To the Editor.—
Choosing an appropriate control group is difficult in the experimental design of studies of patients with schizophrenia. Many investigators choose to match patient and control groups with respect to education as well as age and gender. The dilemma posed by the decision to match for education is highlighted by the recent report by Goldberg et al1 In pairs of identical twins discordant for schizophrenia with the same genotype and socioeconomic background of origin, the 16 ill twins completed an average of almost 2 years less education than their unaffected cotwins (mean for affected twins, 13.8 years; mean for unaffected twins, 15.6 years). This difference reflects the discrepancy between educational potential and actual attainment in individuals with schizophrenia, presumably a consequence of the disease process. The twin data offer additional supportive evidence for the hypothesized downward drift in social class associated with schizophrenia.2,3To the extent