Compared 2 methods for selecting national samples of Blacks as part of the 1982 General Social Survey (GSS)–area probability sampling with the selection probability for an area proportional to the size of its Black population and sampling extra dwelling units in areas selected for the regular GSS. Findings indicate that the 1st sample proved less expensive to interview and required fewer screening interviews to locate households with Black members than that of the 2nd method; however, the 2nd sample exhibited a smaller design effect. Factors in the decision to use one method or the other to sample rare subgroups are discussed. (7 ref) (PsycINFO Database Record (c) 2007 APA, all rights reserved)