The problems of improvising with inadequate data are not new to most researchers, but this paper treats these problems in the context of microeconomic research in an underdeveloped area. The paper presents the detailed procedures used to process primary data secured from a sample of farm families in the State of Minas Gerais, Brazil. Essentially a descriptive case study, the paper highlights the difficulties of estimation and imputation encountered in the preparation of a series of indexes of agricultural output and input. The results are particularly designed for researchers who plan to undertake studies of agricultural economic problems in underdeveloped areas.