In a large (n = 1,827) multi-site study examining effectiveness of consumer operated service programs as an adjunct to traditional mental health services, we examined individuals' preferred term describing their status as service recipients, and we applied logistic regression to examine whether preference varied by gender, race or diagnosis. Preferred terms were client (39%), patient (22%), consumer (16%), survivor (11%), other (11%) and ex-patient (1%), varying by site. Controlling for site, preferences did not vary by gender, race, or diagnosis. The lack of consensus suggests clinicians, researchers, program administrators, and policymakers should be sensitive to individuals' preferences.