A discrepancy has emerged in the literature suggesting statistically significant and clinically meaningful gender differences in IQ test performance for clinical samples, but not for normal samples. Wechsler (1958) reported no significant Wechsler Adult Intelligence Scale (WAIS;Wechsler, 1955) IQ differences between males and females of the normative sample. Turner and Willerman (1977), testing a sample of adoptive parents on the WAIS, obtained a significant but meaningless Rill Scale (FS) IQ difference of 2.6 points in favour of males. No significant Verbal (V) or Performance (P) IQ differences emerged.However, the literature addressing gender IQ differences among clinic samples consistently indicates that there are IQ differences between males and females favouring the former. L' Abate (1962) found that males of a psychiatric inpatient and outpatient clinic sample scored approximately 7 IQ points higher than did the females across WAIS V, P, and FS measures. Working with a similar sample, Boor (1975) found that males obtained significantly higher scores than females by approximately 6, 3 and 5 points on V, P, and FS IQs, respectively. In a combined sample representing two psychiatric facilities, Page and Steffy (1984) discovered 6-point WAIS V and P IQ differences favouring males. FS IQs were not reported.A similar situation exists in the area of children's intelligence testing. Analyses of the Wechsler Intelligence Scale for Children (WISC; Wechsler, 1949) and the revised WISC (WISC-R; Wechsler, 1974) normative data have revealed significant differences on V and FS IQs, with males in both samples outperforming females (Kaufman & Doppelt, 1976;Seashore, Wesman & Doppelt, 1950). Despite their statistical significance, however, the discrepancies were small (varying between 1.7 and 2.5 IQ points) and substantively meaningless. Further study of normal samples with the WISC-R has failed to reveal statistically significant differences by gender (Hanley, 1978;Munford & Munoz, 1980;Wersh & Briere, 1981). Likewise, a study of children undergoing outpatient treatment revealed no significant IQ differences (Page & Steffy, 1984).Nevertheless, in another investigation of youngsters referred to child and family clinics (Dash, Dennis, Mueller, Mancini, Snart& Short, 1983), significant differences of 13, 7 and 11 IQ points emerged across V, P, and FS measures, respectively. All discrepancies favoured males.This bias indicates that clinic populations may be misrepresented by the Send reprint requests to L.