Interpreting WISC-R and McCarthy scatter: A caution

Interpreting WISC-R and McCarthy scatter: A caution

CONTEMPORARY interpreting EDUCATIONAL PSYCHOI.OGY 7, 90-94 (1982) WISC-R and McCarthy JACK Norrhern Scatter: A Caution A. NAGLIERI Arizona lJ...

272KB Sizes 0 Downloads 50 Views

CONTEMPORARY

interpreting

EDUCATIONAL

PSYCHOI.OGY

7, 90-94 (1982)

WISC-R and McCarthy JACK Norrhern

Scatter:

A Caution

A. NAGLIERI Arizona

lJt1iversil.v

In this investigation, indices of scatter on the WISC-R and McCarthy Scales were examined for 20 educable mentally retarded and 20 learning disabled children in relation to a similar control group. The scatter found in the two exceptional samples was also compared to the variability found in the standardization samples of the WISC-R and McCarthy. Results indicated that the learning-disabled sample exhibited more WISC-R subtest scatter and McCarthy Scale variability than the standardization sample but nor more than the control group. The implications of these findings are discussed.

Intratest variability such as Verbal-Performance discrepancies and subtest scatter on the Wechsler Scales, and Scale Index fluctuations on the McCarthy are typically used to determine an individual’s intellectual strengths and weaknesses. This information can then be interpreted into educational recommendations and remedial plans. While V-P discrepancies, for example, have also been used to suggest exceptionality (Clements, 1966; Holroyd & Wright, 1963, the degree to which such fluctuations occur in the normal population has not been adequately considered. Kaufman (1979) suggests that only significant variability which also occurs infrequently in the normal population should be considered as evidence of exceptionality. As a means of comparison, Kaufman (1976a, 1976b) presents an analysis of the subtest variability and Verbal-Performance differences for the WISC-R and variability of the McCarthy Scale Indexes (Kaufman, 1976c) for the respective standardization samples. He suggests that these figures can then be used as basal levels of variability against which a particular child’s profile can be compared. However, comparisons with the normative population alone may be misleading without a local comparison group as well. In this investigation, the variability found within the WISC-R and McCarthy Scales for two groups of exceptional children was compared to a similar control group in addition to the respective standardization data presented by Kaufman. METHOD Subjects A sample of 60 children between the ages of 6-4 and 8-l 1 (mean = 7-8; SD = 9 months) was obtained from several elementary schools in a community of approximately 20,000 1 am sincerely grateful for the aid of Dr. Alan S. Kaufman in this investigation which was in partial fulfillment of the doctoral degree. Requests for reprints should be sent to Dr. Jack A. Naglieri, Department of Psychology, Box 6002, Northern Arizona University, Flagstaff, AZ 86011. 90 0361-476X’82/010090-05$02.00/O Copyright @ 1982 by Academic Press, Inc. All rights of reproduction in any form reserved.

91

WISC-R AND MC CARTHY SCATTER TABLE I DESCRIPTIONOF THE SAMPLES

Age Group

N

Range

Mean

SD

Educable mentally retarded Learning disabled Normal

20 20 20

6-6 to 8-11 6-6 to 8-l 1 6-4 to 8-8

7-10 8-l 7-2

IO months IO months 9 months

Sex and race Males BVWh* Educable mentally retarded Learning disabled Normal

8 9 9

Females BliWh 7 5 6

3 5 2

2 I 3

Total Ma/Fe+ I5 14 I5

5 6 5

Total BVWh 11 I4 II

9 6 9

Note. * Black/white.

+ Male/female. located on the Georgia coastline. This sample was organized into three groups of 20 children previously identified as learning disabled (LD), educable mentally retarded (EMR), and normal so that the three samples would be equal according to age, sex, and race. The LD children were identified and placed in special educational programs by selection committees according to state and federal guidelines. Criteria included at least one IQ score within the average range, a discrepancy between achievement and intellectual ability, a significant WISC-R Verbal-Performance discrepancy, and WISC-R subtest scatter. The EMR children were identified and placed in special education programs by selection committees on the basis of IQ scores between 55 and 69, impaired adaptive functioning, and school failure. The normal children were defined as those enrolled in regular classrooms who were not referred for an evaluation. A description of the samples is presented in Table I.

Procedure The McCarthy and WISC-R were administered to all children by six certified white examiners (3 males and 3 females) in two sessions (median interval = 2.0 days). Approximately half of the children in each group were administered the WISC-R followed by the McCarthy and the remainder of the children were given the McCarthy followed by the WISC-R. The difference between the WISC-R Verbal and Performance IQ scores were obtained by subtracting the smaller IQ from the larger one, and the subtest scatter was obtained by subtracting the lowest subtest scaled score from the highest (excluding the supplementary tests). The difference between the highest and lowest McCarthy Scale Index was similarly obtained using the Verbal, Perceptual-Performance, Quantitative, Memory, and Motor Scales. This difference was also computed excluding the Motor Index. The significance of the difference between these means for each group and the respective standardizations data reported by Kaufman (1976a, 1976b, 1976~)was tested by t tests for the difference between uncorrelated means (Walker & Lev, 1953,p. IS). The Bonferroni technique (Myers, 1972,p. 366) was employed to determine the appropriate level of significance needed to reduce the

92

JACK A. NAGLIERI

TABLE 2 COMPARISONOF THE OBTAINED WISC-R AND MCCARTHY SUBTEST AND SCALE VARIABILITY WITH THE STANDARDIZATION SAMPLE

Educable mentally retarded (N = 20) Variable WISC-R verbal performance difference WISC-R subtest scatter (10

Learning disabled (N = 20)

Normal (N = 20)

Standardization sample (N = 2,200)

Mean

SD

Mean

SD

Mean

SD

Mean

SD

9.6

9.0

13.6

9.6

12.6

1.1

9.7

7.6

6.6

3.1

8.5

3.0

8.0

2.1

1.0

2.1

12.4

8.1

22.2

1.1

16.6

1.2

14.4

6.0

9.9

6.0

16.6

8.1

13.2

6.1

11.9

5.4

subtests)

McCarthy indexes (5)” McCarthy indexes (4)”

n McCarthy Scale Indexes included: Verbal, Perceptual-Performance, Quantitative, Memory, and Motor. b McCarthy Scale Indexes included: Verbal, Perceptual-Performance, Quantitative, and Memory (Motor is excluded).

probability of at least one erroneous rejection to .06. Therefore, only those contrasts significant atp < .Ol were accepted. The significance of the difference between the EMR, LD, and normal group means was determined by t tests for the difference between uncorrelated samples (Guilford & Fruchter, 1978, p. 1.51).

RESULTS

AND DISCUSSION

The data presented in Table 2 provide two points of reference for comparisons of the measures of variability on the WISC-R and McCarthy obtained by the exceptional children-the normal and standardization samples. Examination of this table reveals that the mean WISC-R V-P discrepancies for the LD and normal groups were higher than the mean for the standardization sample, although only the LD group’s mean approached significance (t = 2.25,~ < .05). In addition, there was no significant difference between the LD or normal group V-P means (t = 0.34). Similarly, the means for the WISC-R subtest scatter for the LD and normal groups were significantly greater than those for the standardization sample (t = 3.06, p < .Ol and t = 2.11, p < .05, respectively), but not significantly different from each other. Therefore, while the LD group appeared to evidence a WISC-R V-P discrepancy and subtest scatter greater than that of the standardization sample, it did not manifest more scatter than the control group.

WISC-R AND MC CARTHY SCATTER

93

Analysis of the variability found between the McCarthy Scale Indexes, including the Motor Index, reveals that the LD sample exhibited significantly more variability than the standardization sample (t = 5.69, p < .OOl) and the normal sample (t = 2.35, p < .05). When the Motor Scale was excluded from the analysis, the mean difference between the highest and lowest of the four cognitive scales remained significantly higher than that of the standardization sample (t = 3.78, p = .Ol) but not significantly different from the control group. Thus, the LD sample did manifest considerable variability in their cognitive and motor abilities, but not more than the control group when the Motor Index was excluded. The results of the present investigation suggest that the LD sample exhibited more WISC-R subtest scatter and McCarthy Scale variability (including and excluding the Motor Scale) than the standardization sample. However, the LD group did not demonstrate more WISC-R subtest scatter, a greater V-P IQ discrepancy, or greater variability between the four cognitive scales of the McCarthy compared to the normal group. This suggests that the scatter found in the LD sample was not a unique attribute of these learning-disabled children. Hence, one is faced with a choice. Compare the results of intratest variability to the standardization sample, which may not reflect local characteristics, or to a local norm group of limited sample size. Kaufman (1979) has argued that experimenters and psychologists should be aware of normal scatter before a diagnosis of exceptionality should be considered. Kaufman continues, saying that the variability found in the standardization sample for the WISC-R and McCarthy should be used as basal levels for comparison. However, as is evident from this investigation, comparison to the standardization sample alone can be misleading. The local normative population, in this case the normal group, offered valuable insight into what was typical for children who were very much like the exceptional samples. Both of these comparisons are necessary if one is to make well-informed decisions on the status of children’s abilities. The most decisive statements of exceptionality will follow when the comparison with the local and national norms are in agreement. REFERENCES S. D. Minimal

in children: Terminology and identijicationPhase one. Washington, D.C.: Department of Health, Education and Welfare, 1966.

CLEMENTS,

brain dysfunction

(NINDB Monograph No. 3, GUILFORD,

J. P., & FRUCHTER,

U.S. Public Health Service Publication B. Fundamental statistics in psychology

No. 1415) and education.

New York: McGraw-Hill, 1978. HOLROYD, J., & WRIGHT, F. Neurological implications of WISC Verbal-Performance discrepancies in a psychiatric setting. Journal of Consulting Psychology, 1965, 29, 206-212.

94

JACK A. NAGLIERI

KAUFMAN, A. S. Verbal-Performance

IQ discrepancies on the WISC-R. Journal of Con1976, 44, 160- 168. (a) A. S. A new approach to the interpretation of test scatter on the WISC-R. of Leurning Disabilities, 1976, 9, 160- 168. (b) A. S. Do normal children have “flat” ability profiles? Psychology in the 1976, 13, 284-285. (c) A. S. Intelligent testing with the WISC-R. New York: Wiley-Interscience,

sulting and CIinicul Psychology,

KAUFMAN, Journal

KAUFMAN, Schools,

KAUFMAN, 1979.

MYERS,J. L. Fundamentals of experirnentnl design. Boston: Allyn & Bacon, 1972. WALKER, H. M., & LEV, J. Statistical inference. New York: Holt, Rinehart & Winston, 1953.