Welcome to the International Institute of Tropical Agriculture Research Repository
Representativeness of farmers and sites in onfarm trials
Abstract/Description
In the farming systems approach to technology development, trials are carried out on a limited number of farmers’ fields. Results of these trials are extrapolated to the entire target area and used either to identify directions for further research or to develop recommendations for farmers. Extrapolation is obviously not valid if farmers and fields are not representatives of the target area. In practice however, it is difficult to select a sample of farmers according to predetermined criteria. Researchers are constrained by farmers’ misgivings about the efficacy of the technologies being tested. As a result, trials are carried out on fields of farmers who volunteer or are persuaded to participate. This paper examines the type of bias that is to occur in this situation and reasons for its occurrence. This is illustrated with data from on-farm trials in Southwest Nigeria. The interpretation of data in situations where bias exists is discussed, and procedures which may reduce bias in future trials are presented.