A design-based procedure for the analysis of experiments embedded in complex sample surveys

This paper proposes a procedure to test hypotheses about differences between sample estimates observed under alternative survey methodologies.