OBJECTIVE: To investigate if a) the previously observed link between weight gain and better clinical response during treatment of schizophrenia has a non-pharmacological component, and b) if the magnitude of treatment-related weight gain parallels therapeutic effect size across different treatments. METHODS: We compared olanzapine (N = 187) and placebo (N = 61) during the first 6 weeks of a randomized double-blind trial in the treatment of schizophrenia. Pearson correlations assessed the association between weight change and therapeutic response as measured by BPRS total score. Weight change within treatment group was contrasted between patients who improved and those who deteriorated. Analyses were repeated controlling for treatment duration and baseline weight. RESULTS: Weight gain was significantly correlated with better therapeutic response for placebo (r = 0.36, p = 0.004) and olanzapine-treated patients (r = 0.43, p = 0.001). Improved patients gained significantly more weight than patients who deteriorated (p = 0.0001). After controlling for treatment duration and baseline weight, every 1-point improvement on BPRS was associated with 0.056 kg weight gain for olanzapine and 0.089 kg weight gain for placebo. The correlation between therapeutic effect size, per Cochrane's meta-analytical reviews, and previously published magnitude of the corresponding treatment's weight gain (clozapine, olanzapine, risperidone, haloperidol, placebo) was high and significant (r = 0.91, p = 0.03). CONCLUSIONS: An association between improvement in psychopathology and greater weight gain was observed in schizophrenia patients treated with olanzapine and placebo. Findings suggest that treatment-related weight gain has, in part, a non-pharmacological basis. Further studies are needed to evaluate the hypothesis that weight gain, regardless of treatment, is associated with improving psychopathology in schizophrenia patients.