ABSTRACT
The realization that subject selection criteria significantly influence results when investigating socially stigmatized populations is one of the earliest and most frequently documented lessons learned by sexologists. Examples of the biasing influence of subject selection are legendary, including the early discovery that recruitment venue could determine outcomes, such as age of coming out, number of gay friends, and self-esteem level (Harry, 1986), and, more recently, that selecting gay youth from support groups (Savin-Williams and Ream, 2003) or with sex-atypical behavior (McDaniel, Purcell, and D'Augelli, 2001) escalates levels of reported suicidality. Reviewing problems inherent in sampling “homosexuality,” Sandfort (1997) argued that “findings in a specific study depend heavily on the definition and operationalization of homosexuality adopted, and on the way the sample has been put together” (p. 261). Yet, sample selection remains one of the “unresolved issues in scientific sexology” reviewed by McConaghy (1999) several years ago in this Journal.
