A randomized, double-blind controlled trial is reported comparing phenobarbital and clonazepam for the purpose of sedative-hypnotic taper in inpatients with chronic, nonmalignant pain. After receiving the Minnesota Multiphasic Personality Inventory (MMPI) and a standardized psychiatric diagnostic interview, patients' baseline sedative-hypnotic use was assessed over 48 hours. Baseline use was converted into phenobarbital or clonazepam equivalents and administered in four doses daily using a blinded liquid pain cocktail. Baseline dose was maintained for two days and then tapered by 10% per day. Over the first week of taper, differences in mean and maximum Beck Anxiety and Benzodiazepine Withdrawal scores were not significant. However, when scales 1, 3, or 8 of the MMPI were taken as covariates, differences on the Withdrawal Scale only increased to a trend level for mean scores and to a significant level for maximum scores. These findings support the superiority of benzodiazepines over barbiturates for sedative-hypnotic taper for symptoms of withdrawal but not of recurrent or rebound anxiety.