BOOKS
309
reasonable to suppose that black children do less well in school, on
the average, than white children in part because their teachers expect
less of them. (Books like
Bias in Mental Testing
help shape such
expectations. )
How does Jensen deal with the literature on expectancy effects? "I
have not found in the literature a single bona fide study showing an
expectancy effect for IQ," he writes, and he goes on to cite 13 studies,
none of which reported a statistically significant effect. Four of these
studies, however, measured expectancy effects on something other than
IQ and one of them did not measure an expectancy effect at all. Jensen
does not explain how he chose this flawed and unrepresentative
sample.
If
IQ tests measure an inborn trait, IQ scores should remain
reasonably constant in the course of individual development. Do they?
The question is purely empirical, and it has been answered. The
definitive study was carried out by R. B. McCall, M.1. Appelbaum, and
P. S. Hogarty of the Fels Research Institute. They administered modern
IQ tests
to
a cohort of middle-class children at regular intervals over a
period of fifteen years. They found that "normal home-reared middle–
class children change in IQ performance during childhood, some by a
substantial amount. ... The average individual's range of IQ between
2Y2
and 17 years of age was 28.5 IQ points, one of every three children
displayed a progressive change of more than 30 points, and one in
seven shifted more than 40 points. Rare individuals may alter their
performance as much as 74 points." The study also showed that
children's test scores vary systematically and in different ways as they
mature. Some children do progressively better on the tests, some do
progressively worse. Some reach a plateau and stay there, others peak,
still others improve after a steady decline. Oddly enough, Jensen does
not mention this (or any comparable) study in his long chapter on the
reliability and stability of standardized tests, nor does it appear among
the book's 800-odd references to the technical literature. Yet he asserts
that "the stability of mental test scores during the period from infancy
to maturity is roughly comparable to the stability of height and weight
measurements, although mental measurements more closely resemble
weight measurements in this respect."
The notion that intelligence or "reasoning ability" is a simple
trait has always appealed to those psychologists who believe that some
groups (the British upper class, northern Europeans) are more richly
endowed with intelligence than others (the British working class,
outhern Europeans, the Irish, Jews, blacks). But does it have any
