Psycholog)
in
the
Schools
1981.
f8.
139-143
ITEM ANALYSES
OF
THE CONCEPT ASSESSMENT KIT-CONSERVATION
FOR EDUCABLE MENTALLY RETARDED CHILDREN'
A. B. SILVERSTEIN, LINDA BROWNLEE, AND GREG LEGUTKI
MRRC-Lanterrnan State Hospitaf Research
Group,
UCLA
School
of
Medicine
A series
of
item analyses of the CAK-C was conducted for a sample of
155
educable
mentally retarded children. The probability of a correct response was found to differ
from task to task, and there was evidence that the order of difficulty of the tasks for
this sample resembled that for nonretarded children.'The probabilities
of
the two in-
correct responses were generally not equal, and the choice of one
or
the other incorrect
response showed some relation to CA, MA, and
IQ,
particularly the last two
variables.
Scales constructed within the conceptual framework of Piagetian theory represent
one alternative to traditional intelligence tests, and the Concept Assessment Kit-
Conservation (CAK-C) (Goldschmid
&
Bentler,
1968)
is
one scale
of
this sort. Two
parallel forms of the scale, A and B, include tasks designed to measure the conservation
of
two-dimensional space, number, substance, continuous quantity, weight, and discon-
tinuous quantity. An example should make clear the nature of the tasks. The examiner
makes two equal balls of Play-Doh (following Elkind,
1967,
we designate them
S
and V,
denoting standard and variable), and establishes that the child regards them as the same.
One ball is then rolled into a hot dog
or
flattened into a pancake (V-+V') and the child is
asked whether
S
and
V'
have the same amount of Play-Doh,
or
whether one of them has
more, and to explain why. The scale yields a behavior score, an explanation score, and a
total (behavior plus explanation) score, and the child's performance is taken as an indica-
tion
of
his
or
her level of intellectual development.
In a previous study (Silverstein, Brownlee,
&
Legutki,
1980),
we investigated the
reliability of the CAK-C for educable mentally retarded children. In the present study,
we conducted a series of item analyses of the scale for subjects from this population.
Specifically, these analyses were addressed to the following questions: (a) Does the
probability of a correct response (S=V') differ from task to task? (b)
Is
the order
of
difficulty of the tasks the same for educable mentally retarded children as it is for non-
retarded children? (c) Are the probabilities
of
the two incorrect responses (S>V'
&
S<V')
equal? (d) Is the choice of one
or
the other incorrect response related to CA, MA,
or
IQ?
Piagetian theory suggests that the first two questions should be answered affir-
matively, since the theory posits an invariant developmental sequence en route to master-
ing the concept of conservation. Further, retarded children are supposed to traverse
precisely the same sequence as nonretarded children, differing only
in
the rate at which
they progress and the ultimate ceiling they attain (Weisz
&
Zigler,
1979).
The third and
fourth questions are more exploratory
in
nature. In traditional intelligence testing, atten-
tion usually is focused on the child's ability to produce correct responses, but Piaget
himself is said to have discovered-while working as a young man at the Binet
Requests for reprints should be sent to A.
B.
Silverstein, MRRC-Lanterman State Hospital Research
'This
study
was
supported
in
part by
NICHD
Research Grants
No.
HD-04612 and
No. HD-05540.
Group,
P.O.
Box
100-R,
Pomona, CA
91769.
139