Objective: The aim of the present study was to provide a useful and quick estimate of premorbid intelligence that can be used in the population of schizophrenia patients.Method: Regression analysis (stepwise procedure) was used to identify which of five demographic variables significantly predicted National Adult Reading Test (NART) scores in schizophrenia patients. The new regression equation was subsequently assessed in an independent patient sample. The results were compared with scores obtained for an equation developed in healthy populations.Results: Age, occupation, and education contributed significantly to the equation, which accurately predicted NART scores in the patient sample. The equation developed in healthy people, however, consistently underestimated NART scores.Conclusions: Equations developed in healthy people are not accurate predictors of NART scores. By contrast, the new equation acknowledges the educational and occupational characteristics typical of individuals with schizophrenia. The equation should be used cautiously, but can be used as a rapid estimate when it is not possible to conduct a cognitive assessment using more traditional approaches.