Group discussion and prediscussion assessment of preferences in the risky shift

Abstract

For this experiment the Choice Dilemma Questionnaire risk norm items were divided into two equivalent sets (Set A and Set B). The effect of the measurement sequence found in the typical risky shift study on the risky shift was assessed by comparing the postdiscussion risk preferences of subjects in three conditions on the Set A items: subjects in the standard pretest-group discussion-posttest paradigm (A-A condition); subjects who did not express prediscussion preferences, going directly into group discussions followed by a posttest (O-A condition); and a group of subjects who expressed pretest preferences on the Set B items and then had group discussions and a posttest over both Sets A and B items (B-AB condition). The postdiscussion risk preferences of subjects in the A-A condition were significantly riskier than those of subjects in the O-A and B-AB conditions. Also, the post-discussion preferences of subjects in the O-A and B-AB conditions did not differ from the prediscussion preferences in the A-A condition. The results indicated that the pretest-discussion-posttest paradigm used in risky shift studies is, in and of itself, instrumental in producing the typically reported findings.