A Comparison of the Wide Range Achievement Test and the Peabody Individual Achievement Test for Educationally Handicapped Children

Abstract

PIAT; Dunn & Markwardt 1970) and the Wide Range Achievement Test (WRAT; Jastak, Jastak, & Bijou 1965) have been widely used in clinical and special education settings. These measures appear to overlap both in the knowledge tested and in the input-output modalities used by the child, although there are important differences in the two tests.The PIAT samples the variety of classroom activities, including math, spelling, and reading recognition and comprehension, and it contains an additional test of general information and a total test score. The PIAT requires that the child use auditory and visual perception in recognizing, matching, discriminating, reading, and naming items. Other than the PIAT-WRAT congruency reports described in this report, no content validity studies of the PIAT were found.The WRAT measures reading, arithmetic, and spelling via visual and auditory inputs, and it contains a measure of rapid symbol copying, but it does not measure reading comprehension or general information. Previous studies assessing the content validity have tended to support the WRAT's subtest construction; the test compares well with teacher ratings (Hopkins, Dobson, & Oldridge 1962) and with the California Achievement Test (Oldridge 1964), especially for children in grades 1 through 4.The purpose of the present study is to compare the PIAT and the WRAT on two 74 dimensions -content (academic knowledge tested) and information processing (inputoutput modalities required). Specifically, an attempt is made to estimate the content validity of the PIAT and WRAT with a sample of children designated by the public school system as educationally handicapped (EH).A problem with measuring content validity by correlating two tests with each other is that a third independently validated measure is necessary as a criterion to adequately validate either of the first two. However, there are two types of content analysis that can provide estimates of validity: (1) subtest convergent correlation by which tests measuring similar traits show relatively high coefficients and (2) subtest discriminant correlation by which different tests show nonsignificant correlation coefficients. These measures reflect the relative "pureness" of test content, at least as defined by the test used.Sittington (1970) compared PIAT and WRAT scores in educable mentally retarded (EMR) adolescents. In general, both convergent and discriminating validity were supported, since similar tests (in content) showed higher correlations than dissimilar tests. Ysseldyke, Sabatino, and Lamanna (1973) reported similar results using test scores of EMR children, and again convergent validity was demonstrated by high correlations of like tests, but little discriminant validity was found. Considering