Do participants' reports of symptom prevalence or severity vary by interviewer gender?