Consistency of reports of violence from northern Rakhine state in August 2017

In August 2017, Myanmar’s Armed Forces, the Tatmadaw, launched an orchestrated attack on hundreds of Rohingya-majority villages in northern Rakhine state. This study seeks to validate the consistency of previous reports of violence against the Rohingya people in the region carried out by the Tatmadaw, Border Guard Police, and Rakhine villagers in the late summer and early fall of 2017.

Internal validation data is from a three-armed study. Data analyzed in the external triangulation was sourced through a literature review of known, publicly available surveys and interviews. Both sets of data documented instances of violence against the Rohingya people in northern Rakhine state during the late summer and early fall of 2017. Consistency was evaluated across five indicators of violence: arson, presence of mass graves, reports of sexual violence and human injuries, as well as human fatalities, across 611 locales in northern Rakhine state. Further analysis was conducted to measure consistency of reports by locale and across locales by indicator.

Overall, an internal validation of 94 hamlets found that 98% of these locales were consistent across at least four of the five indicators (80% + consistency). Arson and reports of human injuries were the most consistent indicators across locales (100% and 99% consistency, respectively) and sexual violence was the least consistent indicator, with 84% of participating locales exhibiting consistent reports of sexual violence between the qualitative and quantitative data. Similarly, an external validation of 57 locations found that 50 of the 57 locations (88%) were consistent across indicators. Arson was the most consistent across sources (96%), whereas source agreement across locations was the least consistent for reports of sexual violence (58%).

The government of Myanmar has denied involvement in the 2017 attacks on Rohingya communities in northern Rakhine state and purports that reports of the violence and destruction are overstated. However, consistent reporting from multiple sources on the same locales clearly underscores the veracity of the evidence documented, both by investigative groups and as recounted by Rohingya survivors of violence. It is our hope that this cataloging and comparison of available data, along with this study’s assessment of its consistency, will aid ongoing accountability efforts.