Fredrick Wamalwa and Justine Burns
This paper examines the effect of private schools on literacy (language) and numeracy (maths) skill acquisition among children drawn from lower primary grades in Kenya. We use a comprehensive household survey data that allows us to apply a number of econometric techniques to deal with the challenge of the endogeneity of private school choice. We begin with the OLS as a baseline model. We then estimate the village and household fixed effects (FE) models that control for unobservables at the village and household levels, respectively. We supplement the OLS and FE models with the propensity score matching (PSM) technique.
We find positive and signi cant private school effect throughout all these methodologies. However, assessing the impact of omitted variable bias on the estimated coefficient of private schools by use of recent techniques, we find that the estimated bias in household FE is quite small in magnitude relative to the bias based on other estimation techniques. Since (private) schooling decision is made at the household level, it is likely that a substantial part of the unobservable component is pertaining to the household.