In Honduras, dry beans are the second most important staple crop, next to maize, in terms of both production and consumption. During the past decade, agricultural scientists in Honduras, in collaboration with Bean/Cowpea CRSP scientists, have released numerous new varieties, developed improved bean-management practices, and actively worked with governmental agencies, NGOs, and farmer groups to ensure that these improved technologies are widely available to farmers. However, data on national bean production in Honduras fail to demonstrate that agricultural research has had an impact on bean production, yields, and area harvested. Given the situation, profitability analysis represents an alternative approach for assessing the farm-level impact of research. This study analyzes the record keeping data collected from Honduran bean farmers in the main bean-growing regions during the period 1998-2000. The study assesses cost and pattern of input and labor use, and analyzes the profitability of bean production for farmers growing traditional and improved bean varieties. Further, the study identifies ways to improve record keeping studies to reduce the cost of future data collection. The analysis showed that among the sample of farmers included in the record keeping surveys, farmers growing modern varieties had higher average yields and earned higher profits or suffered less loss than the farmers growing traditional varieties. However, the difference in yield for the traditional and modern farmers was statistically not significant (at 5% significance level) for three out of five of the data sets. The sensitivity analyses on enterprise gross margin showed that for traditional farmers, gross margins were more sensitive to yield and price changes than for modern farmers. The study found that none of the farmers in the sample completely followed the recommended practices for bean production and that the major share of the total production cost consisted of labor cost. It recommends that in the future, efforts to assess profitability should utilize a single-round (post-harvest) or a two-round (mid-season and post-harvest) survey and the sample size should be increased to at least 26 farmers for each farmer group (i.e., adopters and non- adopters of improved bean varieties). Furthermore, it is recommended that the standard labor parameters--which were computed from the pooled data sets--should be used to estimate total labor cost and labor cost by type of farming operations. The study suggests that in the future, possible farmers to be included in the survey should be pre-screened to insure that all "traditional" and all "modern" farmers are relatively homogenous with respect to inputs used and the sample should include only farmers with a bean area of 0.50 hectare or more. Finally, additional research is needed to better understand what factors are responsible for the high variability in farmers yields, why few farmers follow the recommended bean production practices, whether or not these recommended practices are appropriate for limited-resource farmers who grow beans in marginal environments, and the potential for identifying labor-saving technologies that are appropriate for small-scale bean formers in Honduras.