This study addresses the questions if and for which patients expressed emotion (EE) has a predictive validity in an 8 year follow-up study and if this is dependent on the residential form, especially on patients and relatives living together. Eight years after a baseline examination (Camberwell Family Interviews [CFI], global assessment scale [GAS], symptom score) 69 of 99 patients were reexamined. The number and duration of rehospitalizations as well as the symptom and GAS scores were recorded. These was no difference between outcome of high and low EE patients with a duration of illness of less than 4.5 years and for those living with partners. Patients from high EE parents with a longer duration of illness (> 4.5 years) at the outset of the study were significantly more often and longer in hospital. They had significantly higher symptom and GAS scores at the outset of the study and after 8 years. The EE effect was equal in the first and second 4 years of the follow-up and was independent from patients and parents living together. The results emphasize the significance of the EE index as a long-term predictor for the course of schizophrenic illness. The independence of the prediction from living together is an argument against a simple causal interpretation of the connection between EE and the course of schizophrenia.