In Africa, there have been successes in cassava research in terms of the development of production technologies, particularly improved varieties with high yield potential. The study addresses the question of whether and to what extent adoption of improved cassava varieties has led to rural poverty reduction in four African countries, namely Tanzania, Democratic Republic of Congo, Sierra Leone and Zambia. Data for the study come from a household survey conducted in the above-mentioned countries through a multinational-CGIAR support to agricultural research for development of strategic crops (SARD-SC) project in Africa. Given the observational nature of the data, a parametric approach (endogenous switching regression model) is applied. The results indicate that the model detects selectivity bias. Accounting for the bias, we find that adoption of cassava technology has resulted in an approximately 10 percentage point reduction in the poverty rate. Given an adoption rate of 34 per cent and a 10 percentage point reduction in the poverty rate, an estimated 24,309 households (equivalent to 194,469 individuals) have managed to move out of poverty in these four countries as a result of adoption of the technology. We also find that adoption of the technology has benefitted non-poor and female-headed households, relative to poor and male-headed households. The results present important evidence in favour of promoting cassava technology in a targeted fashion as part of an effective poverty reduction and sustained agricultural growth strategy in Africa. Considering the large realised and even more pronounced potential impacts of the adoption of cassava technology on poverty reduction, it is vital that regional and global development organisations should continue supporting the existing cassava improvement programme to sustain the technology development efforts in the continent.