ON THE USE OF THE POSITIVE AND NEGATIVE SYNDROME SCALE IN RANDOMIZED CLINICAL TRIALS.