U.S. Department of Defense


Date of this Version



Published in Lancet (2011) 378: 915–24


Background Breakdowns in the ethical conduct of soldiers towards non-combatants on the battlefi eld are of grave concern in war. Evidence-based training approaches to prevent unethical conduct are scarce. We assessed the eff ectiveness of battlefi eld-ethics training and factors associated with unethical battlefi eld conduct.

Methods The training package, based on movie vignettes and leader-led discussions, was administered 7 to 8 months into a 15-month high-intensity combat deployment in Iraq, between Dec 11, 2007, and Jan 30, 2008. Soldiers from an infantry brigade combat team (total population about 3500) were randomly selected, on the basis of company and the last four digits of each soldier’s social security number, and invited to complete an anonymous survey 3 months after completion of the training. Reports of unethical behaviour and attitudes in this sample were compared with a randomly selected pre-training sample from the same brigade. The response patterns for ethical behaviour and reporting of ethical violations were analysed with chi-square analyses. We developed two logistic regression models using self-reported unethical behaviours as dependent variables. Factors associated with unethical conduct, including combat experiences and post-traumatic stress disorder (PTSD), were assessed with validated scales.

Findings Of 500 randomly selected soldiers 421 agreed to participate in the anonymous post-training survey. A total of 397 soldiers of the same brigade completed the pre-training survey. Training was associated with significantly lower rates of unethical conduct of soldiers and greater willingness to report and address misconduct than in those before training. For example, reports of unnecessary damage or destruction of private property decreased from 13·6% (54 of 397; 95% CI 10·2–17·0) before training to 5·0% (21 of 421; 2·9–7·1) after training (percent difference –63·2%; p<0·0001), and willingness to report a unit member for mistreatment of a non-combatant increased from 36·0% (143 of 397; 31·3–40·7) to 58·9% (248 of 421; 54·2–63·6; percent difference 63·6; p<0·0001). Nearly all participants (410 [97%]) reported that training made it clear how to respond towards non-combatants. Combat frequency and intensity was the strongest predictor of unethical behaviour; PTSD was not a significant predictor of unethical behaviour after controlling for combat experiences.

Interpretation Leader-led battlefield ethics training positively influenced soldiers’ understanding of how to interact with and treat non-combatants, and reduced reports of ethical misconduct. Unethical battlefield conduct was associated with high-intensity combat but not with PTSD.