Collaborating with small-scale farmers can help develop research agendas that accurately reflect farmer interests while exploiting researcher and farmer knowledge. In this context, participatory research can help farmers develop tools to improve bean yield, which can have a significant impact on development in rural areas. Data from the Honduran government indicate that annual national bean yields are around 700 kg/ha, while experiment station data shows that bean yields can reach over 2500 kg/ha, especially when using the Pan-American School of Agriculture's improved varieties, such as Amadeus-77 and Carrizalito. However, these maximum yields are based largely on data from experimental trials done in research plots, and not in production systems that are controlled by farmers whose management is based on their own local situations. In order to examine which farmer management practices have the most influence on yield, we used a participatory on-farm research model that incorporated farmers at all levels. The experimental model was based on "yield gap" analysis. The "yield gap" is the difference in yield between what is produced on experiment stations and what is produced in farmer fields, and can be divided into two groups. Yield Gap I encompasses non-transferable technology and differences in the environment, that is, factors out of the control of the farmers. Yield Gap II includes biological constraints (variety, weeds, disease, pests, and soil fertility) and socioeconomic constraints (costs, credit availability, tradition, knowledge, and input availability). This experiment looked at the biological and social constrains contained in Yield Gap II. The results of research from 2003 and 2004 will be presented.