“…A study of actual long-term, abstinent sobriety--its nature, dynamics, variables; its experience and meaning to those involved--is possible, if one considers using an A.A. population. However, while any study of long-sober alcoholics may have to include members of A.A. (McBride, 1991;Pettinati et al, 1982;Vaillant, 1983), researchers must continue to be mindful of the dangers of ''self-selection'' and ''affiliation bias'' in using A.A. members for research purposes (Bebbington, 1976;Seiden, 1960). Indeed, ''long-term recovery'' itself, independent of A.A., may function as a biasing variable, since those with long-term sobriety are a select group even in A.A. (Edwards et al, 1967).…”