ERIC Number: ED395008
Record Type: Non-Journal
Publication Date: 1988-Jun
Reference Count: N/A
Achievement Test Scaling.
Cook, Linda L.; And Others
Scaling is carried out in an effort to increase the comparability of scores obtained on different tests. This study explored the relationships between College Board Achievement Test scores and potential scaling covariates for various subgroups of the test-taking population with the goal of providing several alternatives to traditionally used scaling procedures. The alternatives vary components such as scaling covariates, scaling samples, and characteristics of the hypothetical reference group. The traditional method of scaling the Achievement Tests uses scores on the Scholastic Aptitude Test (SAT) verbal and mathematics tests as scaling covariates for all 14 achievement tests. To determine additional covariates that might be used, correlations of test scores with 17 potential covariates were examined, and stepwise regressions were carried out. As a result, the Achievement Tests were grouped into five clusters. Two alternative scaling procedures were proposed. The first consists of scaling the tests within the five independent clusters; and the second consists of a two-stage process of scaling within the clusters independently and using the results of the initial scaling to scale the tests again with SAT verbal and mathematics scores as common covariates. (Contains 9 figures, 12 tables, and 3 references.) (SLD)
Publication Type: Reports - Evaluative
Education Level: N/A
Authoring Institution: Educational Testing Service, Princeton, NJ.
Identifiers - Assessments and Surveys: College Board Achievement Tests; SAT (College Admission Test)