ciation between memory and intelligence. Even after the effects of age have been
removed, general memory accounts for more than 17% of the variability in gen-
eral intelligence. In addition, general memory contributes significantly to general
intelligence even after both age and general RT have been controlled.
General RT, although also correlating highly with general intelligence, does
not correlate as highly as does general memory. RT and intelligence share ap-
proximately 18% common variance. However, after age is partialed out, RT no
longer contributes significantly. The variance overlapping between intelligence
and RT can be accounted for almost entirely by age.
These findings persist when subjects are divided into three age groups and
regression analyses are performed separately within each group. For 4-year-olds,
5-year-olds and 6-year-olds, general RT fails to predict any significant variance
in general intelligence. Alternatively, general memory accounts for significant
variance in general intelligence both when entered into the equation first and
when entered second, accounting for as much as 45% of the variance in 4-year-
olds, 48% in 5-year-olds, and 58% in 6-year-olds. Interestingly, the variance in
intelligence accounted for by memory shows an increasing trend across the three
age groups, implying that as age increases from 4 to 6 years, memory and intel-
ligence tend to overlap increasingly. Miller and Vernon (1992) found that general
memory accounted for 21% of the variance in general intelligence in adults. It
would therefore appear that the trend of increasing overlap between memory and
intelligence in children from 4 to 6 years is either artifactual, or indicative of a
trend in which the memory-intelligence overlap increases to a particular point in
development and then declines or plateaus. It would be necessary to test the
relationship at numerous ages over a wide age range to fully examine the devel-
opmental changes in the relationship between memory and intelligence.
It should be noted that the memory span tests used in this study were some-