This paper examines the extent to which levels of education of a wage employee account for wage difference in a selected sample of workers in Kenya, Tanzania, Uganda, Madagascar, Ghana, Niger, Guinea Conakry, Rwanda, Benin and Togo. The paper uses country specific factors to control for omitted variable bias. Using OLS and Control function estimates, the paper confirms significant wage difference accounted for by education attainment. Estimates of the models suggest a positive correlation between education and wages.The square term of schooling is highly significant indicating nonlinearity of the returns to education, such that at higher levels of education the marginal returns to schooling is greater than the marginal returns to schooling at lower levels of education. The paper shows strong evidence that at primary level of education there is wage difference of 4.82 per cent due to one year of education, while middle school graduates have 6.3 per cent wage premium and so on. Country specific attributes have a positive influence on the extent of wage differences by country. Thus, it is likely that wage differences of education are a result of self-selectivity.