Delivering reliable information on the lives of a population is dependent on taking a representative sample.
This is straight-forward when working in the USA, the UK, Japan, or Germany, when there is reliable census information with which to design sampling frames, but what happens when you are working in a country where the last census was completed in the 1970s, or where many citizens are nomadic, or because of war there are many Internally Displaced People (IDPs)? This is a scenario that is often faced by ORB, when working in conflict zones, post conflict zones, or countries that face crippling poverty or government dysfunction.
This was the situation faced in recent studies ORB conducted in Hargeisa, Somaliland and Mogadishu, Somalia. Due to the volatile situation in these countries, the population is extremely transient, and there was no reliable government information on the structure of the population. In this situation - working with previously un-researched populations - ORB designed sample strategies unrecognizable to organizations working in the developed world.
Most of the literature on sampling methodology assumes that the population being sampled live in fixed abodes, and that their housing is set out in a reasonably orderly manner so that standard randomized selection procedures can be implemented. In these Somali studies, we had to gather the views of nomads and those Mogadishians that reside in formal or informal IDP settlements. It is not possible to apply standard sampling procedures to those living in arbitrarily placed scattered tents, or even to interview these individuals in their homes due to the lack of privacy in their settlements. Thus, with the nomadic people, we managed research teams that used local intelligence to predict where they would be taking their herds to graze, so that the sample was both predictable and representative of the various nomadic groups. With IDP settlements, we first mapped the location, number and population density of the settlements, and then devised a sampling strategy that ensured the views of these people were polled in a representative way, alongside those living in more formal villages, towns and cities.
For more information on conducting research in complex environment, please email ORB International