Publications & Resources

Construct Validation of Mathematics Achievement: Evidence From Interview Procedures

Jan 1999

Haggai Kupermintz, Vi-Nhuan Le, and Richard E. Snow

This study investigated the validity of measures derived from a large-scale multiple-choice achievement test in mathematics, using evidence from introspective think-aloud protocols of students as they attempted test items. In a small-scale study of 21 local high school students, we sought to identify and describe cognitive processes underlying performance on test items, and to examine their utility in supporting validity claims about the achievement dimensions tapped by the test. We examined differences and similarities in solution strategies and sources of knowledge used to solve items representing five achievement dimensions. The results provided further evidence for the plausibility of interpretations of the dimensions derived from a large-scale factor analysis and support the conclusion that 12th-grade mathematics achievement is factorially and cognitively complex. Test scores that do not capture such complexity may mask important achievement information.

Kupermintz, H., Le, V.- N., & Snow, R. E. (1999). Construct validation of mathematics achievement: Evidence from interview procedures (CSE Report 493). Los Angeles: University of California, Los Angeles, National Center for Research on Evaluation, Standards, and Student Testing (CRESST).