Weight Calculations for Panel Surveys with Subsampling and Split-off Tracking

Abstract

The Living Standards Measurement Study—Integrated Surveys on Agriculture project collects agricultural and livelihood data in seven countries in Sub-Saharan Africa. To maintain representativeness as much as possible over multiple rounds of data collection, a subsample of households is selected to have members that have left the household tracked and interviewed in their new location with their new household members. Since the subsampling occurs at the level of the household but tracking occurs at the level of the individual, a number of issues arise with the correct calculation for the subsampling and attrition corrections. This article is based on the panel weight calculations for the initial rounds of the Integrated Surveys on Agriculture surveys in Uganda and Tanzania, and describes the methodology used for calculating the weight components related to subsampling, tracking, and attrition, as well as the criteria used for trimming and post-stratification. It also addresses complications resulting from members previously classified as having attrited from the sample returning in later rounds.