Nalven, Hofmann, and Bierbryer t2) found that psychologists gave higher estimates of effective intelligence when the child was described as Negro than when he was described as Caucasian (M IQs of 100.4 and 97.3, respectively) and when he was described as lower class than when he was described as middle class !M IQs of 101.4 and 96.3, respectively). The child's age (8 or 14 years) and sex were not significant factors in the estimates. Because the Nalven, et al. study employed only one WISC protocol, there is no way t o evaluate the extent to which the pattern of subtest scores (scatter) played a part in the psychologists' judgment^'^). The present study was designed to remedy this difficulty as well as (1) to include Mexican Americans in the study of bias; (2) to evaluate the effects of a Verbal-Performance Scale discrepancy on the estimated I&; and (3) to determine how psychologists evaluate the validity of the WISC for the testing of minority-group children.
The study also is a partial replication of the Nalven, et al. study, but leaves out sex and age as variables because these were not found to be significant factors. It is hypothesized that psychologists would give higher I Q estimates (1) to black and to Mexican-American children than to white children; (2) to lower-class than to middle-class children; and (3) to profiles with much scatter than to profiles with limited scatter. No hypothesis was formulated as to the direction of the Verbal-Performance discrepancy on the estimated I&. I n addition, it is hypothesized that psychologists would judge the WISC as more valid t o test white children than to test black and Mexican-American children.