Pyyychology
in
the
Schools
Volume
31.
Ocroher
1994
GF-GC THEORY OF HUMAN COGNITION:
DIFFERENTIATION OF SHORT-TERM AUDITORY AND
VISUAL MEMORY FACTORS
RON McGHEE LEWIS LIEBERMAN
Auburn University Columbus College
A
principal component factor analysis with varimax rotation was performed for
50
subjects on
18
cognitive and achievement variables. Four, five, six, and seven factors
were specified
for
extraction. Variables were chosen
to
allow for investigation of factor
loadings among measures of short-term auditory and short-term visual memory.
Also
included in the analysis were tests representative of other Gf-Gc factors. The results
provide support for a model of Gf-Gc theory with distinct and separate short-term
visual and short-term auditory memory factors.
In the field of intelligence assessment there has been increased interest (McGhee,
1993;
Woodcock,
1990)
in verifying the factorial structure of human abilities hypothesized
by Horn and Cattell
(1966)
in their Gf-Gc theory. Carroll and Horn
(1981)
presented
evidence that about
80%
of intellectual ability could be measured and predicted in terms
of as few as
30
basic processes. Most of the variation represented in primary processes
could be further organized in terms of as few
as
8
or
9
second-order abilities. More
recently, Horn
(1988)
presented a model of Gf-Gc theory based on
10
abilities, including
two sensory detectors (Gs), quantitative ability (Gq), visualization capacity (Gv), general
perceptual speed (Gs), memory capacity (Gm), fluid (Gf) and crystallized (Gc) abilities,
general retrieval capacity (Gr), and auditory capacity (Ga). Table
1
provides descrip-
tions of factors in current Gf-Gc theory.
McGhee’s
(1993)
joint confirmatory factor analysis of the Woodcock-Johnson-
Revised (WJ-R; Woodcock
&
Johnson,
1989),
Differential Ability Scales (Elliott,
1990),
and Detroit Tests of Learning Aptitude-3 (Hammill,
1991)
indicated that an eight-factor
model of intelligence based on Gf-Gc theory best fit the data. Horn and Cattell’s Gs,
Gr, Ga, Gv, Gc, and Gf factors proved stable. However, the Gsm factor split into short-
term auditory (Gsma) and short-term visual (Gsmv) factors. McGhee hypothesized that
these results were contrary to previous research (Woodcock,
1990),
which suffered from
“underfactoring.” Most factor analytic studies of intelligence tests have not included
enough marker variables to distinguish between the two types of memory. For example,
the WJ-R contains two “pure” measures of short-term auditory memory but no “pure”
measures of short-term visual memory. A similar argument can be leveled against the
WISC-R (Wechsler,
1974),
WAIS-R (Wechsler,
1981),
and, presumably, the WISC-I11
(Wechsler,
1991).
These two intelligence tests, when factor-analyzed with other in-
telligence batteries (Woodcock,
1990),
demonstrate underlying Gc and Gv factors. The
arithmetic subtest is a lone measure of Gq, the Coding subtest is a lone measure of
Gs,
and the Digit Span subtest is
a
lone measure of Gsma. Because there are no other markers
for these factors, they do not emerge when factor-analyzed. McGhee suggested that
further investigation into the differentiation of short-term auditory and visual memory
factors might take the form of factor-analyzing data from tests requiring only recall
of
previously seen material (Gsmv), tests requiring only recall
of
previously heard material
Kequehts
for reprints should be sent
lo
Ron
McChee,
304
Martindale
Lane,
Albany, CA
31707.
297