In this discussion the author deals with the most frequent questions that come to him, as statistical consultant at BAE, relative to designing samples for social science surveys. Although he draws heavily on the recent developments in sampling theory, he presents this material in a non-mathematical language.