The effects of risperidone on affective symptoms were determined by an analysis of pooled data from six double-blind trials of risperidone versus haloperidol in 1254 patients with chronic schizophrenia. Symptoms indicating mania were assessed by the Positive and Negative Syndrome Scale (PANSS) excitement and grandiosity items and by the excited cluster (excitement, hostility, uncooperativeness, and poor impulse control); anxious / depressive symptoms were assessed by the PANSS anxious / depressive cluster (somatic concern, anxiety, guilt feelings, and depression). Mean change scores from baseline to endpoint were compared in patients receiving risperidone, haloperidol or placebo by analysis of variance with factors for trial and baseline score included in the model. In all patients, change scores on excitement and grandiosity items and excited and anxious / depressive clusters were significantly greater for risperidone than for haloperidol or placebo. Dropouts due to inefficacy were less frequent with risperidone (5 of 59; 8%) than with haloperidol (7 of 38; 18%) or placebo (8 of 10; 80%). In patients with anxious / depressive symptoms at baseline (anxiety / depression cluster score > or = the median), anxiety / depression scores decreased significantly more with risperidone than with haloperidol, and symptom reduction occurred faster with risperidone. These results are consistent with previous reports and suggest that risperidone is more efficacious than haloperidol for affective symptoms in patients with schizophrenia.