# Statistical evidence: how to help jurors understand and use it properly.

No matter how skillfully expert witnesses and counsel present potentially bewildering numbers, jurors' psychological aspects must be considered

THE USE of statistical evidence in a jury trial poses specific problems for the parties and the attorneys. This is so in both civil and criminal cases. The effective communication of statistical evidence can be of critical importance to the jury's impression of the case. Statistical evidence in civil cases might involve business concepts, such as market share in an antitrust dispute, or the analysis of the racial or age makeup of the workforce in an employment discrimination case. Mathematical evidence can be important in professional negligence cases that involve engineering specifications, or product liability disputes in which there is exposure to a toxic agent alleged to have caused an illness or injury.

Jurors bring individual beliefs and understandings to the courtroom, whether correct or not, about the use and importance of statistical evidence. These attitudes are the backdrop that must be considered by litigators when shaping their cases for presentation to a jury.

JURORS' UNDERSTANDING OF STATISTICS

Prospective jurors typically understand very little about statistics. Many people have a fear of numbers, and their anxiety causes them to view those who are comfortable with numbers as nerdy or different. "Math geeks" often are ridiculed and mocked. This causes people who are comfortable with numbers to be hesitant to admit it.(1)

When considering how prospective jurors process statistics, it is important to remember that people who are less comfortable with numbers tend to personalize an event more than do those who have a greater understanding of probabilities. If people are uncomfortable with numbers, they are likely to underestimate drastically the existence of coincidence while giving very little weight to statistical evidence on the same subject. In other words, they will perceive causal relationships where none actually exist.

Most people tend to understand probability in a non-mathematical sense and infer plausibility and/or possibility from a mathematical probability. Peter Sedlemeier notes that human beings were not designed to make decisions based on mathematical rules, and as a result, human judgment is often incorrect. One of the ways that human beings deal with statistical problems is to apply heuristics to help them in their decision-making process. These heuristics cause biases or cognitive illusions that might lead to flawed conclusions.(2)

Heuristics are rules of thumb people use to help them simplify the decision-making process. They are processes that save time and effort. Heuristics are an extension of a person's gut reaction and are derived from personal experiences with similar tasks or problems. People rely on heuristics because they believe that the process provides them with reliable outcomes, although that may not always be the case.

One type of heuristic is known as the availability heuristic. It involves the ease with which one can come up with an example of the occurrence of a similar event. If someone has heard of a disease being caused by exposure to a chemical and has a pre-existing notion that many diseases are caused by such exposures, the likelihood that this potential juror will believe the evidence of causation increases. Availability also means that one type of outcome is easier to visualize than an alternative outcome. That is, it is easier to believe that the plaintiff's illness was caused by the chemical than to believe that there is no explanation for the fact that the illness has developed and/or that each person may be susceptible to a random chance of contracting the same illness. The ability to imagine an event's occurrence also makes it more likely in the mind of the decision-maker, even if there is no statistical basis to support that conclusion.(3)

Another psychological concept that impacts a person's ability to use statistics effectively is known as hindsight bias. It is the perception that the likelihood that a result caused by a preceding event is much greater once the result has occurred.(4) This is an important factor that often dilutes one's perception of the importance of statistical evidence offered in support of a defendant. Since it seems obvious to an observer, after the fact, that the two facts have a causal relationship, any statistics presented to the contrary will be overshadowed by the person's reliance on hindsight.

Potential jurors also are influenced by media reports that seem to offer a correlation for every negative event that occurs and imply that these correlating actions may be the cause: of the bad outcome. The number of conflicting reports of the health effects of various foods is an example of the exposure typical jurors have to contradictory stories that may or may not be based on sound statistical evidence. The result is often a cynical response to such research and the idea that "statistics can be used to say anything."

Another response to media accounts of statistics is for people to react strongly to the stories of risks, even though the probabilities associated with these risks often are very small. People do not think about the low probability of the bad outcome. They focus on the human element of the story and think, "But what if I'm the one who becomes ill?" Statistics cannot be fairly evaluated without knowing the sample size, the composition of the sample, the methodology, the confidence intervals, and the significance levels used in the research.

But many potential jurors are not the least bit interested in these concepts, and those who are willing to try to understand them will interpret the information within the context of experiences that are familiar to them. People draw pragmatic implications from the data available and/or understood by them to fill in gaps that exist in their understanding of the evidence.(5)

JURORS' USE OF STATISTICAL EVIDENCE

The use of statistical evidence by jurors has been studied by a number of researchers in recent years. The results of these academic efforts offer useful information for defense counsel.

Edward Lee Schumann has shown that jurors exhibit a basic understanding of how to use statistical information but are easily confused by an attorney's presentation of the evidence. Therefore, they often did not use it properly. Jurors were able to see the logic in each attorney's presentation of the evidence and had difficulty determining which was the weaker argument.

Schumann also found that jury members tend to shift toward the majority, particularly when dealing with complex concepts with which they were uncomfortable. He also found that if there are only a few jurors in the group who are able to recognize the improper statistical arguments put forth by an attorney, there's no indication that the minority can convince the majority of the flaws they find to exist in these arguments.(6)

The importance of jurors' pre-existing beliefs about a given subject matter is noted in the findings of Gary L. Wells.(7) He found that jurors' hypotheses concerning the cause of the outcome described by the plaintiff affects jurors' views of the truth or falsity of the evidence presented. This is described as fact-to-evidence reasoning, and it differs from the typical assumption that the elements of the evidence are considered first, prior to developing a conclusion about the ultimate facts of the case.

For example, jurors often conclude that since someone is sick and was exposed to the defendant's product, the product caused the sickness, rather than relying on the statistical evidence to the contrary that there is no known cause for the plaintiff's illness. If jurors' beliefs about the result in the case are consistent with the evidence presented, that evidence will be believed and relied on. If the evidence presented contradicts their beliefs about the outcome, that evidence will be viewed critically and is less likely to be relied on.

Wells also offers the conclusion that jurors require a greater degree of certainty to hold a defendant liable. However, this author's applied research experience indicates that in most cases jurors want to see an increased probability that the defendant is not liable.(8) Wells's underlying premise is correct, but it works to the advantage of the plaintiff rather than the defendant. Wells does note that his research lacks many of the additional types of evidence and arguments that are typically found in a trial. He recognizes that this prevents his research from being a good predictor of an expected outcome at trial. Rather, it is informative as to the process jurors might use to evaluate statistical evidence.

Wells also defines naked statistics as probabilities that are not case-specific to the events in question, but rather are developed independently--for example, studies that were conducted by researchers unrelated to the litigants and prior to the allegations of wrongdoing involved in the litigation. In situations where these statistics support the defendant's position, they can carry greater weight with jurors than those studies and opinions created specifically for litigation.

As noted, many people will infer possibility or plausibility from the citing of statistical probabilities. This was researched using a legal framework as long ago as 1971 when Laurence Tribe studied the role of mathematics in the legal process.(9) This research found that jurors make verdict determinations based on their subjective understanding of the probabilities involved, not on the actual mathematical odds. They use their life experiences to tell them what the actual likelihood of the occurrence of an event.

Given jurors' reliance on their beliefs and experiences, it is not surprising to learn that Brian Smith and his colleagues found that jurors tend to underuse statistical evidence in making verdict determinations.(10) In contrast to Schumann's findings, Smith did not find that jurors are prone to confusion caused by an attorney's or an expert's improper assertions about the evidence, but rather they simply do not give conflicting statistical evidence much weight when developing their conclusions. When experts with equal credentials present completely contradictory analyses of the available data, it should not be surprising that jurors elect to ignore them both.

Similarly, research conducted by Edward Wright and his colleagues noted that people are reluctant to make liability decisions based on statistical evidence alone.(11) Jurors often rely more heavily on their assessment of motivations of the people involved in the dispute, along with the knowledge they have acquired based on their own relevant experiences, rather than the statistical evidence presented during trial.

The concept of anticipated justification also may impact jurors' assessment of statistical evidence. They may be concerned that their decision will need to be justified to someone in the future. Wright et al. hypothesize that this may be why jurors require such a high objective probability in order to reach their decision. While they suspect that this explains a reluctance to hold a defendant liable, as noted above, this author's research experience would suggest that this more likely explains jurors' hesitation at finding a defendant not liable. Jurors would rather risk holding a wealthy corporation liable in error than risk denying an injured plaintiff compensation that may be needed to address significant medical or financial problems. Many jurors would rather find against the defendant, just in case it did cause the harm complained of by the plaintiff.

HELPING JURORS UNDERSTAND

Here are suggestions for defense counsel based on THE research noted, along with the author's applied research experience:

* Remember the unique perspective of the jurors. The more complex the statistical information being presented, the less ability jurors will have to grasp its meaning and importance. To them, you will be speaking in a foreign language, so you need to help them to translate your words into a language that they can understand. They have been asked to listen to information on subjects they know little or nothing about. Recognize that you are asking the jurors to learn and apply rules to their decision-making process that are new to them.

* Remember the jurors' concern with motive, context and alternative causation. Although the law does not require a defendant to prove what did cause an injury, jurors will be uncomfortable denying the claim of an injured party unless they feel sure that another cause is likely. Regardless of how convincing you believe your statistical data to be, jurors who are less comfortable with numbers will worry about the motives of the parties and want to be assured that some other cause is possible. By carefully evaluating this aspect of your case, you can also address the concerns of jurors who will be bothered by hindsight bias.

* Tell a story that will hold the jurors' attention. One of your greatest challenges will be to make your story as concise as possible. The more levels of detail you need to reach your conclusion, the harder it will be for the jurors to follow you on this journey. You must strive to simplify the message as much as possible.

* Develop graphics that communicate the story. Research by Sedlemeier shows that statistics are taught most effectively by hands-on practical applications and by the use of graphic representations. At trial, the opportunity to learn "by doing" does not exist for jurors, so the use of appropriate visual tools becomes that much more important. A multimedia approach will work best because the variety will help you in your quest to maintain the jurors' interest level.

* Show the jurors why they can trust statistics in general. Most jurors enter the courtroom with a somewhat cynical view of statistics. Don't spend too much time telling the jurors that you have the best statistics without first helping them to believe that numbers in general can be used to draw the correct conclusion. You must create a level of trust that tells the jurors that they can rely on statistics to help them make sound decisions. Only then can you begin to try to persuade them of the merits of your statistics. One way to do this is to show jurors examples of statistical analyses that are evident in everyday life and how the collection and analysis of this data complies with the necessary scientific principles. To be effective, you must use examples that the jurors will be able to confirm as accurate and correct from their own life experiences.

* Teach the jurors why your statistics are trustworthy. Once you have built the foundation that explains basic statistical practices, you can then compare and contrast your data to illustrate its validity. When your data involves numbers that are very large or exceedingly small, you must represent them in terms the average person can visualize. The research by Paulos shows that the more personal and real the examples are to the jurors, the more real the concept you are trying to convey becomes. Help your expert to develop points of reference for the jury that are easy for them to visualize. For example, use the volume of air in the courtroom or the seats in a local stadium to illustrate the size of something.

* Teach the jurors why the opposition's statistics are not trustworthy. Once you have established the threshold for what makes statistics reliable, you then can compare and contrast the data relied on by your opposition with these same standards. If you are to convince the jurors that the statistics cited by the opposing expert are not trustworthy, you must be able to evaluate them on the same simple criteria that you have used to establish trustworthiness in the first place.

* Be critical of the opposing expert's methods or conclusions, but do not attack the witness personally. No matter how poorly you regard the opposing expert, you should refrain from attacking him or her on a personal level. If jurors perceive that you are "beating up on" the expert, it is very likely that it will hurt your case more than help it. Focus on the witness's professional experiences and opinions, but do not let any personal animosity you may feel show in the presence of the jury. Jurors want to see that everyone is treated fairly, including the witnesses who are testifying.

* Choose your expert carefully. Sometimes expert testimony boils down to a battle of credentials, but it's usually not the typical credentials that are most important to jurors. Teaching skills and realworld experience are the things that jurors weigh when assessing the value of an expert's testimony. They want to hear from someone who can communicate with them in a language they can understand. Equally important is someone who has gained his or her knowledge from experience that jurors will consider "hands on" rather than in an isolated academic environment. You must weigh how these aspects of an expert's credentials will be perceived in addition to his or her more traditional credentials.

CONCLUSION

While jurors' comprehension of scientific evidence is limited only by the communication skill of the counsel and witnesses who present that evidence, jurors' use of statistical evidence is influenced by psychological factors that cannot be overlooked. Not only must the trial attorney present an expert that employs strong communication techniques, it is equally important to be cognizant of the pre-existing beliefs that jurors will have when they are presented with statistical evidence by a defendant.

(1). See JOHN ALLEN PAULOS, INNUMERACY: MATHEMATICAL ILLITERACY AND ITS CONSEQUENCES (Vintage Books, 1990).

(2.) See PETER SEDLEMEIER, IMPROVING STATISTICAL REASONING: THEORETICAL MODELS AND PRACTICAL IMPLICATIONS (Lawrence Erlbaum Associates Inc., 1999).

(3). See SCOTT PLOUS, THE PSYCHOLOGY OF JUDGMENT AND DECISION MAKING (McGraw-Hill Inc., 1993).

(4.) Id.

(5.) R.J. Harris & G.E. Monaco, Psychology of Pragmatic Implication: Information Processing Between the Lines. 107 J. EXPERIMENTAL PSYCHOLOGY: GENERAL 1 (1978).

(6.) EDWARD LEE SCHUMANN, EFFECTS OF ATTORNEYS' ARGUMENTS ON JURORS' USE OF MATHEMATICAL EVIDENCE (University of California, Irvine, 1991).

(7.) Gary L. Wells, Naked Statistical Evidence of Liability: Is Subjective Probability Enough? 62 J. PERSONALITY AND SOCIAL PSYCHOLOGY 739 (1992).

(8.) The author has participated in research involving the decision-making processes of thousands of mock jurors involving actual cases pending in venues around the United States. She also has obtained feedback from hundreds of actual trial jurors post-verdict to assess the importance of specific pieces of evidence and elements of testimony.

(9.) L.H. Tribe, Trial by Mathematics: Precision and Ritual in the Legal Process, 84 HARV. L. REV. 1329 (1971).

(10.) Brian C. Smith, Steven D. Penrod, Amy L. Otto & Roger C. Park, Jurors' Use of Probabilistic Evidence, 20 LAW & HUMAN BEHAVIOR 49 (1996).

(11.) Edward F. Wright, Lora MacEachern, Elaine Stoffer & Mancy MacDonald, Factors Affecting the Use of Naked Statistical Evidence of Liability, 136 J. SOCIAL PSYCHOLOGY 677 (1996).

Nancy L. Neufer is senior research associate in the Pennsylvania office of DecisionQuest. She holds B.S. and M.S. degrees in business administration from Pennsylvania State University and has coordinated jury research in more than 250 cases of numerous types. She was senior jury consultant for FTI Corp. and worked in the airline industry prior to joining DecisionQuest.
COPYRIGHT 2000 International Association of Defense Counsels
No portion of this article can be reproduced without the express written permission from the copyright holder.