An assessment of the Young Lives sampling approach in Peru

Summary: This paper outlines the sampling methodology used by Young Lives in Peru and how it differs from that deployed by the teams in the other study countries. The Peru team selected study sites with a pro-poor bias by randomly choosing clusters of equal population from a sub-sample which excluded districts located in the top five per cent of the official poverty map developed in 2000 by the Fondo Nacional de Cooperación para el Desarrollo (FONCODES, the National Fund for Development and Social Compensation). The paper assesses this sampling methodology by comparing baseline data collected in 2002 with data from two nationally representative surveys and from the Census 2005. Examination of the common variables in the different surveys – area of residence and access to electricity and drinking water – indicates that the Young Lives sample includes households with better access to basic services and thus includes noticeable biases. The team then used post-stratification, a technique used in survey analysis to incorporate the population distribution of important characteristics into survey estimates. This reduced many differences between the samples and caused them to be less statistically significant