INTRODUCTION Although depression is a well-established feature of schizophrenia, it is difficult to measure, because it overlaps with negative symptoms and extrapyramidal symptoms (EPS). Routinely adopted depression scales were not designed to be used in--cases of schizophrenia, and are known to perform poorly when trying to distinguish depression from other symptoms. OBJECTIVE The aim of this study was to evaluate the validity of the Brazilian version of the Calgary Depression Rating Scale for Schizophrenia (CDSS). METHOD Outpatients from four mental health units in the city of São Paulo, diagnosed as having schizophrenia by DSM-IV criteria, were evaluated by two independent raters who applied the DSM-IV depression criteria. All patients were assessed by means of the CDSS, the Positive and Negative Syndrome Scale (PANSS), and the Extrapyramidal Symptom Rating Scale (ESRS). RESULTS Eighty patients were recruited for the study. The analysis was carried out by comparing the DSM-IV criteria of depression with the CDSS scores, by means of the receiver operating characteristic (ROC) curves. The area under the ROC curve for major depression was 0.95 (SD = 0.02), and at a cut-off point of 6/7 the validity coefficients were as follows: sensibility 77%, specificity 92%, positive predictive value 67% and negative predictive value 95%. The area under the ROC curve for minor depression was 0.95 (SD = 0.02), and at a cut-off point of 4/5 the validity coefficients were as follows: sensibility 95%, specificity 88%, positive predictive value 75% and negative predictive value 98%. The correlation coefficients between the CDSS scores, the PANSS negative and positive subscale scores, and the ESRS scores were all below 0.50. CONCLUSION It can be concluded that the Brazilian version of the CDSS is a valid research tool to assess depressive episodes for stabilized patients with schizophrenia.