Do Mock Trials Predict Actual Trial Outcomes [In House Defense Quarterly]
1. George Speckart, Ph.D. Do Mock Trials Predict Actual Trial Outcomes?
7. Copyright 2014 Litigation Psychology, LLC. All Rights Reserved.
4. 19 In-House Defense Quarterly n Summer 2010 n 15 who, before entering the industry, were receptionists, paralegals, acting coaches, accountants, and even cooks. In hiring a jury consultant, a wise consumer will ask, “Do you have a background, credentials or training in the prediction of behavior?” Prediction is a vital, established area of psy - chological research, and if a client wants this expertise, the client should and is enti - tled to ask for it, and obtain it. Poor trial outcomes can be tied to fail - ing to establish consulting credentials. Just as a gastroenterologist would not per - form eye surgery, there are different types of psychologists who are qualified to do some things but not others. Many indi - viduals holding Ph.D.’s in psychology have formative training working with autistic children, in counseling and psychother - apy, or chasing rats through mazes. Some may also have managed pain clinics. What a wise consumer needs, however, is a jury consultant with expertise or qualifications in designing and implementing research to predict human behavior. A jury con - sultant’s background does matter. Many jury consultants have Ph. D.’s in Commu - nication, which constitutes an excellent background for witness training and assis - tance with opening statements, but not for research on the prediction of behavior. Of course, many practitioners in the field quite legitimately may not care about prediction at all. Mock trials can be hugely informative but still fail to predict. Our topic here, however, is prediction. Esti - mating damages or exposure is an attempt to predict behavior—namely, the behavior of a group of people making a decision on damages. If we simply want to know how a group of people react to various themes and arguments, that’s certainly a legiti - mate area to research, but if we need to determine a probable trial outcome, expo - sure or potential damages, then we need to draw inferences about a jury’s behavior in the future, and that involves prediction. In those cases, it is reasonable to search for jury consultants with backgrounds of accomplishment in this area, and a litiga - tor should shop as carefully in this area as anyone would before buying anything that would cost $30,000 or more. Finally, just as nothing bars entry to the field or has established professional stand - ards to guide jury consulting, currently, no ethical standards regulate conduct by those in the field. In other words, if there ever was a “buyer beware” industry, jury consulting is it. How Is Predictive Utility Obtained? Progressing from the researchers to research design and implementation, let’s examine how appropriately experienced jury consultants can design and implement mock trials to achieve predictive utility. In research terms, this is also called “valid - ity”—namely, the extent to which research measures what it is intended to measure and can accurately extrapolate generaliza - tions to real world events. In our case, here, we want to generalize about actual court - room verdicts and damages. Constructing a valid mock trial to pre - dict jury behavior is similar to construct - ing a three-legged stool. It consists of three key components: (1) the participants, or “mock jurors,” who must faithfully repre - sent the venire; (2) the presentations, dur - ing which the lawyers must present the same materials that the real jury will hear; and (3) analysis, which must demonstrate methodological soundness. If a mock trial meets those three requirements, generally, the goal of reasonably precise forecasting is achievable. Of course, actual implementation of the research is not quite so simple, but when every decision in a mock trial procedure is resolved by making sure that it follows the “gold standard” criterion of recreat - ing what the real jury will see and hear, the results will become progressively more realistic. Obviously, regarding the first “leg” of the “stool,” the participants, the respon - dents chosen to participate in a mock trial must reflect those individuals who likely will serve on the jury panel of the particu - lar court in question. Satisfying this aspect of preparation involves elaborate recruit - ing, measurement, and screening of pro - spective mock jurors. However, most of the controversy involved in constructing mock trials centers around the second item, the presentations, as many people contend that we cannot possibly simulate the events in a courtroom. Strictly speaking, they are right. We cannot condense an entire trial into an one- to three-day exercise. But cer - tain factors have to be taken into consider - ation before concluding that a mock trial cannot have predictive utility. First, jurors do not deliberate based on what happens in the courtroom. They deliberate based on what they store and retain in their memories , and then retrieve from memory later, and jurors’ memo - ries represent a tiny subset of what has occurred in court. S. Tuholski, When Facts Don’t Fit, Some Jurors Make Up New Facts , Nat’l L.J. , Feb. 4, 2008. This is where expe - rience comes into play in designing and implementing a mock trial. Obviously, a researcher’s credentials alone are not enough to obtain predictive utility. Select - ing the evidence that is pivotal in a case and that should be included in mock trial presentations requires the combined years of judgment of the entire trial team, prior focus group research identifying jurors’ “hot buttons,” if possible, and other sub - stantive considerations based on the case fact scenario and types of claims involved. Second, jurors do not make up their minds on the basis of opening statements, a common myth. They make up their minds when listening to the witnesses . Incidentally, this is why drawing inferences about future verdicts based on spreadsheets of past ver - dicts in a venue does not work. Verdict and damages decisions hinge on witness testi - mony, which varies greatly in appeal and persuasiveness, even across the same types of cases. Therefore, condensing the witness testimony into its essential components, and reflecting this testimony faithfully in a mock trial project, are pivotal elements in achieving predictive accuracy. Selecting n Prediction of not only liability but also damages is possible on a level that far surpasses the accuracy of “guesses” or “hunches.” n
6. 21 66 n In-House Defense Quarterly n Summer 2010 to get it right. Moreover, in cases character - ized by the latter, that is, the trial team was insufficiently involved, this is not meant to disparage the lawyers. Quite unfortunately, many mock trials are conducted under “trial by hurry” conditions on the eve of trial, due to factors beyond the control of the lawyers, simply without sufficient time or necessary personnel resources. Conclusion The emergence of the potential for predic - tive utility over the last 30 years of trial sciences has profound policy implications for settlement practice, which naturally entail considering numerous cost-bene - fit issues in view of the general consensus that, for most cases, research of this type cannot fit into most litigation budgets. In one very fine article about mock trials by J.C. Johnson, the author takes the position that a mock trial is somewhat of a luxury— something that a client should use if he or she can afford it. J.C. Johnson, Mock Juries: Why Use Them? Litig. J. (Winter 2009, Vol. 35, No. 2,). While the achievement of predictive utility appears to provide obvious bene - fits, examining the costs associated with how cases are actually resolved reveals that such benefits are even more extensive than they might at first appear. Most cases cur - rently resolve to settle or proceed to trial based on “intuition.” Conducting scientifi - cally valid research to determine exposure is highly cost effective, since the margin of error in “intuition” is many times greater than the cost of the research. In fact, our analysis based on cases in which settle - ment negotiations were aborted to conduct research indicates that the margin of error in hunches—that is, the amount that pro - posed settlement amounts diverge from scientifically valid estimates about jury awards—is typically more than 10 times the cost of research itself. In short, guessing is not only more expensive than the research—it’s far more expensive than research. The cost effec - tiveness issue has been dealt with in other forums. G. Speckart, Trial by Science , Risk & Insurance, Oct. 2008. While other fac - tors do dictate settlement value, such as Mock Trial Predictiv e page 16 nuisance, risk, and corporate image, what a jury would actually do with a case is still part of the calculation in most instances, and obtaining accurate information in this area can ultimately minimize expenses. Much has also been written about a law - yer’s professional obligations to a client. Discussing with a client whether an appro - priately designed and implemented mock trial could benefit a client faced with the need for obtaining an estimate of exposure certainly would appear to be part of those obligations. If mock trials predict, then an entirely new way of viewing mock trials is required. They are no longer simply a lux - ury. Rather, a mock trial becomes a diag - nostic tool implemented to systematically assess exposure—and the question then becomes not whether a client can afford to implement one, but rather whether the cli - ent can afford not to. The state of the science has progressed so that accurate prediction is achievable. It is time for rational decisions on settle - ments to take over, particularly in a climate in which cost effectiveness is important to clients.
3. 18 14 n In-House Defense Quarterly n Summer 2010 PRETRIAL STRATEGIES ■ n George Speckart received his Ph.D. in Psychology from UCLA in 1984 with a specialization in personality and measurement, and has published extensively in the application of statistical models to the prediction of behavior. He has been active in the jury consulting field since 1983, and has conducted over 800 mock trials and focus groups in pre-trial research for numerous types of litigation. Dr. Speckart has worked with litigators in over 150 jury selections, beginning with Dalkon Shields cases in 1983, the Agent Orange litigation in 1984, and the Exxon Valdez litigation in 1994. Much More than a Hunch To obtain the right information, we must ask the right ques - tions, and we must not dismiss others. For instance, we should not casually dismiss whether mock tri - als can predict trial outcomes. After all, if mock trials did—or could—predict actual trial outcomes, the policy implications for trial planning and settlement decisions would be enormous. Before we can ask whether mock trials do or can predict trial outcomes, we must first clarify our terminology. In particu - lar, we should first define what we mean by “mock trial.” If by “mock trial” we mean a legal exercise designed to provide practice for lawyers, and the associated insights that come from practice, we should not expect prediction, as this is not a realistic goal for such an exercise. On the other hand, if we define “mock trial” as a form of psycholog - ical research , then whether a mock trial has predictive utility carries importance that demands consideration. In this article, we will consider mock trials as a form of psy - chological research, rather than as prac - tice exercises. We then may ask the questions, do mock trials predict, and can mock trials predict? The answer to the first question is, “some - times—it depends on a number of factors.” The answer to the second question is, “yes, to a substantial degree, if they are con - ducted in a certain way.” The purpose of this article is to explore when and how mock trials achieve predic - tion, and whether this amounts to “dumb luck,” or whether something else happens. Do particular systematic factors allow us to achieve predictive utility the accuracy of “guesses” or “hunches,” and in many cases, is surprisingly accurate. Now that we have established that whether mock trials predict depends on how they are conducted, let’s start with the first factor, which is the designer and implementer of the research itself. Who’s Conducting the Research? Whether mock trials predict partly depends on who does the research. First and foremost, this article documents sev - eral exemplars of precise prediction from mock trials, noted below, both in terms of verdict and damages. In each and every one of these cases, the legal team was top- notch, working assiduously to put together the most realistic mock trial possible. Clas - sically exemplifying the phrase “garbage- in-garbage-out,” mock trials are more predictive when the legal team is fully immersed, on board, enthusiastic, and willing to work hard. So, the first com - ponent of “Who’s doing the research?” is always the legal team. Do the members appreciate “balance”—that both presenta - tions in a mock trial must be as persuasive as possible? Are the graphics comparable on both the plaintiff and defendant sides? Has the team chosen truly representative witness excerpts? Are the hearts of legal team members really in the exercise? Now, a few words about the researchers are in order. Jury consultants with established exper - tise in the prediction of behavior are quite rare. Therefore, it is unsurprising that the most prevalent opinion appears to clearly be that mock trials do not predict. Consid - ering the practitioners in the field today, this opinion is certainly understandable. The field of jury consulting has no barri - ers to entry whatsoever, leading to “bar - gain services” conducted by practitioners, in mock trial research? This treatise asserts an affirmative answer and explains how and why, with the proviso that perfect predic - tion still, of course, remains unattainable. In terms of the current state of the in - dustry, some mock trials do not predict at all. Some have moderate “predictive util - ity,” modestly predicting outcomes. For ex - ample, perhaps two of three mock juries in a mock trial research project provide the same verdict as the actual jury. And others have predicted not only liability, but dam - ages very accurately indeed, in particular with a plaintiff-oriented outcome. The real questions then become (1) Can we identify differences in the research design and im - plementation between those mock trials that predict, achieving predictive utility, and those that do not? (2) What creates predic - tive utility in a mock trial? (3) How consis - tent is this predictive utility across multiple projects when the research is appropriately designed and implemented? and (4) What do we want to predict: verdicts finding liability, or damages? While predicting damages is certainly more challenging than predicting liability, presently, mock trials can also pre - dict damages. But again, whether they can and whether they do are distinct questions. The answer to the first question is, “yes, to a substantial degree. However, whether mock trials do predict damages depends on how they are conducted. While prediction is achievable “to a sub - stantial degree,” it is important to acknowl - edge that perfect prediction is obviously impossible. Unpredictable factors often impinge on a trial, from court rulings to volatile witnesses to the mysterious “luck of the draw” in jury selection. However, generally speaking, when the research is designed and implemented correctly, pre - diction of not only liability but also dam - ages is possible on a level that far surpasses By George Speckart Do Mock Trials Predict Actual Trial Outcomes? © 2010 DRI. All rights reserved.
5. 20 16 n In-House Defense Quarterly n Summer 2010 PRETRIAL STRATEGIES key testimony is an area where researchers heavily depend on lawyers and illustrates why we cannot assert that when predictive utility is achieved, it is solely because of the qualifications of the researchers. Achieving predictive utility absolutely requires a team effort . Generally, a good researcher should “coach” the trial team on these issues to get the most out of this team effort. One of the reasons that mock trials in civil cases can predict is because they in - volve depositions. It is doubtful that we can make good predictions in criminal cases because without depositions, we do not know enough in advance about wit - ness testimony. But in civil cases, in which the witnesses are more or less tied to dep - ositions, predictive utility becomes more achievable because witnesses’ testimony is largely known in advance. However, faith - fully distilling vital witness testimony into a mock trial project is exceedingly labor in - tensive, often requiring more than one day. Indeed, based on our records, of all of the instances in which we achieved predictive utility, the highest levels of accuracy were found in multi-day projects in which a great deal of painstaking labor was expended to get the witness testimony “right.” In the area of analysis, the third “leg” of the “stool,” many researchers average the damage awards proffered by each respon - dent in a focus group or mock trial to obtain an expectation of the damages in a case, but this is a faulty method because juries award damages differently from individuals. In particular, research from various sources has identified what has been termed a “severity shift,” which dem - onstrates that damages awarded by a group tend to shift upward, or escalate, com - pared with damages awarded by individ - uals acting alone. Sunstein, et al. , Punitive Damages: How Juries Decide (University of Chicago, 2002). Thus, the proper way to estimate potential damages is to average across juries, not across individuals. In hir - ing a jury consultant, a wise consumer will ask a potential candidate how he or she cal - culates damages estimates. The area of psychological measure - ment, or psychometrics, is a close sister to prediction in psychological research methodology, and various aspects of psy - chological measurement must be appro - priately observed and implemented within the mock trial research to ensure predictive utility. To enumerate these in their entirety is beyond the scope of the present article, but it suffices to say that questionnaires cannot and should not be designed and administered without attention to proper psychometric criteria. Finally, achieving predictive validity in mock trials requires years of making mis - takes and learning from those mistakes to identify various things that simply make research “go wrong.” If a plaintiff has color graphics and animations while the de - fendant simply presents black and white documents on an Elmo display, a percep - tual shift in balance will result, or what psychologists call the “demand charac - teristics” of an experiment, which creates an artificial bias in the results. Bias can be introduced in innumerable other ways, and there is generally no substitute for the watchful eye and experience of an experi - menter who has suffered all of the embar - rassments of research that went awry in various ways over the years. The Record At the bottom of the page is just a par - tial list of notable examples of precise pre - diction in mock trial research from our archives. While the skeptical reader will no doubt conclude that these have been cherry-picked, we admit that many exem - plars also exist that did not predict. As noted earlier, unexpected court rulings, unstable witnesses, vicissitudes in jury panels, what litigators often refer to as the “luck of the draw” when the members of a venire walk into court, and other factors do sometimes play a role. The list below represents a subset of those from our data - base that went to trial and in which dam - ages were awarded. Our position is simply that mock tri - als can predict, and do predict, when cer - tain conditions are met, and that most of these conditions are under volitional con - trol of the trial team and the client. The existence of mock trials that do not predict is not proof that they cannot. More often than not, in our view, when mock trials do not predict, either the researchers were unqualified, or the trial team was unwill - ing to expend the time and effort necessary n In hiring a jury consultant, a wise consumer will ask a potential candidate how he or she calculates damages estimates. n Mock Jury 1 Mock Jury 2 Mock Jury 3 Mock Jury 4 Average Mock Award/ Actual Award ETSI v. Burlington Northern et al. , 1989 $500 million $160 million $ 310 million $323 million/$345 million Newman v. Stringfellow Superfund toxic case, 1992 $175,000 $300,000 $80,000 $185,000/$138,000 Exxon Valdez , 1994 $2 billion $3 billion $4 billion $12 billion $5.2 billion/$5.0 billion AHDC v. Fresno , 2001 $1,000 $1 $10,000 $3,667/$1 Steele v. First Union , 2002 $140 million $275 million $320 million $245 million/$239 million Heavy Equipment Case, 2003 $25 million $37 million $ 112 million $58 million/$55 million Legal Malpractice Case, 2008 $88 million $20 million $140 million $83 million/$73 million Note: Some names have been withheld per the wishes of the client. Mock Trial Predictiv e page 66
- 934 Total Views
- 543 Website Views
- 391 Embedded Views
- 0 Likes
- 0 Dislikes
- 0 Comments
- 5 crm.courtroomsciences.com
- 3 184.108.40.206
- 2 www.courtroomsciences.com
- 1 courtroomsciences.com
- 2 www.csi-exch01.courtroomsciences.com
- 1 courtroomscience.com
- 2 avenel.courtroomsciences.com
- 3 smtp.courtroomsciences.com
- 2 www.smtp.courtroomsciences.com
- 2 sawgrass.courtroomsciences.com
- 3 dunbar.courtroomsciences.com
- 1 courtroomscience.info
- 2 courtroomservices.net
- 1 www.courtroomscience.info
- 1 csi-exch01.courtroomsciences.com
- 1 courtroomservice.com
- 1 www.courtroomsciences.org
- 2 www.courtroomservices.com
- 1 www.courtroomscience.com
Understanding How Jurors Would Value Your Matter1060 Views .
Winning the Environmental Case941 Views .
Derailing the Reptile Safety Rule Attack923 Views .
The Generation X and Y Factors [For the Defense]854 Views .
A Psychopath in the Courtroom847 Views .
Assessing Sympathy in Voir Dire814 Views .
Generations X and Y747 Views .
Social Influence in the Courtroom717 Views .