Can transferring policing power to community members improve legal protection in Papua New Guinea?
This is the title of new research being undertaken in Bougainville by Jasper Cooper from Columbia University. Details can be found here.
It’s not often that we hear about randomised control trials (RCTs) being conducted in PNG – so please excuse my slightly elevated level of excitement. See a previous blog on the subject.
RCTs work by randomly selecting units of analysis (in this case 40 villages in Bougainville) into those that receive treatment (the treatment group) and those that don’t (the control group). This randomisation ensures that we can expect that there are no differences across the two groups except for the treatment – any differences in average outcome measures across the two groups can then be attributed to treatment. Impact evaluation is not an easy thing to do and this is why RCTs are often referred to as the “gold standard”. That being said, there are a number of potential hazards inherent in RCTs, which hopefully this research will address.
It’s also exciting that the research is in the law and justice sector, which is not known for RCTs. In fact, a search of the 3ie Impact Evaluation Repository for “police” or “policing” turns up only 12 papers (out of a total of 4,260), of which only a handful of these are based on RCT methods.
The proposed treatment is for respected and locally accountable individuals to be trained to work as unarmed uniformed officers in the villages they are from as full-time police, conducting arrests, dispelling tension, investigating crimes, and mediating conflicts. Around 2,000 individuals across treatment and control villages will then be surveyed to examine the impact of the program on the quality of local policing, such as the reporting of crime, and changes in attitudes towards the state and other behaviours, such as domestic violence. Hopefully there will be positive effects and, even if there aren’t, at least we will know that money and effort is better spent elsewhere or in some other form.