A TOTAL OF 645 SUBJECTS, 359 MALES AND 286 FEMALES, PARTICIPATED. OF THE TOTAL, 606 PARTICIPATED IN 6-PERSON MOCK JURIES. DATA FOR THE REMAINING SUBJECTS WERE NOT ANALYZED. THE EXPERIMENTAL DESIGN WAS A TWO BY THREE FACTORIAL ANALYSIS, WITH INDEPENDENT VARIABLES OF THE ASSIGNED DECISION RULE (UNANIMOUS VERSUS MAJORITY) AND DEFINITION OF REASONABLE DOUBT. FOR THE LATTER VARIABLE, SUBJECTS WERE EITHER TOLD THAT ESSENTIALLY ANY DOUBT ABOUT THE DEFENDANT'S GUILT QUALIFIED AS A REASONABLE ONE, TOLD THAT A REASONABLE DOUBT MUST BE A SUBSTANTIAL ONE, OR GIVEN NO DEFINITION OF THE CONCEPT OF REASONABLE DOUBT. DURING THE FIRST PART OF THE EXPERIMENT CONDUCTED IN A LARGE CLASSROOM, ALL SUBJECTS PARTICIPATED INDIVIDUALLY. THE SECOND PART OF THE EXPERIMENT TOOK PLACE IN SMALLER ROOMS. IN BOTH PARTS, SUBJECTS COMPLETED A QUESTIONNAIRE ON THEIR OPINIONS ABOUT THE GUILT OR INNOCENCE OF THE DEFENDANT. MOCK JURIES ASSIGNED A UNANIMITY DECISION RULE WERE SIGNIFICANTLY LESS LIKELY TO REACH A VERDICT THAN JURIES ASSIGNED A MAJORITY RULE. MINORITY MEMBERS OF THE JURIES ASSIGNED A MAJORITY DECISION RULE WERE PARTICULARLY DISSATISFIED WITH GROUP DELIBERATION. DEFINITIONAL VARIATIONS IN REASONABLE DOUBT SIGNIFICANTLY AFFECTED BOTH INDIVIDUAL AND GROUP VERDICTS. THE DATA SUGGEST THAT THE DEFINITION OF REASONABLE DOUBT DOES NOT AFFECT MOCK JUROR PERCEPTIONS OF THE WEIGHT OF EVIDENCE AGAINST THE DEFENDANT. EFFECTS OF THE INDEPENDENT VARIABLES IN THE EXPERIMENT ON THE GROUP DECISIONMAKING PROCESS ARE EXAMINED USING A MODEL-FITTING APPROACH TO THE SOCIAL DECISION SCHEME MODEL OF DAVIS. SUPPORTING TABULAR DATA ARE PROVIDED, BUT THE QUESTIONNAIRES ARE NOT INCLUDED. A LIST OF REFERENCE NOTES IS PROVIDED. (DEP)