The effects of orally administered placebo, diphenhydramine, lorazepam, methocarbamol and placebo were studied in volunteers with histories of recreational substance abuse including sedative/hypnotics. Placebo, diphenhydramine (100, 200 and 400 mg), lorazepam (1 and 4 mg) and methocarbamol (2.25 and 9 g) were tested in a randomized, double-blind crossover study using 14 subjects. Psychomotor and cognitive performance and subject- and observer-rated responses were measured daily before and for 5.5 hr after drug administration. The results showed that each of the drugs exhibited a different profile of effects on the test battery. Lorazepam produced significant increases in subjects' ratings of drug effect and liking, increases in measures of sedation and impairment of psychomotor performance. Methocarbamol also produced significant increases in subjects' ratings of drug effect and liking and measures of sedation, but it produced only minor impairment of psychomotor and cognitive performance. Diphenhydramine increased subjects' and observers' ratings of drug effect and measures of sedation, but it produced less psychomotor performance impairment and liking than lorazepam. Diphenhydramine produced the most side effects. The present study clearly differentiated the behavioral and subjective profiles of diphenhydramine, lorazepam and methocarbamol. Consistent with its recognized low abuse liability, diphenhydramine produced fewer increases in measures of positive mood and more adverse effects. The considerable overlap in subjective effect measures of positive mood make further differentiation with respect to abuse liability difficult.