Depressed caregivers who present for parenting assistance often display excess difficulties with maintaining positive parent-child interactions and report that they cannot manage their children's problem behaviours. In addition to this, they often report other life stressors such as marital distress, lack of social support and/or socioeconomic disadvantage. This confluence of problems means that engaging depressed caregivers in parenting services can be challenging and depression is believed to impede successful intervention outcomes. For example, research has shown that depressed participants are at increased risk of intervention dropout and that they more often fail to maintain positive parenting behaviours (Assemany & McIntosh, 2002; Forehand, Furey & McMahon, 1984). However, others have suggested that engagement in parenting interventions in order to improve parent-child relationships may provide additional benefits such as reducing caregiver stress and depressive symptoms (Sameroff, 2004). In this randomised controlled trial of Parent-Child Interaction Therapy (PCIT), we assessed depression using three methods -- an interview, a self-report questionnaire and observation. We anticipated that nonattendance and attrition would be higher in depressed compared to nondepressed caregivers. In addition, those who attended 12 weeks of treatment (n = 68) were compared to those on a supported waitlist (n = 27); we expected that caregivers receiving PCIT would have greater declines in depressive symptoms than those on the waitlist. Participants were female caregivers (age M = 34, SD = 8.9) and their young children (ages 3 to 7). Caregivers were at risk of child maltreatment based on a child maltreatment inventory and reported that their children had clinical levels of externalising symptoms. Survival analysis showed that attrition was similar to previous studies of PCIT and there was no significant difference in attrition rate when depressed and nondepressed caregivers were compared. Measures of attendance such as the number of missed appointments also did not differ between groups. Regarding parenting outcomes, treatment participants showed greater improvements in observed interactions with their children than those on waitlist. Yet, the anticipated difference between depressed and nondepressed caregivers was not found; groups did not differ when we compared observed interactions with children prior to treatment and during treatment, with the exceptions of reflections/descriptions and negative talk. When depression was compared, it declined similarly and rapidly for caregivers regardless of whether they were receiving PCIT or were on the waitlist. Caregiver depression does not correlate with attendance and length of PCIT or observed parent-child interactions when participants are female caregivers with high risk of maltreatment and children with behaviour problems. In summary, study findings suggest that PCIT is an effective intervention for improving the observed parenting skills of both depressed and nondepressed caregivers with young children, but PCIT is not directly implicated in reducing caregivers' depressive symptoms.