Contends that the study of group influences on risk-taking has been overly dependent on a single laboratory instrument, the Choice Dilemmas Questionnaire. A gambling task, blackjack, which avoided various methodological and analytic problems, was employed in the present study to re-examine the group shift to risk phenomenon and to investigate factors overlooked in the past. Ss were 72 male undergraduates. Results demonstrated group-produced shifts to risk. A moderate increase over time in the level of risk-taking for individuals also was found, and the effect of outcomes on individual and group risk-taking was assessed. Contrary to previous work, no differences were found between groups with discussion and groups without discussion. A 2-fold process was postulated to explain the findings, involving both intraindividual and interindividual mechanisms. These mechanisms entail both personal and social values. (17 ref) (PsycINFO Database Record (c) 2016 APA, all rights reserved)