Doctor of Philosophy / Department of Agricultural Economics / Dalton, Timothy J. / Essay one evaluates two farmer field schools aimed at promoting conservation agricultural practices. The field schools were conducted and offered to approximately 1/3 of all individuals surveyed in a baseline in 2010. These same farmers were resurveyed in 2012 in order to determine whether their knowledge of conservation agriculture practices had changed using a double-difference approach. The approach was also used to determine whether innate perceptions and biases against conservation agriculture have changed over time due to training in the field schools. These findings are supported with enterprise budgets of conservation practices to determine whether knowledge or on-farm economics limit adoption of conservation practices. The data showed that farmer-to-farmer communications are effective tools for raising knowledge. Essay two examines the interdependence of sustainable agricultural intensification practices (SAIPs) in order to better understand the constraints and incentives for the adoption of components and “packages” of components. The impact of accumulated knowledge score on the adoption of SAIPs was assessed using data from 168 participant and non-participant farm households that completed a survey in 2014 and 2012 from the Upper West region of Ghana. From a three-step regression, our findings show knowledge of participant household improved with evidence of knowledge spillover to non-participant. Participation, age and gender of the head of household and experience were factors impacting farm household knowledge score change on SAIPs. The study found that, knowledge score through the treatment effect impacts adoption of SAIPs which are complementary. Younger household heads and experience in farming are also found to likely impact adoption. Essay three estimates technical efficiency (TE) scores for millet and sorghum and evaluates the impact of soil and water conservation methods on TE scores. The paper also examines the sensitivity of TE scores on the distributional assumptions of the one-sided error using data from 518 and 754 farm households producing millet and sorghum respectively from a random national household survey in Niger. A Cobb-Douglas stochastic frontier model was used. The mean TE scores range from 52% to 66% and 35% to 60% respectively for adopters and non-adopters of soil and water conservation methods in millet production based on the distributional assumptions of the one-sided error. For sorghum production, the mean TE scores range from 47% to 63% and 39% to 63% respectively for adopters and non-adopters of soil and water conservation methods based on the distributional assumptions of the one-sided error. This suggests inefficiencies in the production of millet and sorghum and hence, the potential to improve output using existing technology. Adopters are relatively more efficient than non-adopters of soil and water conservation methods. The TE score differences in millet production are explained by location of household (rural), educational level and adoption of soil and water conservation. The efficiency score differences in sorghum can be explained by household size, educational level and soil and water conservation adoption. We also found TE scores are sensitive to the distributional assumptions of the one-sided error using the farm household level data.