Ethnolinguistic Background and Enrollment in Primary Education: Evidence from Kenya

In Kenya, educational enrollment rates increased significantly for both girls and boys after 2003, when primary education became free of charge. Unfortunately, approximately one million school-aged children are still not enrolled in school. Earlier literature provides empirical evidence that educational opportunities differ among children, due to poverty, gender, rural area of residence and disability. Our paper con-tributes to the literature by providing empirical evidence of the importance of children's ethnolinguistic background for their probability of being in school. Estimates from a three-level random intercept probit model using data from the Kenya Integrated Household Budget Survey 2005/06 reveal that Somali and Maasai children are least likely to be in school. A separate analysis by child's gender shows that compared to Kikuyu children both girls and boys from the Somali and Maasai groups, but also Mijikenda and Swahili girls, have a lower probability to be in school. This might be an indication that gender norms are stronger in these groups.