In examining the applicability of different methods to different individuals, as well as conducting dedicated trials of specific methods there may be value in considering alternative methods of data analysis. Clinical trials of psychosocial interventions typically compare the average outcomes of individuals randomly assigned to different treatment conditions. However, because no treatment is equally effective for all recipients, the average response, as measured, eg, by clinical rating scales, is less important than determining which participants have a response. In the case of antipsychotics, eg, the average response in comparison with placebo is often statistically significant but masks individual variability in response. New statistical techniques have enabled researchers to isolate the proportion of patients (about 20%) that show a dramatic and rapid improvement in their positive symptoms. These growth mixture modeling or latent trajectory analyses have obvious applications in psychosocial trials and may facilitate the identification of subgroups of voice hearers who are responsive (and nonresponsive) to different psychological therapies.