For many practical reasons there is no escape from expressing economic inequality in terms of households, or farming households, as in Chapter 5. This is more unfortunate than is generally recognized, since so many households in the rural tropical world do not conform to the Western stereotype of the integrated nuclear group that necessarily dictates our statistical approach, inappropriate though it may be. The main purpose of this brief chapter, which further develops some of the arguments previously made, is to look at a few of the ‘defects’ of this statistical unit in terms of the departure from the stereotype, and to suggest the need for a degree of sub-classification of households in certain circumstances.

So that our definition may relate to the varying circumstances in different regions it is best to regard the farming household as being, in vague general principle, a group of kin and affines, which eats from the same cooking pot, lives under the same roof and cultivates the same land, and is commonly based on at least one conjugal unit. But, as we shall see, there are so many variants on this norm, especially in West Africa, and so many reasons why a person's individuality (rather than his/her position as X's spouse) requires emphasis, that one often wishes that the individual adult (such as the mother with her children, the father or the married son) might be the basic statistical unit – though this would not obviate the need for sub-classification.