A new look at the Wonderlic Personnel Test
The purpose of this study was to scrutinize a very popular and widely used mental ability or general intelligence test, the Wonderlic Personnel Test, in a very rigorous fashion in relation to three general questions: (a) Does the Wonderlic Personnel Test favor one ethnic group over another group to any great extent? (b) If the Wonderlic Personnel Test does in face favor one ethnic group over another ethnic group, what might be the basis for this differentiation? (c) Are the claims made by the test author of the Wonderlic Personnel Test in regards to its internal characteristics valid? In order to answer the first question, a sample of White college students (N=35) and a sample of Negro college students (N=27) were administered Wonderlic Personnel Test Form I under both the standard administration time limit of 12 minutes and under an unlimited (power) administration procedure. Each group was composed of summer college students of both sexes who were under 2S years of age and who had obtained 30-90 semester college hours. A comparison of the mean score for each group indicated that the Negro group scored considerably lower (significant at the .01 level) on Wonderlic Personnel Test Form I than the White group under both the standard time limit administration did under the power administration. The difference in mean gain obtained between the two groups under the power administration was not significant, however, and thereby supported the hypothesis Negroes MH not benefit more than Whites on Wonderlic Personnel Test Form I when given additional testing time. Clearly, then the answer to the first question is: Yes, the Wonderlic Personnel Test (Form I) does favor Whites over Negroes. The second question was, 'If the Wonderlic Personnel Test does favor one ethnic group over another, what might be the basis for this phenomenon? To answer this question mean scores were computed for both the Whites and the Negroes based upon test items whose content was verbal in nature. The result of this procedure showed that the Negroes had a significantly lower (at the .01 level) mean score than dad the Whites under both the 12-minute administration and under the power- administration. This finding answered the above question and supports the hypothesis that Negroes will perform less well than Whites on the verbal items contained within Wonderlic Personnel Test (Form I) under a standard administration time limit. It appears as though the verbal items within Wonderlic Personnel Test Form I constitute a major type of item content that distinguishes between the two ethnic groups. In order to verify the internal characteristics claimed for the Wonderlic Personnel Test by its author, Wonderlic test forms I, II, and IV were administered to a sample of White summer school students (N=108). Mean interform correlation coefficients between paired test forms which took into account a noted practice effect ranged from +.68 to +.75, lower than the +.32 to +.94 range given in the Wonderlic Personnel Test Manual (1966). Correlation charts between paired test forms indicated that there was a high degree of variability among subjects taking alternate Wonderlic test forms I, II. and IV. Due to the low interform correlation coefficients obtained and the high degree of variability among subjects taking these test forms, Wonderlic Personnel Test forms I, II, and IV can not be regarded as equal, equivalent, or alternate test forms. And the statistical treatment of the data to check on per cent of difficulty for units of ten questions between test forms showed that Wonderlic Personnel Test Forms I, II, and IV are not equal in terms of per cent of difficulty for units of ten questions. In summary, the results of this study are at odds with the information found in the Wonderlic Personnel Test Manual (1966) and cast doubts on the adequacy of the interform reliability of these tests. Replication of this study in additional samples to test the generalizability of estimates of interform reliability of the Wonderlic should be performed.