This study examines four drug combinations (midazolam, midazolam-midazolam, fentanyl-midazolam, and fentanyl-midazolam-methohexital) in a placebo-controlled double-blind clinical trial of intravenous sedation. It tests the hypothesis that there is no difference between the anxiolytic effect of the four combinations when compared with a saline placebo. Subjects were 207 mildly anxious young adults having their third molars removed. Cognitive measures of anxiety increased from preoperative levels in the placebo and both midazolam groups (P < .05). The anxiety response remained the same in the fentanyl-midazolam and fentanyl-midazolam-methohexital groups (P > .05). The level of successful anxiolysis ranged from 24% in the placebo group to 74% in the barbiturate group. Using the log likelihood method, comparisons suggest that the drug groups (from midazolam alone to the methohexital combination) have increasingly positive anxiolytic effects even when controlling for the effects of dental fear and intraoperative pain. The fentanyl-midazolam group is 8.1 and the methohexital group is 9.0 times more likely to have had a favorable outcome than the placebo group. Additional analyses of behavioral measures of anxiety yielded parallel results. Global evaluations after surgery were related to the success of anxiolysis for subjects in the active drug conditions (P < .05).