Evaluation of new technology on farms: Methodology and some results from two crop programmes at CIAT

The yield gap between experiment station and farm yields in the production of food crops in developing countries has been frequently noted and various quantitative attempts have been made to separate its components in irrigated rice (Gomez et al., 1979; IRRI, 1979; Flinn, 1980). In two food crops in Latin America one principal hypothesis of the authors for the continuation of this yield gap over time is that many successful technologies on the experiment station do not pass a set of reasonable farm level criteria. Farm testing is the logical extension of the research evaluation process once a technology has been identified on the experiment station and regionally tested for adaptation. Farm testing is an especially important component of the research process in developing countries, where communication links between farmers and researchers are weak and farmers often do not have the information or management experience to combine and modify various technology components adapting experiment station or regional trial observations to their own environments and production systems. The research problems at the farm are different from those at the experiment station in regional trials so there are important distinctions in design and analysis in the farm trials. The evaluation process developed here identified the technologies later adopted by farmers. For the unsuccesful technologies, information was provided from the farm trials to the breeders and other scientists on further design requirements. The results of the farm trials substantially modified the recommendations for farmers, which would have been arrived at utilising the results from the experiment station and/or regional trials eng