Ethical considerations

The study received ethics approval from the South African Medical Research Council Ethics Committee. Permission to conduct the study in Bougainville was also granted by the Chief Administrator on behalf of the Autonomous Bougainville Government. Participants (men and women) were invited to participate and were told they were free to decline and that there would be no repercussions from doing so, and that they may withdraw at any stage, skip any question and that there will be no direct benefits for them from participation. All participants who volunteered to participate gave written informed consent.

The UNDP and National Statistics Office staff conducted community mobilisation in the selected census units, first presenting the study to Chiefs in all the selected villages. In keeping with the WHO (2001) guidelines on “Ethical and Safety Recommendations for Domestic Violence Research”, the study was presented to the communities and prospective participants as a ‘Family, Health and Safety Study’ and not as a survey on gender-based violence.

Anonymity was imperative because of the sensitive nature of some of the questions. Anonymity of the participants’ responses was ensured by the use of Personal Data Assistants (PDAs) to collect data from participants. Most of the questionnaire was interviewer administered (all for women), but the section on violence perpetration and conflict was self-administered with responses directly entered on the PDA. This enhanced confidentiality. Although literacy levels were low, we tested PDA use with men in villages outside Buka and found it acceptable. Fieldworkers were present to assist in all self-completed sections if needed and did not report problems. Further information on ethical issues can be found elsewhere [32].