NotesFAQContact Us
Search Tips
ERIC Number: ED222529
Record Type: Non-Journal
Publication Date: 1982
Pages: 20
Abstractor: N/A
Reference Count: N/A
An Investigation of WISC-R Item Bias with Black, Chicano, Native American Papago, and White Children: Implications for Nondiscriminatory Assessment.
Reschly, Daniel J.; Ross-Reynolds, Jane
Two psychometric approaches for the investigation of item bias were employed to examine the performance of four sociocultural groups on six subtests of the Wechsler Intelligence Scale for Children - Revised (WISC-R). The sample of 950 students was from Pima County, Arizona. One approach involved the comparison of the psychometric indices of internal consistency reliability, which were reasonably high and comparable for the four groups; the rank order correlations of item difficulty, which were quite high and comparable for the four groups; and the rank order correlations of the differences in the difficulty of adjacent items, which were less comparable for Native American Papagos suggesting possible items by group interactions. A second approach involved the utilization of transformed item difficulties and a quantitative method of outlier analysis to identify specific items as biased (i.e., as manifesting an item by group interaction). Results were not interpreted as evidence of cultural bias, however, since the possible confounding effects of socioeconomic status and overall ability were not controlled. Further, the items identified did not appear to reflect aspects of the assumption of "interference" implicit in the notion of cultural bias. (Author/PN)
Publication Type: Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Identifiers - Assessments and Surveys: Wechsler Intelligence Scale for Children