ERIC Number: ED386479
Record Type: Non-Journal
Publication Date: 1992-Jan
Reference Count: N/A
Construct Validity Study of the NTE Core Battery Using Confirmatory Factor Analysis.
This study investigated the construct validity and population generalizability of the NTE Core Battery using two different data bases and several analytical designs. Part I of the study used data from the November 1982 administration of the Core Battery at the subscore level for 5,183 examinees. Part II used data from the October 1985 administration of a revised version of the battery at the item level for 13,059 examinees. Both parts of the study used confirmatory factor analysis to model the structure of the test scores in relation to the knowledge and ability (constructs) they purport to measure. In Part I a nine-factor model, one for each subtest, was tested, but factors correlated too highly to be different constructs. The simplest model found to fit the data was a three-factor model of general academic skills, mathematics, and essay factors. This model fit the data well and almost equally well for white males and females and black females. The black male population was small, and results were somewhat unstable. Additional exploration of models in Part II resulted in the conclusion that the Core Battery measures general academic skills, mathematics, and essay writing. Eighteen tables present analysis findings. Three appendixes present information on correlations and six additional tables. (SLD)
Publication Type: Reports - Evaluative
Education Level: N/A
Authoring Institution: Educational Testing Service, Princeton, NJ.