Two experiments demonstrated the efficiency of assessing drug reinforcement in humans by using a novel multiple-choice procedure. The distinguishing characteristic of the procedure is that it arranges intermittent reinforcement for choices between pairs of potential reinforcers. The procedure has three key operations: (1) a subject is exposed to the potential reinforcers; (2) a subject then makes two or more choices on a questionnaire; for each choice, the subject is required to choose one of two potential reinforcers (e.g. drug vs. drug choices and/or drug vs. money choices); and (3) subsequently only one of the choices, randomly selected, is reinforced. In the present experiments, two variations of the multiple-choice procedure were evaluated in twelve male drug abusers. Both experiments assessed the reinforcing effects of three drug conditions (200 and 400 mg/70 kg pentobarbital and placebo) which were presented no more often than every other day. The experiments demonstrated dose-related choice of pentobarbital over money as well as choice of a higher dose of pentobarbital over a lower dose or placebo. Orderly data were generated with a single-session exposure to each drug condition. Multiple-choice procedures should have applicability, not only to the investigation of drug reinforcement, but also to the study of non-drug reinforcement in humans.