Validity and gender issues in the use of the DISCOVER. Assessment for the identification of gifted minorities

Abstract

The purpose of this study was to investigate certain aspects of the internal structure of the DISCOVER assessment checklist (C. Maker, A. Nielson, and J. Rogers, 1994) to assess its construct validity. Other purposes included assessing percentages of identified students and gender differences through the use of the assessment. The sample consisted of 368 Navajo Indian and Mexican American participants from kindergarten, fourth, fifth, and sixth grades. The results showed low and nonsignificant inter-rating correlations, indicating high discriminant validity of the checklist. R-squared analyses revealed low percentages of variance explained, indicating low convergent validity. Items most frequently checked clustered around the underlying intelligence assessed. A pattern of more checks was found for higher ratings. Also, 22% of kindergarten, 27% of fourth- and fifth-grade, and 44% of sixth-grade participants were identified as gifted. Chi-square tests revealed no overall significant differences in the number of boys and girls identified as gifted. The results indicate sound qualities of the DISCOVER assessment and promote the use of performance-based assessments. An appendix contains checklist sample items.