Design/Methodology/Approach: A participatory experimentation approach was followed to arrive at recommendations matching with local preferences and context. In total, 16 groups of 5 farmers were monitored during 4 years.

We monitored research strategy of the farmer groups by considering the following: (1) the type of treatments, (2) the inclusion of responsive treatments, (3) the actual responses achieved and (4) the treatments perceived optimal.

Findings: We found that the farmer groups followed a very rational, context-rooted strategy that, e.g. in its focus on straw production and the use of combinations of organic and mineral fertilizers, differed from that of the researchers.

Practical implications: Farmers often follow research strategies different from standard scientific approaches. Consequently, in participatory experimentation, involvement of farmers in defining the actual experimental design is required to deal with local preferences and context.

Theoretical implications: Outcomes of participatory experimentation are directly relevant for further outscaling of the technologies involved. In addition, insights and understanding obtained also might support upscaling in the form of designing rural development policies.

Originality/Value: Participatory experimentation processes are applied in development work for different reasons but often concentrate primarily on direct outputs. For development workers engaged in such processes, it is important to realize that actual involvement of participants in the whole process is equally important.