Comments on voluntarism: reply to Penk and Robinowitz.

Two articles on tests of the voluntarism hypothesis by Penk and Robinowitz (1976; 1980) suggest that these investigators are convinced that voluntarism is a variable of importance in influencing self-reported psychopathology. In contrast, Sutker, Archer, and Allain (1979), without minimizing the potential effects of voluntarism per se, cautioned that its influence may be modified, if not overshadowed, by person-context characteristics across different settings and among varied groups of drug-dependent individuals (e.g., veteran/nonveteran addicts, chronic opiate/polydrug addicts) and by the consequences of not volunteering for treatment. We suggested that situational, environmental, or physiological factors associated with the street life or long-term use of illicit drugs, as well as the specific conditions of testing or addict perceptions of such, may be more critical than voluntarism in affecting selfreported psychopathology. We also commented that voluntarism may prove important in predicting successfulness of drug abuse treatment participation and outcome. Although Penk and Robinowitz (1980) turned attention to the treatment implications of voluntarism, the lengths to which they extend themselves to argue the powerfulness of this variable are surprising. Penk and Robinowitz (1980) took the Sutker et al. (1979) research to task for methodological flaws and suggested that the time had come to move forward from "preexperimental" research to use of "true experimental design." This is a vulnerable stance in view of the general type of research conducted, problems in design conceptualization, and methodological flaws in their 1980 article. First, we suggest that for Penk and Robinowitz (1980) to label their design as "true" or "experimental" is inappropriate as defined by Campbell and Stanley (1963). Penk and Robi-