Disclosure bias for group versus individual reporting of violence amongst conflictaffected adolescent girls in DRC and Ethiopia

Methodologies to measure gender-based violence (GBV) have received inadequate attention, especially in humanitarian contexts where vulnerabilities to violence are exacerbated. This paper compares the results from individual audio computer-assisted self-administered (ACASI) survey interviews with results from participatory social mapping activities, employed with the same sample in two different post-conflict contexts. Eighty-seven internally displaced adolescent girls from the Democratic Republic of the Congo and 78 Sudanese girls living in Ethiopian refugee camps were interviewed using the two methodologies. Results revealed that the group-based qualitative method elicited narratives of violence focusing on events perpetrated by strangers or members of the community more distantly connected to girls. In contrast, ACASI interviews revealed violence predominantly perpetrated by family members and intimate partners. These findings suggest that group-based methods of information gathering frequently used in the field may be more susceptible to socially accepted narratives. Specifically, our findings suggest group-based methods may produce results showing that sexual violence perpetrated by strangers (e.g., from armed groups in the conflict) is more prevalent than violence perpetrated by family and intimate partners. To the extent this finding is true, it may lead to a skewed perception that adolescent GBV involving strangers is a more pressing issue than intimate partner and familybased sexual violence, when in fact, both are of great concern.