Comment: Snowball versus Respondent-Driven Sampling

Abstract

Leo Goodman (2011) provided a useful service with his clarification of the differences among snowball sampling as originally introduced by Coleman (1958–1959) and Goodman (1961) as a means for studying the structure of social networks; snowball sampling as a convenience method for studying hard-to-reach populations (Biernacki and Waldorf 1981); and respondent-driven sampling (RDS), a sampling method with good estimability for studying hard-to-reach populations (Heckathorn 1997, 2002, 2007; Salganik and Heckathorn 2004; Volz and Heckathorn 2008). This comment offers a clarification of a related set of issues. One is confusion between the latter form of snowball sampling, and RDS. A second is confusion resulting from multiple forms of the RDS estimator that derives from the incremental manner in which the method was developed. This comment summarizes the development of the method, distinguishing among seven forms of the estimator.