This was created in partnership with Environment for Development .
This paper evaluates the ex-post impact of adopting improved groundnut varieties on crop income and rural poverty in rural Uganda. The study utilizes cross-sectional farm household data collected in 2006 in seven districts of Uganda. We estimated the average adoption premium using propensity score matching (PSM), poverty dominance analysis tests, and a linear regression model to check robustness of results. Poverty dominance analysis tests and linear regression estimates are based on matched observations of adopters and non-adopters obtained from the PSM. This helped us estimate the true welfare effect of technology adoption by controlling for the role of selection problem on production and adoption decisions. Furthermore, we checked covariate balancing with a standardized bias measure and sensitivity of the estimated adoption effect to unobserved selection bias, using the Rosenbaum bounds procedure. The paper computes income-based poverty measures and investigates their sensitivity to the use of different poverty lines. We found that adoption of improved groundnut technologies has a significant positive impact on crop income and poverty reduction. These results are not sensitive to unobserved selection bias; therefore, we can be confident that the estimated adoption effect indicates a pure effect of improved groundnut technology adoption.