curacy level, such as 75%. Be clear that inspection time does not refer to how
long it takes a subject to make this simple discrimination; rather, it refers to
how long the stimulus was exposed in order that the subject might reach an
acceptable level of performance.
The main finding of interest is that inspection times correlate with perfor-
mance on standard tests of intelligence (e.g., Nettelbeck & 1976; Deary, 1993;
see Deary & Stough, 1996). People whose inspection times are short tend to
score higher on the intelligence tests. Across a variety of studies the correlation
is usually around .5, a moderately strong correlation (Deary & Stough, 1996).
One interpretation of the correlation is that inspection time measures a basic
information processing component—namely, the speed with which informa-
tion is taken in or initially perceived.
Other information processing measures have also been correlated to IQ.
These include estimates of the span of working or short-term memory (Hunt,
1978; Schofield & Ashman, 1986; Daneman & Carpenter, 1980; Dark & Benbow,
1991; see Dempster, 1981), the speed with which subjects supposedly scan short
term memory (Keating & Bobbit, 1978; Vernon, 1983), the speed with which
people mentally rotate a visual stimulus (Mumaw Pellegrino, Kail, & Carter,
1984), the speed with which people access the name of a letter (Hunt, 1978,
1983; Hunt, Lunneborg & Lewis, 1975), and the speed with which subjects ac-
cess the meaning of a word in memory (Goldberg, Schwartz, & Stewart, 1977;
Vernon, 1983). Measures of the speed of information processing tasks correlate
with scores on IQ tests even when the IQ test itself is not timed (Vernon &
Kantor, 1986).
Problems with the Information Processing Perspective on Intellectual Differences
There are, however, problems with the information processing account of indi-
vidual differences in cognition. One problem is that not every researcher finds a
correlation between measures of the speed or efficiency of a component and IQ
performance (e.g., Keating, 1982; Ruchalla, Scholt & Vogel, 1985; see Long-
streth, 1984; Barrett, Eysenck, & Luching, 1989). Further, when a correlation
is found, that correlation is often achieved by comparing college students to
mentally retarded people. When the studies are done using subjects who are
not mentally retarded, the correlation between any estimate of the speed with
which a cognitive component is executed and IQ scores is usually quite modest,
in the .3 to .4 range (see Kline, 1991; Mackintosh, 1986). The correlation be-
tween inspection time and IQ scores seems a bit more robust, however (Deary
& Sough, 1996).
A more fundamental problem is that the information processing approach
relies too much on establishing correlations between measures of information
processing and IQ scores. What is generally lacking from this line of inquiry are
demonstrations that measures of information processing can predict perfor-
mance on real life tasks better than conventional IQ tests (Richardson, 1991).
Another difficulty with the information processing approach to individual
differences is that of establishing cause and effect. Is the efficiency with which
information is initially processed the cause of intelligence, or is speed of pro-
cessing the effect of intelligence, whose cause is undetermined? Even if it is
conceded that perception speed, as measured in tasks like the inspection time
786 R. Kim Guenther