IN THE KESSLER EXPERIMENT, COLLEGE STUDENTS WERE RANDOMLY ASSIGNED TO 6-MEMBER OR 12-MEMBER JURIES. AFTER VIEWING THE SAME VIDEOTAPED TRIAL, EACH JUROR PRIVATELY RECORDED HIS OR HER OWN PREDELIBERATION VERDICT. RESPECTIVE JURIES THEN DELIBERATED UNTIL A VERDICT WAS REACHED OR UNTIL JURORS REPORTED THAT THEY COULD NOT RESOLVE THEIR DIFFERENCES. THE CRITIQUE CONSIDERS THE FOLLOWING EXPERIMENTAL VARIABLES: VERDICTS, DELIBERATION TIME, ISSUES DISCUSSED, JUROR PARTICIPATION, AND JUROR SATISFACTION. THE EXPERIMENTERS FAILED TO REACH CONCLUSIONS ABOUT THE SIGNIFICANCE OF DIFFERENCES BETWEEN VERDICTS, DELIBERATION TIMES, AND NUMBER OF ISSUES DISCUSSED FOR THE TWO DIFFERENT JURY PANEL SIZES; THE SAMPLE SIZE; THE COSTLY NATURE OF JURY RESEARCH; AND THE DISTRIBUTION OF PREDELIBERATION VERDICTS AND THE FIVE-SIXTHS DECISION RULE FOLLOWED IN THE STUDY. THE APPROPRIATE DESIGN FOR THE EXPERIMENT WOULD HAVE BEEN TO ASSIGN JURORS TO PANELS ON THE BASIS OF THEIR PREDELIBERATION VOTE, THEREBY CREATING TWO GROUPS WITH SIMILAR SETS OF VOTE DISTRIBUTIONS. IT IS CONCLUDED THAT THE RELATIONSHIP BETWEEN BEHAVIORAL SCIENCE RESEARCH AND THE LEGAL SYSTEM MAY BE ADVERSELY AFFECTED IF RESEARCH IS NOT OF HIGH QUALITY. (DEP)