The last decade has been marked by increasing attention to rigor in the evaluation of interventions that seek to promote the mental health and psychosocial well-being of children in crisis contexts. One of the key markers of such rigor has been the increased adoption of strong quasi-experimental designs, where children receiving an intervention are compared to children not receiving the intervention. Although usually not randomly assigned to such conditions, children in the ‘intervention’ and ‘comparison’ groups are generally assumed to have had similar experiences other than in relation to exposure to the intervention. The use of such designs—when planned and implemented on a sound ethical basis—can significantly strengthen the capacity to make appropriate attribution of any changes observed by exploring the counterfactual case: what happens when there is no intervention? This question is usually supplementary to questions about the outcomes observed in children receiving the services being evaluated and crucial to inference regarding such outcomes. However, this paper argues that data from comparison groups has major value in its own right. Indeed, it represents a major untapped source of reflection on processes of resilience in humanitarian contexts. We use as a foundation for our analysis 3 studies completed over the last decade which examined the impact of protective and psychosocial interventions for war-affected children in Sierra Leone and Uganda. The interventions considered include programs fostering reintegration of formerly abducted children, prompting structured activities in schools, and establishing child-friendly spaces in refugee settlements. In each case, however, our focus is not on the group that received greatest attention in the original reports—the children receiving the intervention—but on those that did not. Analysis indicates the powerful forces which promote recovery in situations of conflict and the need for interventions to be more mindful that their core function is to bolster such engagement and not seek to drive recovery.