Psychiatry Research, 24, I I-71 Elsevier 71 Ventriculomegaly in Schizophrenia: Is the Choice of Controls Important? Sarah Raz, Naftali Raz, and Erin D. Bigler Received January 26, 1987; revised version received April 20, 1987; accepted May 5. 1987. zyxwvutsrqpon Abstract. The issue of adequate controls for the study of ventriculomegaly in schizophrenia has been explored initially by Andreasen et al. (1982), who proposed that healthy volunteers, a group with presumably smaller ventricles than medical controls, be considered the optimal comparison group in this area of research. Recently, Smith and lacono (1986) suggested that group differences may appear to be greater when controls who are patients are used, as a result of the systematic exclusion of medical patients with large though normal ventricles. We conducted a meta-analysis of 37 studies of lateral ventriculomegaly in schizophrenia to clarify this issue. The results indicate that the average effect size in studies with healthy controls does not differ significantly from that obtained in studies employing medical, neurological, or psychiatric (nonpsychotic) controls. Thus, any group without gross neurological damage or severe psychopathology may be used for establishing ventriculomegaly in schizophrenia. Key Words. Schizophrenia, computed tomography, cerebral ventricles. Since the appearance of the first computed tomography (CT) study in which ventriculomegaly in schizophrenics was documented (Johnstone et al., l976), mul- tiple replications have provided a basis for new hypotheses about the etiology of the disorder (Crow, 1980; Weinberger, 1984). Several studies (e.g., Jernigan et al., l982), however, failed to replicate this finding. Andreasen et al. (1982) were the first group of investigators to explore the relationships between type of controls and findings of ventriculomegaly in schizophrenia. They concluded that healthy vol- unteers were preferable to nonpsychiatric patients as controls in studies of CT abnormalities in schizophrenia: not only did healthy volunteers have smaller ventricle-brain ratios (VBRs) than headache controls, according to their report, but the variance also was smaller in the former group. Recently, however, Smith and lacono (1986) reached an opposite conclusion. According to these investigators, healthy volunteers have larger ventricles than “controls who are patients,” perhaps because “medical patients with normal but large ventricles are systematically Sarah Raz., M.A., is a Ph.D. candidate, Department of Psychology, University of Texas at Austin, Austin, TX 78712. Naftali Raz, Ph.D., is Assistant Professor of Psychology, University of Health Sciences/The Chicago Medical School, North Chicago, IL 60064. Erin D. Bigler, Ph.D., is Neuropsychologist, Austin Neurological Clinic, 718 West 38th St., Austin, TX 78705. (Reprint requests to Dr. N. Raz, Dept. of Psychology, University of Health Sciences/The Chicago Medical School, 3333 Green Bay Rd., N. Chicago, IL 60064, USA.) Ol65-1781/88/$03.50 @ 1988 Elsevier Scientific Publishers Ireland Ltd.