Construct and content validity bias for the universal nonverbal intelligence test for African Americans and European Americans

Author

Barzanna Anita White

Date of Award

12-1996

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Education

Major Professor

R. Steve McCallum

Committee Members

Donald Dickinson, Michael K. Smith, Lynn Cagle, Thomas W. George

Abstract

The difference in lay and professional definitions of bias are presented. The issue of bias is presented as it relates to instrument design and standardization of the Universal Nonverbal Intelligence Test (UNIT). The primary purpose of this study is to determine whether the UNIT shows evidence of content and construct validity bias. A second purpose is to explore other basic psychometric properties of the UNIT based on mathematical and statistical analysis of data of two groups: African Americans and European Americans.

Data from 83 African Americans and 581 European Americans provided evidence that the UNIT is a psychometrically sound instrument. Reliability figures are high (from .79 to 92) and there is minimal evidence of content and construct validity bias. A few biased items were identified and recommendation was made to delete those. With further development, the UNIT may be used to assess the intelligence of a variety of individuals – especially those whose verbal ability is impaired.

Recommended Citation

White, Barzanna Anita, "Construct and content validity bias for the universal nonverbal intelligence test for African Americans and European Americans. " PhD diss., University of Tennessee, 1996.
https://trace.tennessee.edu/utk_graddiss/9884