Out to Sea in an Ocean of Distortion: Cognitive Psychology in the Jury Box

-- By MichaelBerkovits - 10 Feb 2008

Introduction

Juries in criminal trials are tasked with processing the evidence adduced at trial and passing it through a special cognitive lens called the "reasonable doubt" standard in order to determine the guilt or innocence of the accused. Among the many factors that can distort this process are a number of cognitive tendencies that recent work in cognitive psychology predicts to be present in the average juror. At least three factors operate to distort jurors' processing and weighing of the evidence: the tendency to overbelieve eyewitness testimony, to overestimate one's lie-detection abilities, and to overfocus on the crime in question to the exclusion of other, relevant considerations. And at least one factor - the inability to properly interpret the defendant's probability of innocence - warps the application of the reasonable doubt standard.

Weighing the Evidence

Jurors' processing of the evidence presented at trial is handicapped by the systematic shielding of other, relevant information. This information is meta-evidence - information crucial to the interpretation of the trial evidence - and is highly relevant.

Eyewitness Testimony and Faulty Memory

First, jurors receive scant information and no official instruction as to the unreliability of eyewitness testimony. Jurors are repeatedly told, by defense counsel and by the judge, that witnesses may lie; they are essentially never told that witnesses may be "telling the truth" about what they saw and yet still be spewing misinformation. Lawyers undoubtedly make this argument sometimes, but it so violates people's preconceptions that it probably is not worth the effort. Yet modern research shows that memory is notoriously malleable, link, all the more so when the eyewitness was involved in a terrifying or violent crime Schacter excerpt. This would not be a problem if jurors already believed that eyewitness testimony was highly error-prone; instead, jurors come in believing it is the evidentiary gold standard.

Lie-Detection

Similarly, jurors, like all people, believe that are at least decently skilled at telling when someone else is lying. Indeed, given the human tendency to inflate one's perception of one's talents, link, the average juror most likely believes that she is a better-than-average lie detector. Yet, research suggests that ordinary people cannot detect lies at better-than-chance levels: link. Even so, the jury is not told to disregard a witness's demeanor when guaging the truth of what was said. To the contrary, juries may well be encouraged by judges and counsel to take close account of such evidence.

Error Rates

A third meta-evidentiary fact kept from the jury is evidence of errors made by other juries. DNA evidence and confessions by other criminals have exonerated many wrongfully incarcerated defendants, and no one doubts that many wrongfully convicted defendants remain un-exonerated. Unlike the two topics just discussed, jurors' preconceptions in this regard are roughly, if not numerically, accurate. Jurors know that some people innocent of the crime with which they are charged are found guilty. But no judge would instruct a jury to consider this fact. A judge that did might find herself reversed on appeal. Furthermore, defense counsel would clearly be barred from introducing such evidence on relevance grounds. But, of course, if reasonable people applying a "no reasonable doubt" standard get it wrong more than a negligible proportion of the time, then evidence to that effect is highly relevant to a jury made up of such people. Unfortunately, even jurors who consider the fact that erroneous convictions do occur are subject to "focusing" errors: because the primary issue at trial, defense counsels' efforts notwithstanding, is a single (or multiple) bad act by the defendant(s), jurors are prone to view the evidence through that prism and overestimate the strength of the prosecution's case. Jurors might then reason that, even if other cases were wrongly decided, this "strong" case could not fall into that category. (For some examples of modern psychological research on focusing errors, see Gilbert, Kahneman and Tversky (wheel of numbers)).

Underweighting Low Probabilities of Innocence in Applying the Standard of Reasonable Doubt

Among the many recognized formulations of the reasonable doubt jury instructions, Am. Jur. 2nd Trial § 1159, none contain language encouraging the jury to escape the cognitive biases that attend the evaluation of low probabilities. Prospect Theory Prospect Theory suggests that people tend to neglect or underweight outcomes that occur with very small probabilities. Citation Suppose that the sum total of the evidence presented to an ordinary juror at trial suggest, to the juror, a 1% probability that the defendant is actually innocent of the alleged crime. Because jurors underweight low probabilities in their decisionmaking, the juror will tend to "value" this probability at a negligible amount, perhaps zero. A juror in this position would convict. Of course, this violates what most people would want the reasonable doubt standard to do: at least for serious crimes, we would be very uncomfortable knowing that one of every hundred people jailed was innocent.

This problem is endemic to all cases involving low, but not zero, perceived probabilities of innocence. Assume that the jurors in any given trial construct some probability of innocence based on the evidence and that these subjectively-determined probabilities, are in the same ballpark but vary around a mean. In cases where the probability of innocence is low, jurors will systematically underweight those probabilities downward. To the extent that subjective determinations of the probability of innocence go into jurors’ decisions of whether or not to acquit, the cognitive tendency to underweight low probabilities will cause more jurors to vote guilty than is justified by their own subjective weighing of the evidence.

Where Does This Leave Us?

Psychologists as judges? Discuss Frank or other readings? Why don't we just say, "absolutely no doubt," knowing that jurors tend to overconvict anyway?

Some ways to alleviate within the system But, overhauling of the system?

Lying - but questioning can still trip jurors up

are we just moving dials up and down to attain a conviction rate proper for comfort level? we're not after truth at all!

Subsection B

---

3) We keep lots of "external" evidence from the jury: parol evidence rule, unreliability of witness testimony, expert testimony about the number of wrongfully convicted people exonerated by DNA evidence, expert testimony about jurors' own cognitive biases (all we have are instructions from the trial judge about disregarding biases and judging the evidence). Logically, if some people are convicted even though a group of 12 reasonable women and men supposedly finds that there is no reasonable doubt as to their guilt, then jurors should never vote to convict - because they should not be able to trust their own impulse that there is no reasonable doubt, because other groups of 12 reasonable people get it wrong. Furthermore, do the jury instructions on reasonable doubt - that doubt that would give pause before making a life decision or a business decision - encourage jurors to bring to bear faulty cognitive systems? Prospect theory tells us that we tend to assign low-probability events a zero probability - is this the cognitive system we want governing decisions about guilt or innocence? Do our jury instructions encourage jurors to assign to low probabilities of innocence a zero value? What a "reasonable doubt" means can be affected in at least two ways.

First, do we as a society - and hence, judges and lawyers, as representatives of that society - mean to convey by "reasonable doubt" that level of reasonableness that gives us sufficient comfort with our verdicts to live with ourselves, and do jurors (implicitly?) know to interpret the phrase through that lens? Along those lines, do the demands for an adequate comfort level change over time? In 1600's Salem, the populace appeared to be relatively comfortable (but maybe not, given some historical accounts) with tests of guilt that were truly random - does she float? Today, we probably do somewhat better in determining whether the accused actually did what she or her is accused of, but we obviously don't do it perfectly - so have the demands of achieving a sufficient comfort level increased with improving crime-detection technology? Does the definition of "reasonable" change in accord with the times?

Second, do jurors themselves interpret the phrase as to accord with intuitive notions of justice? Does the phrase reasonable doubt drop out and get displaced by gut feelings? Are jurors doing anything different than what readers of sports news doing right now as they determine whether Brian McNamee? or Roger Clemens is telling the truth?

5) How do the alternatives to the jury system play into some of these things? One alternative - trial by judge - is frowned upon for several reasons. First, it is seen as elitist, and it is presumed that the people would not be happy if their liberty depended solely on state actors. Second, there is a fear that trial by judges would get us closer to a police state. But what about the power judges already have to "take the case away from the jury" through JML. Is this power exercised often enough that we do, in effect, have a system of trials by judge? Or do we really allow the community to have its say through jury-driven decisions? Clearly, some lawyers in certain types of cases continue to believe that jury trials produce much higher likelihoods of success, though some studies cited in our Civil Procedure casebook found that judges and juries tend to reach verdicts of "liable" about the same proportion of the time (I believe that the confounding factor - that the cases that are tried by judges might differ in relevant ways from the cases tried by juries - was not taken into account).

7) Many, many trials turn on the competing credibility of various witnesses. Studies show that most people are not better than chance at telling when someone is lying (I think - have to look this up). One exception was secret service agents. Why don't we have secret service agents as jurors? Or rather, why don't we have professional jurors, with the kind of training that makes secret service agents good at seeing who's telling the truth? This presumably creates the same criticisms directed at a system with all bench trials - but isn't it better, because at least we're making some attempt to see who's actually telling the truth? Still, even secret service agents obviously were not fallible. When most cases turn on competing credibility, and when we have no lie detectors, isn't it clear that trials are just probabilistic, and if so, aren't there cheaper ways of rolling the dice?

---

# * Set ALLOWTOPICVIEW = TWikiAdminGroup, MichaelBerkovits