Some theoretical explanations of aversive conditioning posit specific emotional stales or "cognitive medialors" to account for behavior suppression. Although "anxiety" is widely recognised as much a cognitive mediator, few studies have attempted to obtain measures of "state anxiety" or of alternative competing hypotheses to test the model. This study probed the effects of two aversive conditioning paradigms on subjective responses to both alcohol-related and non-alcohol related stimuli to test the anxiety model against other competing alternatives. A strong social desirability response bias obscured measurement on three outcome dimensions; adjective ratings of (I) General Evaluation, (2) Dangerousness, and (3) Appetitiveness. Both paradigms [i.e. Punishment and Mixed) appeared capable of producing increased " Aversiveness" for alcohol-related stimuli; "Mixed'" paradigm in which CS-UCS contingencies were unpredictable was capable of increasing state anxiety displayed toward both test stimuli. Due to differential subject attrition, a quan-experimental design was imposed which complicated application of statistical tests, A point prevalence follow-up at 90 days suggested that the experimental groups were slightly more successful al avoiding alcohol. An observation was made about a personality dynamic of alcoholics which is thought to engender psychological reactance toward treatment interventions which do not allow opportunity for internal control.Some theoretical explanations of aversive conditioning assume the presence of specific emotional states or "cognitive mediators" to explain the suppression of deviant behavior . At present, "anxiety" is the most widely recognized of the various cognitive mediators. Its function in aversive conditioning is seen as the forming of an association between conditioned anxiety, generated by noxious stimuli, and a previously attractive stimulus such that the stimulus becomes anxiety-provoking rather than anxiety-reducing. The ultimate goal of the conditioning process is the elimination of undesirable behavior under contingent control ofthe previously attractive stimulus. Tbe process generally cited is that in aversive conditioning, noxious stimulation produces both generalized emotional arousal (anxiety) and escape-withdrawal responses whicb are incompatible with the previously on-going behavior (Bandura, 1969). Although the role * This research contributed to a doctoral dissertation submitted to the Graduate Office, Florida State University, by the senior author. Sincere thanks are extended to members of the doctorale committee and to all those who contributed to the committee, especially to Wallace Kennedy, Chairman; Elvis Smith, ARC