skip to main content skip to footer

Aligning Score Scales for Achievement Tests in Multiple Content Areas

Author(s):
Cook, Linda L.; Eignor, Daniel R.; Burton, Elizabeth B.
Publication Year:
1990
Report Number:
RR-90-30
Source:
ETS Research Report
Document Type:
Report
Page Count:
98
Subject/Key Words:
College Board, Achievement Tests, College Board Achievement Tests, Models, Scaling

Abstract

For this study, the College Board Admissions Testing Program Achievement Tests were separated into two clusters: a language test cluster and a non-language test cluster. Tests within the non-language test cluster were then subjected to a first-stage scaling, using scores on SAT-V and SAT-M as covariates. For the language tests, a similar procedure, but one making use of semesters of study in addition to SAT-V and SAT-M scores, was implemented. Scaled scores resulting from the first-stage scaling procedure were subjected to a second-stage scaling using scores on SAT-V and SAT-M as covariates employing empirically derived reference group values based on data from a combined sample for each language and non-language Achievement Test included in the study. The results of the study indicate there is some reasonable evidence to suggest that the use of empirical values for the reference group and use of the procedure that involves scaling the test in two stages may improve the alignment of the Achievement Test scales. (98pp.)

Read More