person is likely to succeed in school or in certain jobs.
When some people have had more opportunity than others to develop their abilities,
the difference will be reflected in higher IQ scores. From this point of view, intelligence
tests are fair measures of the cognitive abilities people have developed while living in
a society that, unfortunately, contains some unfair elements. In other words, the tests
may be accurately detecting knowledge and skills that are not represented equally in all
groups. That doesn’t mean that the tests discriminate unfairly among those groups.
To some observers, concern over cultural bias in intelligence tests stems from a tendency to think of IQ scores as measures of innate ability. These psychologists suggest
instead that intelligence tests are measuring ability that is developed and expressed in
a cultural context—much as athletes develop the physical skills needed to play certain
sports (Lohman, 2004). Eliminating language and other cultural elements from intelligence tests, they say, would eliminate a vital part of what the term intelligence means
in any culture (Sternberg, 2004). This may be the reason that “culture-fair” tests do not
predict academic achievement as well as conventional intelligence tests do (Aiken, 1994;
Lohman, 2005). Perhaps familiarity with the culture reflected in intelligence tests is just
as important for success at school or work in that culture as it is for success on the
tests themselves. After all, the ranking among groups on measures of academic
achievement is similar to the ranking for average IQ scores (Sue & Okazaki, 1990).
■ What evidence would help to evaluate the alternatives?
If the problem of test bias is really a reflection of differences between various groups’
opportunities to develop their cognitive skills, it will be important to learn more about
how to reduce those differences. Making “unfair” cultures fairer by enhancing the skill
development opportunities of traditionally disadvantaged groups should lead to smaller
differences between groups on tests of cognitive ability (Martinez, 2000). It will also be
important to find better ways to encourage members of disadvantaged groups to take
advantage of those opportunities (Sowell, 2005).