Intraindividual differences in verbal and numerical abilities have been related to personality characteristics and to psychopathology. R. Dana et al (1959) used extreme Verbal (V) and Quantitative (Q) SCAT scores; "control" subjects had smaller difference scores. Groups were compared on 22 MMPI scales; 14 of 176 t tests were significant and no estimation of expected number under the null hypothesis was given. Blind analysis of group profiles indicated greater subjectivity, repression, projection, and distorted thinking for extreme verbal subjects regardless of sex. These results were consistent with past research. However, percentile difference scores were used with no control for inequality of percentile units. Contrasting particular personality variables rather than entire profiles makes determination of chance expectation difficult. Percentile scores may be used with careful selection of groups. A split-plot analysis of variance design for repeated measurements employs the entire profile.