The study is the latest to estimate returns to education after the introduction of free primary education in 2003 in Kenya, simultaneously addressing two sources of biases due to endogeneity of schooling and earnings, and sample selection. Using the 2005-2006 Kenya Integrated Household and Budget Survey, the paper finds that (a) returns to an additional year of schooling are 14.9% for males and 13.5% for females with a continuous education variable, but the returns to females are consistently higher than males when returns are estimated by the level of education, (b) returns to education increases for higher levels of education i.e., the classical pattern of diminishing return to schooling does not hold for both males and females in Kenya, and (c) the use of joint IV-Heckman method adjust the endogeneity and sample selection biases introduced by OLS and IV.