In comparing the two, we first ask: How similar are their verdicts? Given the same information, would judges presiding at a trial render the same decision that juries, given this responsibility, make? There is an empirical answer to this question, and it is reassuring although it is flawed.
More than fifty years ago, a law professor and a statistician/sociology professor at the University of Chicago, Harry Kalven, Jr, and Hans Zeisel, began a series of studies of the jury system in the United States. One of their challenging tasks was to answer the above question. Each trial judge in the United States (whether a state or federal judge) was sent a questionnaire by mail. The judge was asked to select a recent jury trial over which he or she had presided, and to respond to a number of questions, including the judge's hypothetical verdict as well as the jury's actual verdict. The first wave of questionnaires was distributed in 1954 and 1955; somewhat more structured questionnaires were sent to judges in 1958. A total of 555 judges responded; some judges described only one trial, but others provided a large number. The result was a database of 3576 criminal trials, clearly an impressive number but by no means a random sample of all such trials.
In these criminal trials, the judge's hypothetical verdict agreed with the jury's actual verdict in approximately 75% of the cases (the actual percentage depends on whether one includes the 5 to 6% of trials in which the jury had a hung verdict). When the two disagreed, the jury more often voted for acquittal and the judge more often voted for conviction. (In 19% of the trials the jurors voted not guilty when the judge would have convicted, while in only 3% of the cases was the outcome the opposite.) A similar level of agreement, 78%, was obtained in civil trials; both the jury and the judge ruled in favor of the plaintiff about 58% of the time (Kalven and Zeisel, 1966).
These findings are sometimes interpreted as supporting a conclusion that the jury system in the United States 'works', that juries agree with judges' verdicts to a significant and hence satisfactory degree. Perhaps implicit in such a conclusion is the treatment of the judge's verdict as a criterion, as the correct decision. But such an assumption is too simplistic; first of all, judges possess certain information about the case that jurors may not have; for example, in the United States if a criminal defendant does not testify, the jury will not know whether he or she has a criminal record. Judges have defendants' past records in their files, and Kalven and Zeisel's study was not able to control how many individual judges used such information in determining their hypothetical verdicts.
A second source of uncontrolled variation was the reason or reasons for selecting particular trials by the judge. Did the judge include a particular trial in the survey because the jury's verdict was consistent with the judge's? Or because it was inconsistent? Or simply because it was a recent case? Certainly, different judges had different motivations, but this aspect of the procedure makes interpretation of the results more problematic.
Regardless, Kalven and Zeisel chose to focus on some of the possible reasons that jury verdicts sometimes differed from the judges' hypothetical verdicts. Under a category that they called 'jury sentiments' (which accounted for at least half of the disagreements), they included acquittals by juries based on their feeling that the case was too trivial for formal review, plus acquittals based on a belief that the defendant had already suffered enough, as well as other non-evidentiary reactions to the defendant, the prosecution, and the legal system. In contrast, judges' verdicts were seen as sticking closer to the law and the weight of the evidence. However, they noted that such 'sentiments' were more likely to occur when, according to the judge, the evidence in the case was close and capable of supporting either verdict. Thus they proposed a 'liberation hypothesis,' arguing that the closeness of the evidence freed the jurors to give weight to their own feelings of justice (Kalven and Zeisel, 1966).
The difference in the determinations of verdicts by judges and by juries reflects the expert-novice distinction that achieved some visibility in experimental psychology in the 1980s (Chi, Glaser and Farr, 1988). Experts are better able to discern the 'big picture' and meaningful patterns in their areas of expertise, they can process information faster than can novices, and they see deeper implications of the material in their domain. But a faster processing of information does not necessarily mean a more valid one, and the advantage of the expertise of the judge must be balanced against the motivation to do a good job present in most juries. This is especially true when one considers that judgments of credibility of witnesses are central to the factfinding in many criminal trials, and judicial 'expertise' is probably not a contributor to accuracy in this regard.
Was this article helpful?