The integral relationship between economic inequality and social inequality made obtaining statistically meaningful, quantitative data on per household wealth, income, and expenditure a central task of fieldwork. Fortunately, the basic demographic and social organizational characteristics of the population made obtaining and analyzing such data fairly straightforward. Sampling was facilitated because the basic census unit, the household, is also the basic unit of production and consumption. Collecting data presented no major difficulties, for household production and consumption centers on a nuclear family. Even though many households incorporate lineal or collateral kinsmen, only rarely are two married couples, with independent productive activities, domiciled together. Likewise, analyzing and comparing data was made easier as the considerable variation present in household size and composition is random with respect to social status.

There were, of course, departures from the ideal. Some unmarried youths were semi-independent, having their own sources of cash income but eating meals in their parents' households. Households of similar genealogical composition (e.g., those with an incorporated spouse's parent or child-in-law) differed in their eating arrangements. But the procedures outlined here readily accommodated such variations, and I was not faced with the problems confronting fieldworkers in societies where the boundaries separating households and delineating production units are more ambiguous.

Given the nature of the data I desired and a total community population of 112 households, a sample of approximately one-third of the households seemed to be the largest that I could comfortably handle.