An exploratory investigation of ethnic differences within an industrial selection battery

This study was a comparison of the test performance of a group of ninety-one white applicants for employment in a petroleum refinery and a group of ninety-one Negro applicants, matched with respect to age and years of formal education. The employment selection battery consisted of five tests: Test of Learning Ability; Test of Non-Verbal Reasoning; Test of Mechanical Comprehension, Form BB; Test of Chemical Comprehension; and the Advanced California Achievement Test Form W, math section only. Each of the tests had cut-off scores set at approximately the 70th percentile based on company norms. Each applicant to be considered for the interviewing step in the employment process had to exceed the cut-off score on each test. Although the battery was not intended to be factorially pure, it can be said to provide a rough measure of two factors: a spatial reasoning factor (Non-Verbal Reasoning and Mechanical Comprehension) and a verbal factor (California Achievement and Chemical Comprehension). The Learning Ability test contains items measuring both factors. The data indicate that matching on age and years of formal educational attainment does not eliminate the observed differences in test performance between ethnic groups since t tests were significant at the .001 level for all tests in the battery. While a higher percentage of Negroes than whites failed each of the tests; the difference between these percentages for Negro and white was greatest in the case of the tests which appear to have a high spatial factor loading. It was also observed that the difference between the mean scores for the Negroes and for the whites decreased as education increased, but that this decrease was not to the same degree for each of the tests. A statistic, PI, was computed for each test representing the percentage increase of the white mean score over the Negro mean score. On this basis the Mechanical Comprehension and Non-Verbal Reasoning tests show the largest increase, the California Achievement and Learning Ability tests are intermediate and the Chemical Comprehension test shows the smallest increase. The major hypothesis of this study was rejected: there are significant differences between Negro and white test performance even when reported education is controlled. The second hypothesis was also rejected. It appears that tests heavily loaded with a spatial factor contribute more to the differences between Negro and white than do verbally oriented tests.