The studies reported herein investigate the feasibility of using a serial choice reaction time (SCRT) task as the basis for an alcohol ignition interlock. The two measures of performance used are the total time to complete the task plus the number of errors (TTC), and the number of response times exceeding 1.1 s duration plus the number of errors (PS). Both measures are found to increase in proportion to the square of the blood alcohol level (BAL) with the PS measure increasing more rapidly. A sequential criteria task that minimized task duration (<7 s) is proposed. This task would reject approximately 30% of the subjects at 0.10% BAL while only failing 1% when sober. SCRT tasks show promise as a basis for an ignition interlock. It was concluded that additional work is indicated.