My talk at this conference in honor of Lars Lyberg on 1 Dec: There are two steps of sampling: design and analysis. Analysis should respect design (for example, accounting for stratification and clustering) and design should anticipate analysis (for example, … Continue reading …