This paper describes the design and methodology used to assess the impact of free primary education (FPE) policy in Nairobi, Kenya. The key outcome of the study was to assess the impact of FPE on schooling outcomes among the urban poor. The study assesses the impact of FPE by examining how two non-comparable groups responded to the introduction of the policy. Retrospective data were collected for 2000–2004, while prospective data has been collected annually since 2005. This household survey has created both cross-sectional and longitudinal data sets. This provided an opportunity to assess factors, other than lack of fees, which associate with school enrolment and retention. The strength of this paper is in demonstrating how survey design can be utilized to assess the impact of a public policy between different population groups. Based on the design and methodology used in this Nairobi case study, the paper provides lessons on how to utilize household surveys to assess policy impacts.