A human experimental model to test Harrington's hypothesis that psychometric procedures generate systematic test bias against minorities /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1993.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | The purpose of the study was to test empirically the tenability of the hypothesis put forth by Harrington (1975) that predicted that faulty classical psychometric and sampling procedures in test construction could generate systematic bias against groups with smaller representation in the test construction sample. The empirical evidence to test the hypothesis in this study was obtained by constructing test development models under strict experimental controls that systematically examined and replicated the impact of ethnic group representation in the test construction sample on subsequent ethnic group performance on the test for new independent samples. Test data from the Texas Assessment of Academic Skills (TAAS), a criterion-referenced test battery, were used in the study. Two test construction models were developed: one with exclusive representation of each ethnic group (White, African-American, Hispanic, and Asian) in each test construction sample, and the other with differential representation of ethnic groups in each test construction sample. Test scores were constructed for new independent sample of each ethnic group. The impact of group representation on the test item selection and on the ethnic groups' performance on the test forms thus developed was examined systematically using both parametric and nonparametric statistical techniques. Each procedure and analysis was independently replicated. Very large sample sizes allowed for high power in testing the theory, as the entire 11th grade population of TAAS test takers in Texas was accessible. The findings based on the data consistently revealed that the test bias pattern as predicted by Harrington's hypothesis failed to emerge from any of the analyses. The findings lead to the conclusion that it is possibly inappropriate to generalize from the results of Harrington's animal testing model to human populations. At the theoretical level, the findings indicate that the biological theory of genetic-environment interaction, which Harrington adopted to account for the dynamics of human testing, may not be applicable in human testing situations of the kind studied here. |
|---|---|
| Item Description: | Vita. "Major subject: Educational Psychology." |
| Physical Description: | xii, 154 leaves : illustrations ; 28 cm |
| Bibliography: | Includes bibliographical references. |