This paper presents the results of an investigation of the possible effects of methodological differences between controlled-dosage and ad lib drinking sessions on blood alcohol concentration (BAC) estimation data, and of the extent to which individual drinkers show stable error patterns over time. 72 male drivers participated in at least one experimental alcohol drinking session. In ad lib sessions, 47 subjects, with free access to drinks of their choice, were instructed to tell the experimenter when they reached their `usual' intoxication level. In controlled dosage sessions, subjects were required to drink enough alcohol within a given time to achieve given peak BACs. 32 subjects aimed for 50mg/dl, and 33 subjects aimed for 100mg/dl. At the start of each data collection trial, after drinking, subjects were asked: (1) whether they would drive at that moment; (2) what their driving ability would be; (3) how much their drinks would affect it; (4) how intoxicated they felt. Tables 1 to 3 tabulate the comparison results obtained during the three types of sessions. In ad lib drinking sessions, the subjects seemed to achieve comparable subjective status by self-dosing to different BACs. In controlled dosage studies, the subjects at 100mg/dl predominantly underestimated their BACs and the subjects at 50mg/dl mostly overestimated them.
Abstract