This study examined the time-course of treatment response to haloperidol in chronic schizophrenia. Furthermore the predictive value of baseline psychopathology and early therapeutic changes for the identification of the eventual treatment outcome was examined. After a two-week drug-free period forty-three chronic schizophrenic patients were treated with haloperidol for five weeks. Psychopathology was assessed on the last drug-free day and on the third and eighth day from the initiation of treatment, and then at weekly intervals. At the end of the study based on a priori criteria patients were classified as responders or non-responders to haloperidol. Seventeen patients met criteria for treatment response at the end of five weeks of treatment, while 26 did not. Already by the third day of treatment, in the responders there was a significant decrease in total BPRS and in the subscales scores for psychosis, tension and anergia, but not for hostility-suspiciousness and depression. These decreases represented approximately half of the eventual improvement obtained by the end of the study. Discriminant function analysis showed that severity of symptoms at baseline and improvement by day 3 correctly classified overall outcome in 72% of the cases.