Small ruminant production in mixed farming systems : case studies in research design

First, the paper describes two small ruminant production systems, one in Java, Indonesia, the other in the humid zone of West Africa, to highlight key features of small farm systems that should determine the way in which on-farm research is designed. Second, the experiences of small ruminant research teams working in these areas, and of other research programs working in similar farming system contexts, are used to demonstrate problems in conventional trial designs and the value of a broader view. Specifically, discussion is given to four methodological problems which derive primarily from viewing on-farm research as the trial of technologies by individual producing units. These problems concern: (1) identifying the units of production which will be the focus of research; (2) collecting data sufficient for conventional statistical analysis where small ruminants are a secondary or minor enterprise; (3) classifying producers for the purposes of sampling and trials given the variability in producers" livestock, management, and type and use of labor; and (4) designing on-farm trials to ensure that results can be replicated by farmers. A 38-item list of mainly post-1980 references is appended.