Agricultural development support II (project A.D.S.-II) : midterm evaluation

Midterm evaluation covers the period 7/83-5/86 and is based on review of existing data and on site visits. The project, which for political reasons has been made independent of its original home in the Faculty of Agriculture and Veterinary Medicine, has a bright future if needed mid-course corrections are made. Achievements include designing (and testing at the departmental level) an agricultural information system based on a national agricultural statistics survey (NASS) and establishing a potentially valuable comprehensive resource inventory and evaluation system (CRIES), although CRIES is so new and complex that a systematic campaign is needed to mobilize its potential for future users. More importantly, farming systems research and extension (FSR&E), despite two major successes, has focused exclusively on on-farm, agronomic testing and has lacked input from project socioeconomists (who were placed in a separate unit). As a result, the project has not fully characterized farming systems within zones nor identified farmer problems and constraints. Further, extension efforts have not yet studied rural institutions and their effect on technology adoption. The NASS, an ambitious effort, lacks the expertise needed to assure quality data. Initial efforts have suffered from serious technical flaws. The questionnaire is written in the wrong language (French), is overcrowded, at times conceptually unclear, and demands excessive interviewer training. Pretests have been inadequate, and data verification is nowhere attempted. No indepth, topical surveys - needed to verify the range of responses and as an aid in data interpretation - are planned. Until these defects are remedied, work on the NASS should be halted; the project is in a privileged position to carry out smaller surveys with more direct bearing on FSR&E. Four major lessons were learned. (1) FSR is effective only when interdisciplinary. The project's attempt to complement on-farm agronomic testing with a separate, city-based socioeconomic survey was fundamentally flawed and largely failed. (2) The choice of four different zones in two geographically distant regions made it very difficult to establish interdisciplinary local FSR teams. (3) Lack of a full-time administrator can distract the TA team leader from primary tasks. (4) The project has let quantitative work dominate most of its socioeconomic analysis, to the neglect of the more comprehensive analysis needed for FSR&E. This resulted from the unwise attempt to build two entirely different tasks - data collection and FSR&E - into one project. Such an attempt should not be repeated. For USAID/H's response to the evaluation, see the abstract of PD-AAU-024.