Multistage Clustering: An Efficient Technique in Socioeconomic Field Experiments

A multistage clustering methodology to partition large developing societies into a collection of communities with distinct characteristics is introduced. The methodology is based on multiple attributes expressing the socioeconomic status. In the context of our approach, no index of development is needed. An excellent application of the developed methodology lies in investigating the telecommunication needs for rural development in Egypt. The merits of our approach are reflected in overcoming the data inavailability on the village level and providing for an appropriate size of communities for field experimentation and analysis. An additional benefit is that the intricate sampling problem is indirectly solved.