Joint work with: Jean-Marc Fillion (Statistics Canada), Lenka Mach (Statistics Canada), Philip T. Reiss (New York University).

Abstract: Coordinating samples is an important step of survey planning, as statistical agencies need to reduce response burden and the costs of data collection. We formulate and formalize mathematically for the first time an important sample coordination problem. To solve it, we use linear programming techniques as optimization tools and classical statistical methods for minimizing variances. We construct a discrete probability space that sustains these mathematical methods. Within this probability space we can solve other sample coordination problems.