Compensating Plaintiffs and Punishing Defendants: Is ... - UCCS

3 downloads 0 Views 66KB Size Report
(1993). The plaintiff, a 34-year-old woman, was hired as an independent contractor ..... Defense attorneys and tort reform advocates have touted bifurcation as a.
Law and Human Behavior, Vol. 24, No. 2, 2000

Compensating Plaintiffs and Punishing Defendants: Is Bifurcation Necessary? Edith Greene,1,4 William Douglas Woody,2 and Ryan Winter3

Critics of the civil jury have proposed several procedural reforms to address the concern that damage awards are capricious and unpredictable. One such reform is the bifurcation or separation of various phases of a trial that involves multiple claims for damages. The purpose of this study was to assess the effects of bifurcating the compensatory and punitive damages phases of a civil tort trial. We manipulated the wealth of the defendant and the reprehensibility of the defendant’s conduct (both sets of evidence theoretically related to punitive but not to compensatory damages) across three cases in a jury analog study. We wondered whether jurors would misuse the punitive damages evidence in fixing compensatory damages and whether bifurcation would effectively undo this practice. Our findings indicated that mock jurors did not improperly consider punitive damages evidence in their decisions about compensation. Moreover, bifurcation had the unexpected effect of augmenting punitive damage awards. These findings raise questions about the merits of bifurcation in cases that involve multiple claims for damages.

INTRODUCTION One prominent scholar of the civil jury put it this way: ‘‘Over more than two decades, so many writings, both scholarly and journalistic, have been devoted to criticizing the institution of the civil jury that it becomes boring to recite the claims’’ (Vidmar, 1998, p. 849). The claims, briefly, are that civil juries are incompetent, biased, and driven by sympathy and passion (for a review of these claims, see Vidmar, 1995). In addition, and of relevance to the present study is the contention that civil juries are excessively generous in awarding compensatory damages and extravagant in setting punitive damages. Punitive damage awards, in particular, have been criticized as a ‘‘major 1

Department of Psychology, University of Colorado, Colorado Springs, Colorado. Department of Psychology, University of Wisconsin-Eau Claire, Eau Claire, Wisconsin. 3 Department of Psychology, St. Louis University, St. Louis, Missouri. 4 Correspondence should be addressed to Edith Greene, Department of Psychology, University of Colorado, Colorado Springs, Colorado 80933; e-mail: [email protected]. 2

187 0147-7307/00/0400-0187$18.00/1  2000 American Psychology-Law Society/Division 41 of the American Psychology Association

188

Greene, Woody, and Winter

deformity’’ in the civil justice system. In 1890, the Texas Supreme Court questioned juries’ abilities to set punitive (or exemplary) damages: ‘‘A power such as may be exercised by juries in awarding exemplary damages is liable to great abuse . . . juries, with commendable impulses, but with judgment warped by passion, no doubt often render excessive verdicts’’ (Tynberg v. Cohen, 1890, p. 317). Criticism has continued to the present day: ‘‘The fact that higher courts have reversed or reduced a great many punitive damages awards is of little consolation to defendants. It does, however, show how often juries are getting it wrong’’ (Maskin, 1998, p. 490). Research conducted since the late 1980s has shed considerable light on the contours of both compensatory and punitive damages and seems to suggest that many criticisms may be largely unfounded.5 Nonetheless, concerns about unwarranted and excessive damage awards have produced a variety of reform proposals. These include imposing a statutory cap or ceiling on awards for noneconomic (Saks, Hollinger, Wissler, Evans, & Hart, 1997) and punitive (Eisenberg & Wells, 1998) damages and raising the burden of proof during the punitive phase of a trial (Koenig & Rustad, 1993). Yet another reform proposal—limiting the evidence to which jurors are exposed by bifurcating the trial—is the subject of the present research.

Bifurcating Trials Bifurcation involves segregating discrete issues for trial and providing factfinders with evidence relevant only to those issues. Bifurcation can take various forms including separating the guilt and punishment phases in capital cases; the causation, liability and damages phases in civil cases; and the compensatory and punitive damages phases in cases that entail punitive damage claims and issues. Federal Rule of Civil Procedure 42 (b) delineates three conditions under which the trial judge can order separate proceedings: (1) for reasons of convenience,(2) when conducive to expedition or economy, and (3) to avoid prejudice. With respect to the third condition, bifurcation could prevent evidence that is relevant to one phase (e.g., punitive damages) from prejudicing the jury’s decision in another phase (e.g., compensatory damages). Consider the bifurcation of compensatory and punitive damages decision making. In theory, compensatory damages are to be assessed in light of the severity of the plaintiff’s injuries and are intended to provide compensation for economic and noneconomic losses. Punitive damages, however, are to be determined with 5

The data indicate, for example, that punitive damages are awarded infrequently and that an increase in the mean punitive award level is related in large part to a few high awards (Galanter, 1998); that jurors are loathe to award plaintiffs more than they deserve (Hans, 1996); that juries may be superior to judges in awarding damages (Vidmar & Rice, 1993); and that the nature and severity of the plaintiff’s injuries have a strong effect on the award for pain and suffering (Wissler, Evans, Hart, Morry, & Saks, 1997).

Bifurcation

189

reference to the culpability and/or maliciousness of the defendant’s conduct and to be set with knowledge of the defendant’s financial worth.6 Mogin (1998) suggests that unless the amount of punitive damages is determined in a separate phase of the trial, evidence regarding the defendant’s wealth will skew the jury’s assessment of the amount of compensatory damages. Other commentators (e.g., Ghiardi & Kircher, 1995; Koenig & Rustad, 1993) suggest that without bifurcation, evidence regarding the defendant’s reckless or malicious conduct and his financial status may prejudice the jury’s thinking about compensatory damages and that compensatory evidence can inappropriately influence punitive judgments. According to these arguments, bifurcation is necessary to avoid prejudice.7 Our study was conducted to test this assumption. We examined the extent to which jurors in three mock personal injury cases misapply evidence regarding the defendant’s wealth and the malicious nature of that defendant’s conduct (issues that are theoretically relevant only to punitive damage decisions) when they assess compensatory damages. We also wanted to know whether bifurcation would effectively undo this practice and result in more-equitable decisions regarding damages. Jurors’ Use of Evidence Regarding the Defendant’s Wealth The notion that a jury would be prejudiced in its assessment of compensatory damages by evidence regarding the defendant’s wealth rests in part on the assumption that wealthy defendants are made to pay more in damages than poorer defendants (the so–called deep pockets effect). What do we actually know about the fate of deep-pocketed defendants? The widespread perception that the tort system is biased against defendants with extensive financial resources stems partly from comparisons of the treatment of corporate and individual defendants in archival and jury analog research (MacCoun, 1996). For example, Chin and Peterson (1985) analyzed 20 years of verdicts in Cook County, Illinois, and found that juries awarded significantly more money in cases with corporate defendants than in cases with individual defendants. In a mock jury study, Hans and Ermann (1989) found that the defendant ‘‘Jones Corporation’’ was assessed higher damages in a toxic tort case than was the defendant ‘‘Mr. Jones.’’ More recent work (e.g., Vidmar, 1993; MacCoun, 1996) clarifies this issue and casts doubt on the deep-pockets assumption, however. For example, MacCoun varied the identity of the defendant in a variety of mock personal injury cases by describing the defendant as a corporation, a wealthy individual, or a poor individual. 6

The Restatement (Second) of Torts (American Law Institute, 1977) provides that the factfinder can consider evidence pertaining to the wealth of the defendant, and most courts permit a plaintiff to introduce evidence of the defendant’s wealth on the issue of assessing punitive damages. Evidence of a defendant’s wealth is generally considered to be relevant to the assessment of punitive damages based on the theory of marginal utility of money: it takes more money to punish a rich person than a poor one. Thus, for the law to inflict the same level of punishment on these two, it needs to deprive the former of more money than the latter (Owen, 1994). 7 Several states have enacted statutes that require bifurcation of the damages phases of a civil case. The California statute is particularly relevant to our concerns. It provides that, ‘‘The court shall, on application of any defendant, preclude the admission of evidence of that defendant’s profits or financial condition until after the trier of fact returns a verdict for plaintiff awarding actual damages and finds that a defendant is guilty of malice, oppression or fraud’’ (California Civil Code, Section 3295 (d), 1987).

190

Greene, Woody, and Winter

Although corporations were indeed treated differently than individuals, the verdicts on damages were insensitive to differences in perceived defendant wealth. MacCoun suggests that jurors may treat corporations differently because they find it easier to impose costly sanctions against an impersonal entity such as a corporation than against an individual and because they may hold corporations to a higher standard than individuals. Corporations are indeed treated differently, but not, apparently, because of their financial status. Jurors’ Use of Evidence Regarding the Defendant’s Conduct The notion that a jury would be prejudiced in determining compensatory damages by evidence of the defendant’s egregious conduct (relevant to punitive damages) is based on the equally uncertain assumption that jurors generally misuse such evidence. Cather, Greene, and Durham (1996) manipulated the reprehensibility of the defendant’s conduct and the severity of the plaintiff’s injuries and assessed whether mock jurors used these sources of information appropriately (i.e., whether the compensatory award was predicated only on the severity of injury and whether the punitive award was influenced only by the defendant’s level of reprehensibility). Results showed that the degree of reprehensibility of the defendant’s conduct had no effect on compensatory damage awards. More recent work by Landsman, Diamond, Dimitropoulos, and Saks (1998), however, suggests that jurors may indeed misuse evidence of the defendant’s behavior. The punitive damages evidence in their mock product liability trial included information about the conduct of a defendant manufacturer, misbehavior that was removed in time and focus from the injuries suffered by the plaintiff (and thus was theoretically unrelated to compensatory damages). Nonetheless, in some versions of the mock trial, compensatory awards were augmented when jurors were privy to punitive damages information. The Effects of Bifurcating Damages Phases Two studies have directly examined the effects of bifurcating claims for compensatory and punitive damages, and their conflicting findings suggest the need for further investigation. Horowitz and Bordens (1990) assigned mock toxic tort jurors to either a unitary or bifurcated trial condition. Participants in unitary trials heard evidence pertaining to compensatory and punitive damages before making either decision, whereas those in bifurcated trials heard the evidence in segments and made a decision (e.g., about the amount of compensatory damages) at the end of each segment. Compensatory damage awards were higher in the bifurcated trials than in the unitary trials and punitive damage awards were not affected by trial structure. Landsman et al. (1998) asked jurors to make four separate decisions in a mock product liability lawsuit: compensatory liability, compensatory damages, punitive liability, and punitive damages. Bifurcation affected these tasks differently. In particular, compensatory damages were higher in the unitary trials than in the bifurcated cases, but only when the plaintiff’s claim was moderately strong (as opposed to weak). The addition of punitive information produced significantly higher compen-

Bifurcation

191

satory awards when the case was close. However, punitive damages were higher in the bifurcated trials than in the unitary trials. Landsman and colleagues suspect that these results may be due to attrition effects: in bifurcated trials, jurors who favored the defendant departed the trial when they determined that he was not liable for compensatory damages. As a result, the jurors who remained to make decisions in the punitive phase had all favored the plaintiff in the compensatory phase. In a unified trial, all jurors were still present during the punitive phase and defense-oriented jurors could argue for restraint in the amount of punitive damages awarded. Our procedures were comparable to those of Horowitz and Bordens (1990). We asked mock jurors to read the evidence pertaining to three cases (product liability, automobile negligence, and medical negligence) and all participants made judgments of the amounts of compensatory and punitive damages to award the plaintiff. We varied the wealth and conduct of the defendant as well as the structure of the trial (unitary vs. bifurcated). If jurors misapply evidence pertaining to the defendant’s wealth and conduct, then we would expect jurors in unitary trials to award more in compensatory damages when the defendant is described as very wealthy and the defendant’s conduct is described as highly reprehensible. If, however, jurors can refrain from using this evidence until they assess punitive damages, then we would not expect to find effects of defendant wealth or conduct on compensatory awards in unitary trials. Indeed, if we find that jurors misapply the evidence and use punitive information to set the compensatory damage award (as the work of Landsman et al. would suggest), then we can determine whether bifurcation is effective in reducing such overreliance. In addition, we can assess the effects of defendant wealth, defendant conduct, and trial structure on punitive awards. If jurors are using this evidence as the law intends, we would expect to find that higher punitive damages are assessed against the wealthier and more errant defendants. However, because jurors in both unitary and bifurcated trials had access to identical information at the time that they assessed punitive damages, we had no reason to expect differences in these awards as a function of trial structure (as the work of Horowitz and Bordens would suggest). METHOD Participants Participants were 559 jury-eligible students at two moderately sized universities who were given either course credit or $5.00 for their participation. Participants were randomly assigned to one of eight conditions, and there were between 66 and 78 participants in each condition. Design This study involved a 2 (Wealth of Defendant: Wealthy, Less Wealthy) ⫻ 2 (Conduct of Defendant: Highly Reprehensible, Mildly Reprehensible) ⫻ 2 (Trial

192

Greene, Woody, and Winter

Structure: Unified, Bifurcated) between-subjects factorial. Each mock juror read three case summaries from a single cell of this eight-cell design.

Materials Case Summaries We manipulated the wealth and conduct of the defendant as well as the trial structure in three cases based on actual tort cases in which punitive damages had been awarded. The summaries were approximately 1,100–1,400 words long. The case descriptions follow.8 Products Liability Case: Herman v. Sunshine Chemical Specialties, Inc. (1993). The plaintiff, a 34-year-old woman, was hired as an independent contractor to demonstrate and sell cleaning products marketed by the defendant. (The defendant actually purchased the products in bulk from the manufacturer and sold them under the Sunshine Chemical label.) She suffered a coughing fit later diagnosed as a respiratory infection brought on by inhalation of harsh chemicals found in one of the company’s products. She filed a lawsuit, claiming the defendant was negligent in not providing sufficient warnings about the dangers inherent in use of its products. She requested $39,000 for past and future medical bills, $136,000 for past and future lost wages, and $500,000 for pain and suffering (total request ⫽ $675,000). In the Highly Reprehensible Conduct condition, the president and sole proprietor of the company had designed the product’s label by cutting and pasting together the labels from products of different manufacturers. He either did not know of or ignored a safety report that the product contained sodium hydroxide, a caustic soda. Furthermore, the company did not submit either the label or the product to OSHA for approval, as required by federal regulations. Nonetheless, the label indicated that the product met all OSHA specifications. Finally, a chemist employed by the defendant to review labels to ensure that they conveyed accurate information and that the products met OSHA regulations did not recommend any changes. In the Mildly Reprehensible Conduct condition, the sole proprietor had not designed the product’s label. Rather, he used the manufacturer’s label and simply substituted his company and product name. In the Wealthy Defendant condition, the company’s gross sales for the previous year were $14 million, whereas in the Less Wealthy Defendant condition, gross sales were $1.3 million. Automobile Negligence Case: Holben v. Midwest Emery Freight System, Inc. (1981). A car driven by the plaintiff, a 44-year-old man, was struck by a tractortrailer driven by an employee of the defendant. The plaintiff suffered severe back injuries and was partially disabled by the accident. He sued, claiming that the employee operated the tractor-trailer in a careless, reckless, and negligent manner. The plaintiff requested $94,000 in past and future medical expenses, $390,000 for past and future lost wages, and $500,000 for pain and suffering (total request ⫽ $984,000). 8

All case summaries can be obtained from the authors.

Bifurcation

193

In the Highly Reprehensible Conduct condition, the defendant trucking company had been aware that its employee had committed numerous traffic violations and had been involved in five previous accidents, one of which resulted in the death of a pedestrian. In addition, the company president acknowledged that the company had experienced significant difficulty in finding experienced drivers and had hired this employee and several others so as not to lose a profitable hauling contract. Finally, he acknowledged that the company’s policies did not adhere to the U.S. Department of Transportation’s Highway Traffic Safety Administration standards for record checks or safety and training programs. In the Mildly Reprehensible Conduct condition, the defendant trucking company had been aware that its employee had been involved in one previous accident. In the Wealthy Defendant condition, the company was described as employing 450 people with gross profits of $19.7 million for the previous year. In the Less Wealthy Defendant condition, the company employed 4 people and listed gross profits of $1.1 million. Medical Malpractice Case: West v. Vari-Care Inc. (1993). After suffering a debilitating stroke, a 72-year-old man was transferred from a rehabilitation hospital to a long-term care facility operated by the defendant. Two months after his arrival, his condition worsened and he was transferred back to the hospital, where he was found to be suffering from malnutrition and severe bedsores that had become gangrenous. He died of complications resulting from the infection. His widow filed a lawsuit, alleging medical negligence in the death of her husband. She requested $61,000 for medical and funeral expenses, $500,000 for pain and suffering, and $250,000 for loss of companionship (total request ⫽ $811,000). In the Highly Reprehensible Conduct condition, the decedent’s wife had complained to care facility staff about a foul odor in her husband’s room. In addition, an ex-employee testified regarding falsification of records and maltreatment of patients. In particular, she described situations in which patients were not helped to the bathroom or catheterized so that they urinated and defecated in their beds and were forced to lie in the excrement for up to 2 days. Records indicated that patient to staff ratio at the care facility did not meet the requirements established by the state Nursing Home Oversight Commission. Finally, the facility’s director had been offered financial incentives from the parent company for keeping costs down. In the Mildly Reprehensible Conduct condition, the decedent’s wife had described the foul odor, but no other information about the staff or its practices was provided. The Wealthy Defendant was described as owning and operating 37 long-term care facilities across the south and well as several elder day-care facilities and three Alzheimer’s treatment centers. Employees numbered 1,000 and corporate profits for the previous year were $25.3 million. The Less Wealthy Defendant was described as owning and operating two long-term care facilities, employing 15, and reporting previous-year profits of $1.1 million. Juror Questionnaire After reading each case, jurors assessed both compensatory and punitive damages. The timing of these award assessments varied, as explained in the Procedure

194

Greene, Woody, and Winter

section. After determining damages for a given case, they completed a case-specific questionnaire that included five 10-point Likert-scale questions designed to assess their thoughts about the defendant’s actions (e.g., ‘‘To what extent were the defendant’s actions malicious?’’) and three questions related to the defendant’s ability to compensate the plaintiff (e.g., ‘‘How able is the defendant to financially compensate the plaintiff ?’’).

Procedure After completing consent forms, participants read the three case summaries, all with the same combination of independent variables (e.g., wealthy defendant, mildly reprehensible conduct, and unified trial structure). Mock jurors were informed that the defendant in each case had previously been found liable for negligence and that their sole task was to award damages. Procedures differed in the unified and bifurcated trials. In the unified trials, jurors read all of the evidence (concerning the plaintiff’s injuries as well as the defendant’s wealth and conduct) before hearing jury instructions on compensatory9 and punitive10 damages and assessing these damages. In the bifurcated trials, they read evidence about the plaintiff’s injuries and instructions about compensatory damages and then assessed compensatory damages. Only then did they read evidence about the defendant’s wealth and conduct, receive jury instructions on punitive damages, and assess punitive damages. After reading each case and awarding damages, mock jurors answered the Likert-scale questions about the defendant’s conduct and ability to pay. The same procedure was repeated for the other two cases. These procedures are listed in Table 1. 9

The instruction on compensatory damages was taken from the Colorado Civil Jury Instructions and is typical of instructions typically given on compensatory damages. Mock jurors were told that they were to determine compensatory damages based on ‘‘a) any non-economic loss or injury incurred to the present time or which will probably be incurred in the future including pain and suffering, inconvenience, emotional stress, and impairment of the quality of life and b) any economic loss incurred to the present time or which will probably be incurred in the future including loss of earnings or impairment of earning capacity and reasonable and necessary medical, hospital and other expenses.’’ 10 The instruction on punitive damages also came from the Colorado Civil Jury Instructions and is also fairly typical. It read: ‘‘If you award the plaintiff actual damages on the claim of negligence, then you shall consider whether exemplary damages should be assessed against the defendant. If you find that the injury complained of was attended by circumstances of fraud or malice, then in addition to actual damages, you may also assess a reasonable sum as exemplary damages. Exemplary damages, if assessed, are to be assessed as punishment of the defendant, and as an example to others. Malice means despicable conduct which is carried on by the defendant with a willful and conscious disregard for the safety of others. A person acts with conscious disregard of the safety of others when he is aware of the probable dangerous consequences of his conduct and willfully fails to avoid those consequences. Fraud means an intentional misrepresentation, deceit or concealment of a material fact known to the defendant with the intention on the part of the defendant of causing injury. The law provides no fixed standards as to the amount of such exemplary damages. In arriving at any award of punitive damages, you are to consider the following: (1) The reprehensibility of the conduct of the defendant. (2) The amount of punitive damages which will have a deterrent effect on the defendant in the light of defendant’s financial condition.’’

Bifurcation

195 Table 1. Procedures in Unified and Bifurcated Trials Unified trial structure Evidence about plaintiff’s injuries Evidence about defendant’s wealth Evidence about defendant’s conduct Jury instructions on compensatory damage awards Jury instructions on punitive damage awards Compensatory damages assesed Punitive damages assessed Case-specific questionnaire completed Bifurcated trial structure Evidence about plaintiff’s injuries Jury instructions on compensatory damage awards Compensatory damages assessed Evidence about defendant’s wealth Evidence about defendant’s conduct Jury instructions on punitive damage awards Punitive damages assessed Case-specific questionnaire completed

RESULTS Damage Awards Because the data were highly variable and positively skewed, they were normalized with a natural log transformation and all statistical tests were performed on the transformed data. For ease of interpretation, raw dollar amounts are provided in the text and both the transformed and raw amounts are shown in the tables.11 We began our analyses with a five-way mixed ANOVA with Defendant Wealth, Defendant Conduct, and Trial Structure as between-subjects variables and Award Type and Case Type as within-subjects variables. Because there was a significant effect of Case Type, F(2, 1102) ⫽ 80.21, p ⬍ .01, as well as a significant Case Type ⫻ Award Type interaction, F(2, 1102) ⫽ 93.37, p ⬍ .01, the data were analyzed using six three-way ANOVAs (Defendant Conduct ⫻ Defendant Wealth ⫻ Trial Structure), one for each case and award type (compensatory damages, punitive damages). We discuss the effects related to Case Type below. The analyses answered three fundamental questions: (1) What effect does the defendant’s conduct have on compensatory and punitive awards? (2) What effect does the defendant’s wealth have on compensatory and punitive awards? According to legal theory, defendant conduct and wealth should affect the latter, but not the former type of award. (3) What effect does trial structure have on compensatory and punitive awards? Assuming significant effects of defendant conduct and wealth on compensatory awards, are these effects reduced in a bifurcated proceeding where compensatory awards are made without access to such information? 11

In a few instances, raw and transformed means do not correspond well (e.g., see the raw and transformed means in the medical negligence case depicted in Table 3). These disparities are related to the presence of a handful of extremely large awards which affect the raw scores, but not the transformed scores.

196

Greene, Woody, and Winter Table 2. Mean Compensatory Damages as a Function of the Defendant’s Conduct Case type Products liability Automobile negligence Medical negligence

Conduct of the defendant

Mean compensatory award (raw dollars)

Highly reprehensible Mildly reprehensible Highly reprehensible Mildly reprehensible Highly reprehensible Mildly reprehensible

496,018 433,504 973,027 854,709 726,973 1,146,109

Mean compensatory award (transformed) 12.85a 12.57b 13.50 13.34 12.95 13.00

Note: Within each case type, means with different superscripts differ significantly (p ⬍ .05).

Effects of Defendant Conduct on Compensatory Awards In theory, the egregious nature of the defendant’s conduct is not relevant to the determination of compensatory damages. If jurors inappropriately used evidence of the defendant’s reprehensible conduct, then we would have expected to find a Conduct ⫻ Trial Structure interaction, signifying that jurors delivered higher compensatory awards when the defendant acted reprehensibly, but only in unified trials (when they had access to such evidence). We found no such interaction, suggesting that jurors in unified trials were not more likely than those in bifurcated trials to rely on this evidence in awarding compensatory damages. In general, the defendant’s conduct had a significant effect on compensatory awards in only one case, products liability, F(1, 551) ⫽ 6.66, p ⬍ .05, R 2 ⫽ .012. The award was higher when the defendant’s conduct was highly reprehensible (M ⫽ $496,000 vs. $434,000). In the other two cases, compensatory damages did not differ as a function of the reprehensibility of the defendant’s conduct. These awards are shown in Table 2. Effects of Defendent Conduct on Punitive Awards Appropriately, the defendant’s conduct affected punitive damage awards in two of the three cases; products liability, F(1, 551) ⫽ 29.19, p ⬍ .01, R 2 ⫽ .05; and automobile negligence, F(1, 551) ⫽ 37.07, p ⬍ .01, R 2 ⫽ .06. The more reprehensible the conduct, the higher the award. There were no differences in punitive damages as a function of defendant conduct in the medical negligence case. These means are shown in Table 3.

Table 3. Mean Punitive Damages as a Function of the Defendant’s Conduct Case type

Conduct of the defendant

Mean punitive award (raw dollars)

Products liability Automobile negligence Medical negligence

Highly reprehensible Mildly reprehensible Highly reprehensible Mildly reprehensible Highly reprehensible Mildly reprehensible

1,161,463 947,156 1,513,252 1,782,252 5,263,800 10,054,573

Mean punitive award (transformed) 11.77a 9.63b 12.22a 9.81b 13.54 13.21

Note: Within each case type, means with different superscripts differ significantly (p ⬍ .05).

Bifurcation

197 Table 4. Mean Punitive Damages as a Function of the Defendant’s Wealth

Case type Products liability Automobile negligence Medical negligence

Wealth of the defendant

Mean punitive award (raw dollars)

Mean punitive award (transformed)

Wealthy Less wealthy Wealthy Less wealthy Wealthy Less wealthy

1,721,176 404,535 2,863,678 490,050 14,280,626 1,486,623

11.05a 10.27b 11.81a 10.15b 13.99a 12.77b

Note. Within each case type, means with different superscripts differ significantly (p ⬍ .05).

Effects of Defendant Wealth on Compensatory Awards Evidence of the defendant’s wealth is also irrelevant to compensatory damages. If jurors used this evidence in compensatory award decisions, however, we would have expected to find a Defendant Wealth ⫻ Trial Structure interaction, presumably indicating that jurors delivered higher compensatory awards against wealthier defendants, but only in unified trials when they had access to this evidence. As with conduct effects, no such interaction was apparent, suggesting that jurors in unified trials were not likely to factor the defendant’s wealth into their compensatory damage awards. In fact, there were no effects of defendant wealth on compensatory damages in any of the three cases (all ps ⬎ .5). Effects of Defendant Wealth on Punitive Awards By contrast, the extent of the defendant’s wealth is expected to influence the amount of punitive damages assessed against the defendant. Indeed it did. In all cases, higher punitive damages were assessed against the wealthier defendant: products liability, F(1, 551) ⫽ 3.18, p ⫽ .05, R 2 ⫽ .006; automobile negligence, F(1, 551) ⫽ 17.90, p ⬍ .01, R 2 ⫽ .03; and medical negligence, F(1, 551) ⫽ 28.26, p ⬍ .01, R 2 ⫽ .05. Mean punitive damage awards as a function of the defendant’s wealth are shown in Table 4. Effects of Trial Structure on Compensatory Awards The lack of any interaction involving Trial Structure on compensatory awards forecasts our findings: Jurors apparently do not use evidence of defendant conduct or wealth to determine compensation, so there is little reason to expect that awards Table 5. Mean Punitive Damages as a Function of Trial Structure Case Products liability Automobile negligence Medical negligence

Trial structure

Mean punitive award (raw dollars)

Mean punitive award (transformed)

Bifurcated Unified Bifurcated Unified Bifurcated Unified

1,498,581 589,842 2,561,437 722,556 10,858,406 4,733,778

11.12a 10.17b 11.49a 10.41b 13.64a 13.09b

Note. Within each case type, means with different superscripts differ significantly (p ⬍ .05).

198

Greene, Woody, and Winter

would be lower in bifurcated trials. Indeed, trial structure had no effect on compensatory damages in any of the three cases (all ps⬎ .4). Effects of Trial Structure on Punitive Awards Because jurors in both Trial Structure conditions had identical evidence relevant to punitive damages, we had not anticipated any differences in their punitive damages awards. Surprisingly, we found that punitive damages were higher in the bifurcated versions of all three trials than in the unified versions: products liability, F(1, 551) ⫽ 5.70, p ⬍ .05, R 2 ⫽ .01; automobile negligence, F(1, 551) ⫽ 6.69, p ⬍ .01, R 2 ⫽ .01; and medical negligence F(1, 551) ⫽ 5.97, p ⬍ .05, R 2 ⫽ .01. Mean punitive awards as a function of trial structure are shown in Table 5. Effects of Case Type Results of the MANOVA showed a significant effect of Case Type and a significant Case Type ⫻ Award Type interaction. In general, the awards were higher in the medical negligence case than in the products liability case. In addition, the difference in punitive awards between the medical negligence case and the products liability case was larger than the difference in compensatory awards between the two cases. We speculate on reasons for these differences below. Perceptions of Conduct and Wealth To assess the construct validity of two of the independent variables (conduct and wealth), we asked mock jurors to answer five Likert-scale questions concerning their perceptions of the defendant’s conduct (‘‘How careful was the defendant?’’ ‘‘To what extent were the defendant’s actions morally wrong?’’ ‘‘To what extent were the defendant’s actions malicious?’’ ‘‘To what extent were the defendant’s actions selfish and greedy?’’ and ‘‘To what extent should the defendant be punished?’’) and three questions about the defendant’s wealth (‘‘How wealthy is the defendant?’’ ‘‘How likely is it that a $1 million damage award would bankrupt the defendant?’’ and ‘‘How able is the defendant to financially compensate the plaintiff?’’). They used a 1–10 scale. The questions were the same for each case, and participants answered them immediately after reading each case. Across cases, mean responses to the five conduct questions showed significant differences as a function of the reprehensibility of the defendant’s conduct.12 Similarly, across cases, mean responses to the three wealth questions showed significant differences as a function of the defendant’s stated wealth. These data are shown in Tables 6 and 7. The findings suggest that the manipulations were perceived as intended. This analysis also revealed Case Effects and Case ⫻ Conduct interactions on the ratings for several questions. In general, conduct-related ratings were higher in the medical negligence case than in the other two cases (e.g., the defendant in the medical negligence case was perceived to be more malicious than the defendants 12

There was one exception to that general rule: In the medical negligence case, the ratings of the defendant’s carefulness did not differ significantly as a function of reprehensibility.

Bifurcation

199

Table 6. Mean Responses to Conduct-Related Questions (on 1–10 scale) as a Function of Reprehensibility of Defendant Conduct: Results of t Tests Careful?

Malicious?

Morally wrong?

Selfish?

Punished?

Products liability case (df for error terms range from 546 to 550) High 1.15 5.79 7.45 7.34 Low 1.95 4.43 5.80 6.29 a a a t 5.88 5.18 7.29 4.58a

7.80 6.55 7.44a

Automobile negligence case (df range from 544 to 549) High 1.90 5.12 5.90 Low 2.34 3.91 3.95 t 2.60a 5.28a 8.78a

6.29 3.26 13.13a

7.69 6.35 7.75a

Medical negligence case (df range from 549 to 551) High 1.74 7.94 8.70 Low 1.92 7.05 8.06 2.75a t 1.37 3.59a

7.92 7.06 3.57a

9.38 8.93 3.47a

p ⬍ .01.

a

in the products liability and automobile negligence cases), particularly when the defendant’s conduct was not highly reprehensible. We wondered whether perceptions of the defendant’s conduct and wealth would correlate with the punitive damage award, but not with the compensatory damage award (as they should) and whether these effects would differ as a function of trial structure. For each case, we combined responses to the five conduct questions into a 50point Defendant Conduct Scale (Cronbach’s alphas were .79, .82, and .80 for the products liability, automobile negligence, and medical negligence cases, respectively). We combined responses to the three wealth questions into a 30-point Defendant Wealth Scale (Cronbach’s alphas were .43, .78, and .69, respectively). We then correlated these scaled responses with jurors’ raw damage awards. Table 7. Mean Responses to Wealth-Related Questions (on 1–10 Scale) as a Function of Defendant Wealth: Results of t Tests How wealthy?

Able to compensate?

Likely to bankrupt?

Products liability case (df range from 549 to 550) Wealthy 8.01 8.45 Not wealthy 7.16 7.17 t 5.66a 7.83a

2.82 4.80 4.96a

Automobile negligence case (df range from 548 to 551) Wealthy 8.31 8.49 Not wealthy 6.60 6.26 t 11.45a 13.05a

2.06 5.66 16.80a

Medical negligence case (df range from 546 to 551) Wealthy 8.72 8.77 Not wealthy 7.11 6.73 a t 10.87 10.90a

1.76 5.72 17.72a

p ⬍ .01.

a

200

Greene, Woody, and Winter

We found that perceptions of the defendant’s conduct were significantly correlated with punitive damage awards, but not (with one exception) with the compensatory damage award. This finding mirrors the results from the assessed damages: Jurors award more in punitive damages when they perceive the defendant’s conduct to be more reprehensible, but those perceptions do not affect the compensatory damage awards. Thus, even when jurors know about the reprehensible nature of the defendant’s conduct, they do not factor that information into their compensatory judgments. These data are shown in Table 8. Perceptions of the defendant’s wealth also did not correlate with compensatory damage awards. Even when jurors had access to this information at the time that they assessed compensatory damages (in the unified condition), their thoughts about the defendant’s wealth had no influence on their judgments. The relationship between perceptions of the defendant’s wealth and punitive damage awards is less clear. Because defendant wealth should be considered in determining punitive damages and because all jurors had access to that evidence when doing so, we had no reason to expect differences in these correlations as a function of trial structure. Yet clear differences emerged: perceptions of the defendant’s wealth had no relationship to punitive damages when the trial was presented in unified fashion, yet correlated significantly with that award when the trial was bifurcated. So, although we found no effects of Trial Structure on the amount of punitive damages assessed against wealthy and less wealthy defendants, we did find that jurors’ perceptions of the defendant’s wealth were more closely related to their awards in bifurcated trials than in unified trials. These data are also shown in Table 8.

DISCUSSION Bifurcation is assumed to reduce the risk of confusion and prejudice that may arise when discrete legal claims are tried together. More relevant to the present study is the notion that when claims for compensatory and punitive damages are Table 8. Correlations of Measures of Defendant Conduct and Defendant Wealth with Compensatory and Punitive Damage Awards, Shown by Trial Structure Award type Compensatory Unified Defendant conduct Products liability Automobile negligence Medical negligence Defendant wealth Products liability Automobile negligence Medical negligence p ⬍ .05. p ⬍ .01.

a

b

Bifurcated

.14a .01 .05

.05 .09 .08

.02 .00 .01

⫺.01 ⫺.02 .03

Punitive Unified

Bifurcated

.50b .44b .23b

.46b .40b .30b

.04 .06 .08

.16b .27b .29b

Bifurcation

201

severed, jurors will be prevented from using evidence relevant to one decision while making the other. This study provides an empirical perspective on the issue. We start by asking whether jurors tended to use evidence related to punitive damages when awarding compensation. By multiple measures (i.e., the amounts of assessed damages as well as the correlations between perceptions of wealth, perceptions of conduct, and compensatory damage awards), mock jurors in our study did not improperly consider punitive damages evidence in their decisions about compensation. Compensatory awards did not differ as a function of trial structure, suggesting that jurors in unified trials were no more likely to use evidence related to defendant wealth or conduct than were jurors in bifurcated trials who had no access to this evidence. Contrary to the assumptions of many commentators, evidence related to punitive damages did not prejudice mock jurors’ thinking about compensatory damages. One explanation of this null effect is that the punitive damages evidence (i.e., evidence relevant to the extent of the defendant’s wealth and reprehensibility of the defendant’s conduct) was simply not important enough to influence jurors’ verdicts and was therefore essentially overlooked. That possibility is dispelled, however, by our findings related to jurors’ use of this evidence in punitive award decision making. Mock jurors paid considerable attention to both defendant wealth and conduct when they assessed punitive damages. As expected, higher punitive damages were assessed against wealthier defendants and against those who acted more reprehensibly. Furthermore, perceptions of the defendant’s wealth and conduct correlated significantly with the amount of assessed punitive damages. In general, then, evidence related to punitive award decisions, was appropriately considered in the context of that decision, but was essentially ignored in the process of assessing compensatory awards. Horowitz and Bordens (1990) found that compensatory awards were higher in bifurcated trials than in unitary trials. We suspect that differences in our procedures could explain some of the discrepancy in our findings. Participants in bifurcated versions of their study first gave judgments of liability and causation. Only those mock jurors who found in favor of the plaintiff on these issues were requested to assess compensatory damages. Thus, all of the jurors who awarded compensatory damages in the bifurcated trials were already favorably disposed toward the plaintiff, whereas fewer of the jurors in the unitary trial conditions (in which jurors heard all of the evidence related to liability, causation, and compensatory and punitive damages before making any decisions) were so inclined. The higher compensatory awards in the bifurcated conditions may have been affected by attrition. Our findings, as well as others which tend to show that jurors make appropriate use of punitive damages evidence (e.g., Cather, Greene, & Durham, 1996), raise questions about the merits of bifurcation in cases that involve multiple claims for damages. (We acknowledge, however, that our study did not examine another contention of tort reform advocates—that without bifucation, jurors inappropriately use evidence related to compensation to fix their punitive damages award. However, the findings of Cather et al. question this assumption as well.) Indeed, if the results of these studies can be replicated in other contexts, the call for bifurcation should perhaps be muted, as least as a remedy for the presumed prejudice that occurs in its absence.

202

Greene, Woody, and Winter

Landsman et al. (1998) demonstrate that bifurcating the damages phases of a tort case can have unexpected and even paradoxical effects. Like them, we found that punitive damage awards were higher in bifurcated cases than in unified cases. In fact, setting aside data from the medical negligence case that resulted in exceedingly13 high verdicts, the mean raw dollar award for punitive damages was $656,000 in unified trials and $2,030,000 in bifurcated trials. This finding is difficult to interpret in our study, given that all jurors had identical facts at the time they rendered punitive judgments. (One can more easily explain this finding in the Landsman et al. study as an effect of attrition; because jurors who favored the defendant were dismissed earlier in the trial, they were not available to counteract the proplaintiff sentiments expressed by jurors who remained to assess punitive damages. Our study did not involve attrition, so all jurors remained at the end of each case to give a punitive damages award.) Why would punitive damage awards be higher in bifurcated trials? Perhaps, as suggested by Landsman et al., jurors in bifurcated trials were angered by the fact that they were not given all of the evidence before they assessed compensatory damages and therefore chose to augment their punitive awards as an expression of this ire. Although not ruling out the role of motivation, we prefer a cognitive explanation. The effect may be related to the timing and salience of evidence regarding the defendant’s wealth and conduct. Evidence related to these issues is presented closer in time to the punitive damages decision in bifurcated trials than in unified trials. (Recall that in unified trials, jurors heard jury instructions and assessed compensatory damages after hearing punitive damages facts, but before they could go on to the punitive damages decision.) Thus, the relevant evidence is more obviously salient and applicable to the punitive award decision in bifurcated cases. However, by the time that jurors in unified trials assess punitive damages, they have had more evidence to synthesize and multiple decisions to make. As a result, they may have less ability to attend to and focus on the discrete evidence relevant to any particular decision. Some support for this notion comes from the finding that perceptions of the defendant’s wealth (e.g., ‘‘How able is the defendant to financially compensate the plaintiff?’’) correlated with the amount of assessed punitive damages only in the bifurcated cases. The evidence relevant to the defendant’s financial status was relatively short and, we suspect, overshadowed by the lengthier description of the reckless nature of the defendant’s conduct. As a result, it may have been less influential to jurors in unified trials (who had a larger amount of evidence to ponder) than to jurors in bifurcated trials. For purposes of generalizability, we chose to examine the effects of trial structure across three different kinds of cases. Although we had not anticipated finding case differences, such effects did emerge. In general, awards were higher for the medical negligence case than the products liability case. In addition, the punitive awards in the products liability and medical negligence cases were more disparate 13

We suspect that these high awards may be related to the fact that the plaintiff was allegedly killed, and not merely injured, as a result of the defendant’s actions. We return to this point below.

Bifurcation

203

than the compensatory awards. Why? Differences may be attributable to details surrounding the two cases. Recall that the actions of the defendant in the medical negligence case allegedly caused a death, rather than an injury (as in the products liability case). In addition, note that the defendant’s conduct in the medical negligence case—whether reprehensible or not—was deemed more malicious, morally wrong, and punishable than was the defendant’s conduct in the products liability case. Perhaps jurors hold medical care providers to a higher standard of conduct than people in business. Differences could also be attributable to the requests for damages that we provided in the case summaries. (The plaintiff in the products liability case requested $675,000 whereas the plaintiff in the medical negligence case asked for $811,000. These figures may have served as anchors for jurors’ assessments of damages.) Alternatively, the case effects may be attributable to some combination of these factors. Undoubtedly, the defendant’s wealth and conduct (as well as the structure of the trial) can have differential effects on jurors’ judgments as a function of the type of case and the cause of action.14 Future research that more carefully examines these case effects would be welcomed. We acknowledge the usual caveats about generalizing from jury simulation studies. Our results are limited by our use of written materials and a synopsized version of the facts. This abbreviated trial format may have led jurors to use more rational and logical decision-making processes than would be true when the evidence is more complex and detailed. In addition, we concede difficulties in the way we chose to manipulate the defendant’s conduct. Although such conduct was generally deemed more detestable in the Highly Reprehensible Conduct conditions than the Less Reprehensible Conduct conditions, the defendant’s conduct failed to affect the punitive damage award in the medical negligence case and inappropriately affected the compensatory award in the products liability case. We admit to some concerns related to the construct validity of this variable. Finally, mock jurors did not deliberate, and we suspect that some interesting and informative data could be gathered from analyses of such discussions. We have no reason to believe, however, that jury deliberations would result in more reliance on irrelevant evidence. In fact, juries tend to outperform individual jurors in rejecting inappropriate information (Kaplan & Miller, 1983) and are better equipped to follow judicial instructions than are individual jurors (Kerwin & Schaffer, 1994). Landsman et al. (1998) found that juries fixed punitive damages in similar fashion to individual jurors. We suspect that our results would not change if mock jurors had been able to deliberate. In the end, this is an empirical question, of course. Defense attorneys and tort reform advocates have touted bifurcation as a useful mechanism to reduce antidefense bias in civil trials (Landsman et al., 1998). Proponents of bifurcation have assumed that its use will result in more equitable verdicts because jury damage awards would no longer be based on unfair considerations of irrelevant evidence. Results of the current study question the legitimacy of these beliefs and raise the ironic possibility that bifurcation may disadvantage 14

Note, however, that the manipulations of defendant conduct and wealth were consistent across the three cases. For each case, the ‘‘reprehensible conduct’’ was perceived more negatively than the ‘‘less reprehensible conduct’’ and the ‘‘wealthy defendant’’ was perceived as more wealthy than the ‘‘not wealthy defendant.’’

204

Greene, Woody, and Winter

defendants more than it advantages them, at least in cases involving claims for multiple types of damages.

ACKNOWLEDGMENTS This research was supported by a grant from the University of Colorado Committee on Research and Creative Works. Portions were presented at the American Psychology-Law Society Meeting, Redondo Beach, California 1998.

REFERENCES American Law Institute. (1977). Restatement (Second) of Torts. Washington, DC: Author. California Civil Code (1999). Retrieved October 19, 1999, from Lexis–Nexis database on the World Wide Web: http://web.lexis-nexis.com/universe. Cather, C., Greene, E., & Durham, R. (1996). Plaintiff injury and defendant reprehensibility: Implications for compensatory and punitive damage awards. Law and Human Behavior, 20, 189–205. Chin, A., & Peterson, M. (1985). Deep pockets, empty pockets: Who wins in Cook County? Santa Monica: Rand Corporation. Eisenberg, T., & Wells, M. (1998). Punitive damages after BMW, a new capping system, and the reported opinion bias. Wisconsin Law Review, 1998, 387–426. Galanter, M. (1998). Shadow play: The fabled menace of punitive damages. Wisconsin Law Review, 1998, 1–14. Ghiardi, J., & Kircher, J. (1995) Punitive damages law and practice. Deerfield, IL: Clark, Boardman and Callaghan. Hans, V. (1996). The contested role of the civil jury in business litigation. Judicature, 79, 242–248. Hans, V., & Ermann, D. (1989). Responses to corporate versus individual wrongdoing. Law and Human Behavior, 13, 151–166. Herman v. Sunshine Chemical Specialties, Inc. 627 A. 2d 1081 (N.J., 1993). Holben v. Midwest Emery Freight System, Inc. 525 F. Supp. 1224 (1981). Horowitz, I., & Bordens, K. (1990). An experimental investigation of procedural issues in complex tort trials. Law and Human Behavior, 14, 269–285. Kaplan, M. & Miller, C. (1983). Group discussion and judgment. In P. Paulus (Ed.), Basic group processes (pp. 65–94). New York: Springer-Verlag. Kerwin, J., & Schaffer, D. (1994). Mock jurors versus mock juries: The role of deliberations in reactions to inadmissible testimony. Personality and Social Psychology Bulletin, 20, 153–162. Koenig, T., & Rustad, M. (1993). The quiet revolution revisited: An empirical study of the impact of state tort reform of punitive damages in products liability. Justice System Journal, 16, 21–44. Landsman, S., Diamond, S., Dimitropoulos, & Saks, M. (1998). Be careful what you wish for: The paradoxical effects of bifurcating claims for punitive damages.Wisconsin Law Review, 1998, 297–342. MacCoun, R. (1996). Differential treatment of corporate defendants by juries: An examination of the ‘‘deep-pockets’’ hypothesis. Law and Society Review, 30, 121–161. Maskin, A. (1998). Litigating claims for punitive damages: The view from the front line. Loyola of Los Angeles Law Review, 31, 489–502. Mogin, P. (1998). Why judges, not juries, should set punitive damages. University of Chicago Law Review, 65, 179–224. Owen, D. (1994). A punitive damages overview: Functions, problems and reform.Villanova Law Review, 39, 353–413. Saks, M., Hollinger, L., Wissler, R., Evans, D., & Hart, A. (1997). Reducing variability in civil jury awards. Law and Human Behavior, 21, 243–256. Tynberg v. Cohen, 76 Tex 409, 13 SW 315 (1890). Vidmar, N. (1993). Empirical evidence on the deep pockets hypothesis: Jury awards for pain and suffering in medical malpractice cases. Duke Law Journal, 43, 217–266. Vidmar, N. (1995). Medical malpractice and the American jury: Confronting the myths about jury incompetence, deep pockets and outrageous damage awards. Ann Arbor; University of Michigan Press.

Bifurcation

205

Vidmar, N. (1998). The performance of the American civil jury: An empirical perspective. Arizona Law Review, 40, 849–899. Vidmar, N., & Rice, J. (1993). Assessments of noneconomic damage awards in medical negligence: A comparison of jurors with legal professionals. Iowa Law Review, 78, 883–911. West v. Vari-Care, Inc. Baldwin County (AL) Circuit Court No. CV91-617 (1993). Wissler, R., Evans, D., Hart, A., Morry, M., & Saks, M. (1997). Explaining ‘‘pain and suffering’’ awards: The role of injury characteristics and fault attributions. Law and Human Behavior, 21, 181–207.