Aspects of RRA are discussed which suggest how a divergence might occur between its social and scientific value, as a positivist exercise. It often serves only a legitimating function for policies already confirmed by its sponsors; the prior existence of relevant organised knowledge and the extent of formal method are also determinants of the status of the appraisal. The use and shortcomings of indicators are discussed with reference to the example of social relations within an unsupervised credit strategy. A case study of a rapid appraisal in one village in Comilla District in Bangladesh is described.