Productivity gaps of 321 groundnut farmers from Uganda and Kenya were analyzed based on data from the 2009 growing seasons, using the stochastic production frontier model of Battese and Coelli (1995). Productivity gaps associated with the use of improved versus local seed varieties, with the managerial performance of research (RF) versus non-research farmers (NRF), and between male and female farmers were examined. In Uganda and Kenya, farmers who planted improved varieties enjoyed output advantages of 143% and 58.6%, respectively, over those who planted only local varieties. Farmers had a mean technical efficiency (MTE) of 54.6% in Uganda and 54.4% in Kenya. However, no significant differences were found between the MTE of RFs and NRFs, or between male- and female-managed plots. These results suggest that productivity could be improved if high-efficiency households invest more in improved varieties and low-efficiency households make better use of their existing technology. The apparent spillover effect of the technical support received by RF on NRF suggests that farmer education has a multiplier effect. An improvement in extension service delivery could help to enhance managerial skills of both RFs and NRFs. Continued development of improved varieties will help to shift the production frontier outward.