A COMPARATIVE STUDY OF STANDARD SCORES, DISCREPANCY SCORES AND RELIABILITY ANALYSES IN THE LEARNING DISABILITIES EVALUATION PROCESS EMPLOYING THE WIDE RANGE ACHIEVEMENT TEST (1978), AND THE WIDE RANGE ACHIEVEMENT TEST-REVISED (1984)

Reviews of the literature include numerous warnings from school psychologists, measurement specialists, and others, regarding the inadequacy of the most popular instruments and procedures in the LD identification process. However, in current practice data from unreliable and/or invalid instruments and procedures have a decided impact on whether or not a child is determined eligible for LD services. Certain tests selected for LD identification are known to have inflated or deflated scores. The standard scores yielded by the WRAT-R have been noticed by clinicians as lower, in general, than scores for earlier versions of the test.;To address the questions under study, a sample of 60 referrals was selected from a pool of five school psychologists. Each subject was administered the WISC-R and both versions of the WRAT (1978 and 1984 revision) is a counterbalanced design. The Ohio discrepancy procedure was applied to each subject's standard score comparisons of WRAT and WRAT-R subtests with the WISC-R FSIQ. The reliability of each discrepancy score was further analyzed with a standard error of measurement (SEM) procedure to determine the reliability of difference scores for achievement test and intelligence scores.;Significant standard score differences for the two versions of the WRAT, were found. The most recent edition contributed a greater number of students assignable for LD services. Additionally, significantly greater numbers of subjects could be assignable when the reliability analyses (the SEM procedure) were conducted.;Results were discussed in terms of the impact of unreliable and invalid instruments and procedures on placement decisions, and the need for improved analyses by practioners of the instruments and procedures they employ in school settings.;Specific questions under study were: (1) is there a difference between the standard scores obtained on the two versions of the WRAT, in spite of the fact that the format and items are virtually identical? (2) are there differences in numbers of subjects assignable as LD by the formulation mandated by the State of Ohio for assigning students to classes for LD and a reliability analysis of difference scores?, and (3) what is the potential for overdiagnosis and underdiagnosis using the above instruments and procedures?...