Technical Note No. 3. An Assessment of the Young Lives Sampling Approach in Peru.
Young Lives is a longitudinal research project investigating the changing nature of childhood poverty. The study is tracking the development of 12,000 children in Ethiopia, Peru, India (Andhra Pradesh) and Vietnam through qualitative and quantitative research over a 15-year period. Since 2002 it has been following two cohorts in each study country.
The objectives of this report are to describe the first round sample and the sampling design of Young Lives in Peru, and to derive appropriate sampling weights needed to use the data. In addition, by comparing Young Lives with nationally representative surveys that were carried out at the same time that the Young Lives sample was collected, we provide an assessment of potential biases in the Young Lives data. Finally, the comparison between the national census data and Young Lives allows us to suggest potential poststratification weights that may be used to adjust Young Lives sample estimates to a known population at national or regional level.
The report examines how appropriate it is to use Young Lives sample averages without considering the sampling design. The finding of this analysis might also be applicable to the three other Young Lives countries. The comparison of Young Lives data to other datasets allows us to: (i) make Young Lives comparable with national surveys carried out in the country at specific periods; (ii) identify and characterise potential biases in the Young Lives sample; and (iii) evaluate scope for adjusting Young Lives sample estimates to known population figures through post-stratification.
This report is structured as follows. In section 2, we review the sampling strategy used in the four countries to select the Young Lives samples. Then we specify the strategy used in Peru. In section 3, we compare simple, weight-adjusted averages of Young Lives data with data from the Peruvian LSMS (ENAHO 2001) from the Instituto Nacional de Estadística e Informática (INEI, National Institute for Statistics and Informatics) (INEI 2001a) and the Peruvian Demographic and Health Survey 2000 (DHS 2000) (INEI 2001b). The comparison allows us to assess biases in the Young Lives sample. In section 4, we compare data from Young Lives and the DHS 2000 with data from the Peruvian Household and Population Census 2005 (INEI 2006). We use a post-stratification strategy in the comparisons that may be used to adjust Young Lives sampling averages. Finally, section 5 summarises the results and defines the methodological steps needed to do similar adjustments in the other Young Lives countries.
Young Lives, Department of International Development at the University of Oxford, UK. 37 pp.