Relationships of the WISC-R and K-BIT for an adolescent clinical sample.

Slate, John R. ; Graham, Linda Spear ; Bower, Jack 等

Students with disabilities, recognized under Public Law 94-142, who are enrolled in special education are required to have a psychological reevaluation within a three-year or less time period (USDHEW, 1977). Thus, students who have been diagnosed as having a Specific Learning Disability (SLD) or Mental Retardation (MR) will periodically undergo a psychological assessment that will include some form of intelligence test. In the initial evaluation, the intelligence test most likely to have been administered was the Wechsler Intelligence Scale for Children-Revised (WISC-R, Wechsler, 1974; Reschly & Wilson, 1990). Upon re-evaluation, however, readministration of the WISC-R may be unnecessary for two reasons. First, if test data subsequent to the initial evaluation are supportive of the original diagnosis, then administration of the same IQ test may be a less than wise use of limited school funds and professional and student time. Second, unless educational programming for a student is dependent on a new IQ score, administering the WISC-R would not assist school personnel in developing instructional strategies.

Even though readministration of the WISC-R might be unnecessary in a reevaluation, some form of assessment has been deemed essential (USDHEW, 1977). Because of limited resources and time demands, there is a need for assessing intelligence in a shorter time period than that permitted by the Wechsler scales. One such instrument is the Kaufman Brief Intelligence Test (K-BIT; Kaufman & Kaufman, 1990) which purports to assess intelligence. Comparable to the WISC-R, the K-BIT provides three scores: A Composite or IQ score; a Vocabulary subscale (similar to the WISC-R Verbal IQ); and a Matrices subscale (analogous to the WISC-R Performance IQ). The K-BIT and WISC-R scores have the same mean and standard deviation of 100 and 15, respectively.

Use of the K-BIT in special education assessment raises one crucial issue for students who were diagnosed with a disability and enrolled in special education with IQs obtained from use of the WISC-R. Because test scores differ between tests for several reasons (Bracken, 1988), including recency of test publication, using the K-BIT as the intellectual measure in the reevaluation has the potential to affect eligibility decisions. Thus, the extent to which K-BIT scores are comparable to scores on the WISC-R is important both for assessment specialists and for students with disabilities.

A CD-ROM search of Psychological Abstracts and ERIC revealed three studies in which differences in IQs between the WISC-R and the K-BIT were reported. The first study, reported in the K-BIT manual (Kaufman & Kaufman, 1990) on a sample of nondisabled elementary school students, revealed that the K-BIT and WISC-R were significantly related, r = + .80, although the K-BIT provided a Composite score that was approximately six points lower than the WISC-R Full Scale IQ. In the second study (Prewett, 1992a), a sample of 35 students referred because of academic difficulties yielded statistically significant correlations ranging from + .58 to + .92 between the three WISC-R IQs and their K-BIT counterparts. Corresponding to the study using nondisabled students who were not experiencing academic difficulties, the K-BIT Composite was approximately six points lower than the WISC-R Full Scale IQ. The third study (Prewett, 1992b), on a sample of 40 male incarcerated juvenile delinquents, also revealed significant correlations among the WISC-R and K-BIT scores. Unlike the previous two studies, however, the mean difference between the WISC-R Full Scale IQ and the K-BIT Composite score was only .45 points.

These three studies provide evidence that K-BIT scores are generally lower than their WISC-R counterparts. No study, however, examined differences in WISC-R and K-BIT scores for students already diagnosed with a disability. Moreover, none of the studies investigated the longitudinal relationship between these two measures for a clinical sample. Instead, the studies focused on the concurrent validity of these instruments. Thus, the comparability of WISC-R and K-BIT scores for students with disabilities undergoing special education assessment is not known. The purpose of the present study was to address this lack of information by examining the relationship between WISC-R and K-BIT scores over a three-year period for a clinical sample of students.

METHOD

Data were collected on 44 students (26 males; 18 females) from school districts served by one educational cooperative in the Mississippi Delta region of northeast Arkansas. Subjects' mean age was 17 years, 2 months (SD = 1 year, 0 months). All of these students underwent routine three-year psychological reevaluations. Whereas 35 of these students met the criteria for a Specific Learning Disability prior to the psychological reevaluation, only 29 or these students met the criteria for a Specific Learning Disability after the psychological re-evaluation. Nine met the criteria for Mentally Retarded prior to the reevaluation and eight met the criteria for Mentally Retarded following the reevaluation. Thus, a total of seven students did not meet eligibility criteria for special education following the three-year reevaluation.

All test scores were recorded from students' special education folders from the previous evaluation and from the three-year psychological reevaluation. The three WISC-R IQs, Full Scale, Verbal, and Performance, were obtained from the previous psychological evaluation, and the K-BIT Composite, Vocabulary, and Matrices scores were obtained from the psychological reevaluation three years later.

RESULTS

Mean scores on the WISC-R and the K-BIT are shown in Table 1. Scores on the WISC-R were in the upper Borderline to the low end of the Low Average range. Scores on the K-BIT were in the low end of the Low Average range. T-tests revealed a statistically significant difference between the WISC-R and K-BIT for the Verbal IQ and Vocabulary scale, t(44) = +4.97, p [less than] .001, but not for the Performance IQ and Matrices scale, t(44) = -1.66, nor for the Full Scale IQ and Composite score, t(44) = +1.44, ps [greater than] .05. The WISC-R mean Verbal IQ was 6.2 points lower than its K-BIT counterpart, the Vocabulary scale, the WISC-R mean Performance IQ was 3.1 points higher than its K-BIT counterpart, the Matrices scale, and the Full Scale IQ mean was 1.8 points lower than the K-BIT Composite score.

Correlations between students' WISC-R and K-BIT scores are shown in Table 2. Because of restricted range of the standard deviations displayed in Table 1, all correlation coefficients reported were corrected for restricted range with the Guilford and Fruchter (1973) formula. All correlations were statistically significant at the p [less than] .01 level and were very high, especially when the three-year interval between administrations is considered. The WISC-R Full Scale IQ and K-BIT Composite were highly correlated, sharing 64% variance, as were the WISC-R Verbal IQ and K-BIT Composite scores, having 66% common variance. The WISC-R Verbal IQ was most highly related with the K-BIT Vocabulary scale and shared 62% variance. The WISC-R Performance IQ had the lowest correlations and shared only 27% variance with the K-BIT Vocabulary scale.

DISCUSSION

Similar to Prewett (1992b) but in contrast to Kaufman and Kaufman (1990) and Prewett (1992a), the WISC-R Full Scale IQ and K-BIT Composite provided surprisingly similar mean scores, with the Full Scale IQ only 1.8 points lower than the Composite score for a clinical sample of adolescents. Only the WISC-R Verbal IQ was significantly different from the K-BIT Vocabulary scale - 6.2 points lower. Thus, the total scores on both the WISC-R and K-BIT are commensurate over a three-year period.

Coinciding with Kaufman and Kaufman (1990) and Prewett (1992a, 1992b), the WISC-R and K-BIT scores were highly correlated, sharing 62 to 66% variance. This degree of common variance, as well as the evidence from the three previous studies mentioned, suggests that the WISC-R and K-BIT are, for the most part, measuring a similar construct of intelligence. Even so, these tests do not share approximately one-third of their variance in common. Thus, assessment specialists should be cautious in their use of the K-BIT in reevaluations, especially considering the differences between findings in this study involving students with disabilities and other studies involving nondisabled students.

Because of the small sample of subjects from one geographical location, readers should be cautious in generalizing these findings. The findings, however, considered in conjunction with previous studies, support the use of the K-BIT in the assessment of intelligence. These results extend previous research by focusing on the K-BIT's use in reevaluations for students with disabilities. The K-BIT was highly correlated with the WISC-R, even though three years had elapsed between test administrations, and a mean score difference was not present between total test scores. Given the savings in time, assessments specialists would be well advised to consider using the K-BIT.

Table 1

Mean Scores on the WISC-R and the K-BIT.

Test Score Mean SD

WISC-R Full Scale IQ 79.1 9.4
WISC-R Verbal IQ 78.3 9.2
WISC-R Performance IQ 83.6 10.8
K-BIT Composite 80.9 12.1
K-BIT Vocabulary 84.5 9.3
K-BIT Matrices 80.5 15.8
Table 2

Corrected Correlations of WISC-R and K-BIT Scores.

 WISC-R

Test Score Full Verbal Performance
 Scale IQ IQ IQ

K-BIT Composite +[.80.sup.*] +[.81.sup.*] +[.67.sup.*]
K-BIT Vocabulary +[.73.sup.*] +[.79.sup.*] +[.52.sup.*]
K-BIT Matrices +[.85.sup.*] +[.82.sup.*] +[.74.sup.*]

REFERENCES

Bracken, B. (1988). Ten psychometric reasons why similar tests produce dissimilar results. Journal of School Psychology, 26, 155-166.

Guilford, J., & Fruchter, B. (1973). Fundamental statistics in psychology and education. New York: McGraw Hill.

Kaufman, A., & Kaufman, N. (1990). Kaufman Brief Intelligence Test. Circle Pines, MN: American Guidance Service.

Prewett, P. (1992a). The relationship between the Kaufman Brief Intelligence Test (K-BIT) and the WISC-R with incarcerated juvenile delinquents. Educational and Psychological Measurement, 52, 977-982.

Prewett, P. (1992b). The relationship between the Kaufman Brief Intelligence Test (K-BIT) and the WISC-R with referred students. Psychology in the Schools, 29, 25-27.

Reschly, D., & Wilson, M. (1990). Cognitive processing versus traditional intelligence: Diagnostic utility, intervention implications, and treatment validity. School Psychology Review, 19, 443-458.

USDHEW. (1977). Education of handicapped children: Implementation of Part B of the Education of the Handicapped Act. Federal Register, 42, 42474-42518.

Wechsler, D. (1974). Wechsler Intelligence Scale for Children-Revised. San Antonio: The Psychological Corporation.

Linda Speer Graham, an Ed.S. in Counselor Education, is a School Psychology Specialist and Educational Examiner with the Craighead County Special Education Cooperative.

Jack Bower, an Ed.S. in Counselor Education, is a School Psychology Specialist and Educational Examiner with the Craighead County Special Education Cooperative.