Chevron U.S.A., Inc. ("Chevron") petitions this Court for a Writ of Mandamus seeking relief from an order of the district court dated December 19, 1996, containing a trial plan for this litigation. We DENY the petition as it relates to the scheduled trial of the thirty selected plaintiffs referenced in the district court's order, but GRANT the petition as it relates to utilization of the results of such trial for the purpose of issue or claim preclusion.

This controversy arose out of the alleged injuries suffered by over 3,000 plaintiffs and intervenors ("Plaintiffs"), who claim damages for personal injuries, wrongful death, and property contamination allegedly caused by Chevron's acts and omissions. The Plaintiffs and their allegedly contaminated property are located in the Kennedy Heights section of Houston, Texas. The Plaintiffs contend that their subdivision was constructed on land used in the 1920's by Chevron for a crude oil storage waste pit.1 According to the Plaintiffs, when Chevron ceased using the property as a tank farm, it failed to take appropriate measures to secure the site, thereby allowing other waste to be deposited on the land. Later, Chevron sold the property for residential development knowing that the land was contaminated. Various developers filled these waste pits without remediating the land. Plaintiffs assert that the hazardous substances which were stored in the waste pits have migrated into the environment, including the drinking water supply for the Kennedy Heights section. As a result, Plaintiffs claim personal injuries and property damage.

The Plaintiffs brought suit against Chevron in both state and federal court. Subsequent to the federal suit being filed, Chevron removed the state court cause of action to federal court, which was consolidated into this case.2 On December 19, 1996, the district court approved a trial plan. The trial plan provided for a unitary trial on the issues of "general liability or causation" on behalf of the remaining plaintiffs, as well as the individual causation and damage issues of the selected plaintiffs, and ordered the selection of a bellwether group of thirty (30) claimants, fifteen (15) to be chosen by the plaintiffs and fifteen (15) to be chosen by Chevron. Chevron contends that the goal of the "unitary trial" was to determine its liability, or lack thereof, in a single trial and to establish bellwether verdicts to which the remaining claims could be matched for settlement purposes. It is this selection process which Chevron argues will not result in a representative group of bellwether plaintiffs.

Chevron filed with the district court the affidavit of Ronald G. Frankiewicz, Ph.D. which evaluated the district court's trial plan for selecting the thirty plaintiffs, concluding that such a plan was "not representative." Instead, Frankiewicz detailed the "stratified selection process" which should be used by the district court in selecting the bellwether group which would result in a representative group of plaintiffs. The district court however struck Frankiewicz's affidavit as untimely filed and redundant in substance. On January 7, 1997, the district court denied Chevron's request to certify an interlocutory appeal. This Petition for Writ of Mandamus ensued.

Our review of a trial court's plan for proceeding in a complex case is a deferential one that recognizes the fact that the trial judge is in a much better position than an appellate court to formulate an appropriate methodology for a trial. We have consistently noted that a writ of mandamus is an extraordinary remedy and is available in only limited circumstances. See Allied Chemical Corp. v. Daiflon, Inc., 449 U.S. 33, 101 S. Ct. 188, 66 L. Ed. 2d 193 (1980). We have historically reserved the issuance of the writ for "extraordinary" cases, Southern Pacific Transp. Co. v. San Antonio, Tex., 748 F.2d 266, 270 (5th Cir.1984) (citing Ex parte Fahey, 332 U.S. 258, 67 S. Ct. 1558, 91 L. Ed. 2041 (1947)), and will issue the writ where the petitioner has met its burden of proving a clear and indisputable abuse of discretion or usurpation of judicial power by a trial judge. In re First South Sav. Assoc., 820 F.2d 700, 706 (5th Cir.1987) (citing Schlagenhauf v. Holder, 379 U.S. 104, 85 S. Ct. 234, 13 L. Ed. 2d 152 (1964)).

Our traditional reluctance to meddle in the formulation of a district court's trial plan is tempered by the demands placed upon judicial resources and the extraordinary expense to litigants that typically accompanies mass tort litigation. We, therefore, as we proceed, do so mindful of the admonition contained in Rule 1--that what we do should serve the compelling interests of justice, speed, and cost-containment. See FED.R.CIV.P. 1.

We now turn to the focus of Chevron's petition, the December 19 trial plan.

The trial court has in our view quite properly categorized this litigation as complex. The mere fact that there are potentially some 3,000 claimants in and of itself complicates traditional dispute resolution. Additionally, when large numbers of claimants assert both property damage claims and claims for personal injury as well as claims for injunctive relief, it removes any question that may linger regarding the complexity of the task visited upon the lawyers and the trial court.

This case is a classic example of a non-elastic mass tort, that is, the universe of potential claimants are either known or are capable of ascertainment and the event or course of conduct alleged to constitute the tort involved occurred over a known time period and is traceable to an identified entity or entities. When compared to an elastic mass tort where the universe of potential plaintiffs is unknown and many times is seemingly unlimited and the number of potential tortfeasors is equally obtuse, the task of managing the non-elastic mass tort is infinitely less complex. In the non-elastic context, the necessity for the obtainment of maturity as reflected by a series of verdicts over time is not required in order to test the viability of plaintiffs' claims or the defendant's defenses.

The district court, after designating the case as complex, then articulated the goals of its trial plan as seeking to achieve the greatest efficiency and expedition in the resolution of all issues involved in the case. Pursuant to those goals, it structured the trial as follows:

1. Composed of thirty (30) plaintiffs, fifteen (15) chosen by the plaintiffs and fifteen (15) chosen by the defendants. The thirty (30) plaintiffs chosen shall come from the lists submitted by the parties to the state court in April of 1996. However, each side is permitted to substitute or replace not more than five (5) plaintiffs, within its discretion, on or before January 1, 1997.

2. All chosen plaintiffs shall be adults, to the exclusion of minor children, unless the children are part of a household represented by at least one adult.

3. Each individual shall be counted as a single plaintiff, as opposed to a household as a single plaintiff.

4. The trial shall focus on the individual claims of each of the selected plaintiffs and on the issue of the existence or nonexistence of liability on the part of Chevron for the pollutants that, allegedly, give rise to all of the plaintiffs' claims.

Thus, a unitary trial on the issues of general liability or causation as well as the individual causation and damage issues of the selected plaintiff shall occur.

5. The Court reserves the right to: (a) place a time limit on the length of the trial, limit the testimony of certain witnesses, limit the number of witnesses to be called on a particular issue, amend this Order, and issue additional orders.

Initially, we note the obvious. The trial plan, while clearly designed to resolve the issue of liability on the part of Chevron to all the plaintiffs by referring to a unitary trial on the issues of general liability or causation, does not identify any common issues or explain how the verdicts in the thirty (30) selected cases are supposed to resolve liability for the remaining 2970 plaintiffs. It is impossible to discern from the district court's order what variables may exist that will impact on both the property and personal injury claims in this litigation. Similar litigation typically contains property issue variables that are related to time, proximity, and contamination levels of exposure to any pollutants that may be present, and personal injury claims that contain a mix of alleged exposure related maladies that also may be affected by time, proximity, and exposure levels. We, however, may not speculate on the homogeneity of the mix of claims, the uniformity of any exposure that may have existed and what diseases, if any, may be related to that exposure. Instead our review is restricted to the record and to an examination of the district court's order.

The term bellwether is derived from the ancient practice of belling a wether (a male sheep) selected to lead his flock. The ultimate success of the wether selected to wear the bell was determined by whether the flock had confidence that the wether would not lead them astray, and so it is in the mass tort context.

The notion that the trial of some members of a large group of claimants may provide a basis for enhancing prospects of settlement or for resolving common issues or claims is a sound one that has achieved general acceptance by both bench and bar. References to bellwether trials have long been included in the Manual for Complex Litigation. See MANUAL FOR COMPLEX LITIGATION § 33.27-.28 (3d ed.1995). The reasons for acceptance by bench and bar are apparent. If a representative group of claimants are tried to verdict, the results of such trials can be beneficial for litigants who desire to settle such claims by providing information on the value of the cases as reflected by the jury verdicts. Common issues or even general liability may also be resolved in a bellwether context in appropriate cases.

Whatever may be said about the trial contemplated by the district court's December 19, 1996 order, one thing is clear. It is not a bellwether trial. It is simply a trial of fifteen (15) of the "best" and fifteen (15) of the "worst" cases contained in the universe of claims involved in this litigation. There is no pretense that the thirty (30) cases selected are representative of the 3,000 member group of plaintiffs.

A bellwether trial designed to achieve its value ascertainment function for settlement purposes or to answer troubling causation or liability issues common to the universe of claimants has as a core element representativeness--that is, the sample must be a randomly selected one of sufficient size so as to achieve statistical significance to the desired level of confidence in the result obtained. Such samples are selected by the application of the science of inferential statistics. The essence of the science of inferential statistics is that one may confidently draw inferences about the whole from a representative sample of the whole. The applicability of inferential statistics have long been recognized by the courts. See, e.g., Castaneda v. Partida, 430 U.S. 482, 97 S. Ct. 1272, 51 L. Ed. 2d 498 (1977)(using statistical data to prove discrimination in jury selection); Capaci v. Katz & Besthoff, Inc., 711 F.2d 647, 653-57 (5th Cir.1983)(using census data in gender discrimination case); Exxon Corp. v. Texas Motor Exchange, Inc., 628 F.2d 500 (5th Cir.1980)(using statistical sampling in trademark infringement suit); Ageloff v. Delta Airlines, Inc., 860 F.2d 379 (11th Cir.1988)(using evidence of life-expectancy tables to determine damages); G.M. Brod & Co., Inc. v. U.S. Home Corp., 759 F.2d 1526, 1538-40 (11th Cir.1985)(using expert testimony as to profit projections based on industry norms); United States v. 449 Cases Containing Tomato Paste, 212 F.2d 567 (2nd Cir.1954)(approving inspector's testing of samples, rather than requiring the opening of all cases).

The selected thirty (30) cases included in the district court's "unitary trial" are not cases calculated to represent the group of 3,000 claimants. Thus, the results that would be obtained from a trial of these thirty (30) cases lack the requisite level of representativeness so that the results could permit a court to draw sufficiently reliable inferences about the whole that could, in turn, form the basis for a judgment affecting cases other than the selected thirty. While this particular sample of thirty cases is lacking in representativeness, statistical sampling with an appropriate level of representativeness has been utilized and approved. As recognized by the Ninth Circuit, "[i]nferential statistics with random sampling produces an acceptable due process solution to the troublesome area of mass tort litigation." In re Estate of Marcos Human Rights Litigation, 910 F. Supp. 1460, 1467 (D.Haw.1995), aff'd. sub. nom. Hilao v. Estate of Marcos, 103 F.3d 767 (9th Cir.1996) (holding that the random sampling procedures used by the district court do not violate due process).

We, therefore, hold that before a trial court may utilize results from a bellwether trial for a purpose that extends beyond the individual cases tried, it must, prior to any extrapolation, find that the cases tried are representative of the larger group of cases or claims from which they are selected. Typically, such a finding must be based on competent, scientific, statistical evidence that identifies the variables involved and that provides a sample of sufficient size so as to permit a finding that there is a sufficient level of confidence that the results obtained reflect results that would be obtained from trials of the whole. See Hilao, 103 F.3d at 786; Michael J. Saks & Peter David Blanck, Justice Improved: The Unrecognized Benefits of Aggregation and Sampling in Mass Torts, 44 STAN. L.REV. 815 (1992). It is such findings that provide the foundation for any inferences that may be drawn from the trial of sample cases. Without a sufficient level of confidence in the sample results, no inferences may be drawn from such results that would form the basis for applying such results to cases or claims that have not been actually tried.

We recognize that in appropriate cases common issues impacting upon general liability or causation may be tried standing alone. However, when such a common issue trial is presented through or along with selected individuals' cases, concerns arise that are founded upon considerations of due process. Specifically, our procedural due process concerns focus on the fact that the procedure embodied in the district court's trial plan is devoid of safeguards designed to ensure that the claims against Chevron of the non-represented plaintiffs as they relate to liability or causation are determined in a proceeding that is reasonably calculated to reflect the results that would be obtained if those claims were actually tried. Conversely, the procedure subjects Chevron to potential liability to 3,000 plaintiffs by a procedure that is completely lacking in the minimal level of reliability necessary for the imposition of such liability.

Our substantive due process concerns are based on the lack of fundamental fairness contained in a system that permits the extinguishment of claims or the imposition of liability in nearly 3,000 cases based upon results of a trial of a non-representative sample of plaintiffs. Such a procedure is inherently unfair when the substantive rights of both plaintiffs and the defendant are resolved in a manner that lacks the requisite level of confidence in the reliability of its result.

We recognize that our due process concerns seem to blur distinctions between procedural and substantive due process. However, our difficulty in compartmentalization does not detract from the validity of our concern that is ultimately based on fundamental fairness.

The elements of basic fairness contained in our historical understanding of both procedural and substantive due process therefore dictate that when a unitary trial is conducted where common issues, issues of general liability, or issues of causation are coupled with a sample of individual claims or cases, the sample must be one that is a randomly selected, statistically significant sample. See Hilao, 103 F.3d at 782-84, 786.

We express no opinion on whether the mix of claims that collectively make up the consolidated case lend themselves to the sampling techniques required to conduct a bellwether trial or whether this is an appropriate case for a stand-alone, common-issue trial.

We are sympathetic to the efforts of the district court to control its docket and to move this case along. We also are not without appreciation for the concerns a district court might have when it concludes that some of the issues raised may be motivated by delay tactics. However, our sympathies and our appreciation for the efforts of the district court in this case do not outweigh our due process concerns.

The petition, therefore, for mandamus as it relates to the trial of the thirty (30) selected cases is DENIED. Whether the district court wishes to proceed with that trial, to secure thirty (30) individual judgments, is a matter within the discretion of the trial court. Likewise, whether the trial judge wishes to attempt to structure a common-issues trial or conduct a bellwether trial based on a properly selected sample are matters also within the discretion of the district court. The results of any such trials and appropriateness of the requisite findings necessary to so proceed will then be matters for another panel to consider in the event those decisions are subject to appellate review.

The petition for mandamus is GRANTED insofar as it relates to utilization of the results obtained from the trial of the thirty (30) selected cases for any purpose affecting issues or claims of, or defenses to, the remaining untried cases.

EDITH H. JONES, Circuit Judge, specially concurring:

I agree with Judge Parker's conclusions that mandamus must be granted in this case, that the district judge's method of selecting "bellwether" cases is fatally flawed, and that the most expeditious remedy is, without interfering with the setting of these cases, to deprive them of preclusive consequences. I believe, however, that we must elaborate further the basis for the grant of mandamus, lest we risk being consumed by petitions for similar relief and routine trial management problems. I also have serious doubts about the major premise of Judge Parker's opinion, i.e., his confidence that a bellwether trial of representative cases is permissible to extrapolate findings relevant to and somehow preclusive upon a larger group of cases.

This court has a duty not only to encant the proper standard of review applicable to the extraordinary remedy of mandamus, but also to show why that remedy is appropriate in the circumstances before us. The explanation must demonstrate why the facts here are so unique as to warrant mandamus and must reinforce that the remedy is only to be used sparingly and with utmost care. Mandamus is not a substitute for appeal in due course; consequently, the writ should only be invoked if the challenged district court order is not effectively reviewable on appeal. As the Seventh Circuit cautioned, the challenged order must inflict irreparable harm. Matter of Rhone-Poulenc Rorer Inc., 51 F.3d 1293, 1295 (7th Cir.1995). Moreover, the order "must so far exceed the proper bounds of discretion as to be legitimately considered usurpative in character, or in violation of a clear and indisputable legal right." Id. See also In re: Fibreboard Corp., 893 F.2d 706, 707-08 (5th Cir.1990).

In this case, I am persuaded that these stringent criteria are satisfied. First, this is not one case but 3,000 cases filed individually, not as a class action, and aggregated for trial management. The number of cases in which there are 3,000 plaintiffs is, even in these days of frenzied tort litigation, extremely rare. Further, because the cases concern alleged exposure over long periods of time to varying quantities of toxics, the individual circumstances of each plaintiff's claim defy easy aggregated treatment. The district court's selection of 30 "bellwether" cases, whose results would bind all 3,000 plaintiffs on the issues of general liability or causation, is probably not effectively reviewable after trial. The pressure on the parties to settle in fear of the result of a perhaps all-or-nothing "bellwether" trial is enormous.

Second, as Judge Parker's opinion notes, this is an "immature" mass tort action, in which the defendant's liability has not even been tested, much yet firmly established. The use of innovative judicial techniques particularly to resolve immature mass tort actions has been disfavored. For instance, this Court in Castano v. American Tobacco Company, 84 F.3d 734 (5th Cir.1996), refused to certify an immature tort class action brought on behalf of tobacco users. Likewise, in Matter of Rhone-Poulenc Rorer, Inc., the Seventh Circuit granted mandamus to vacate the class certification of hemophiliacs who had contracted the AIDS virus through contaminated blood transfusions. Both opinions note the potentially devastating impact of a class certification decision and its tendency to force defendants to settle even when they might have meritorious defenses. Conducting an imperfect bellwether trial in this case threatens a similar effect. An imperfectly designed bellwether group cannot yield a statistically reliable set of verdicts. Nevertheless, once in place, the verdicts would create enormous momentum for settlement. There would then be nothing to review on appeal and no realistic opportunity for Chevron to appeal.

The lack of correlation here between the bellwether plaintiffs selected and the need for a representative verdict suggests why the court's order represents a usurpation of power. Even if a bellwether trial is an appropriate vehicle for the resolution of mass tort cases, a point I question below, the results cannot serve their function of guaranteeing reliability unless the cases selected are statistically representative of the group of 3,000 plaintiffs. The court made no effort here to assure representativeness. Moreover, as Judge Parker's opinion notes, the determination of reliable representative plaintiffs is difficult in a toxic exposure case. The process involves such questions as quantity, geographic proximity, and temporal exposure to the toxic substance, comparative lifestyles, and physical manifestations of exposure, none of which were explored by the trial judge. The judge allowed the parties to pick faces from the crowd of plaintiffs, and his order forces the parties to expend huge resources preparing for a trial whose results cannot possibly fairly be extrapolated to cover the rest of the crowd. As a "bellwether", the exercise is pointless. Appellate courts can surely remedy the misdirection of resources and the almost guaranteed unfair outcome of a nonrepresentative bellwether trial. For these reasons, I think the compelling circumstances surrounding this extraordinarily large and complex case permit our considering the grant of mandamus relief.

Mandamus relief would also and more emphatically be compelled if the federal courts are not authorized to permit binding verdicts to be rendered against non-parties to bellwether trials or against a defendant with respect to plaintiffs whose cases were not tried in the bellwether group. Although Judge Parker need not have reached this larger question, he appears to have done so, asserting that the notion of a bellwether trial "is a sound one that has achieved general acceptance by both bench and bar." He further asserts that common issues or even general liability may be resolved in a bellwether context in appropriate cases. I have serious doubts about the procedure even where, as here, Chevron agreed to use of a statistically sound bellwether trial process.

The only case cited in the Manual for Complex Litigation concerning a bellwether strategy was tried by Judge Parker when he sat on the district court. Cimino v. Raymark, 751 F. Supp. 649, 653, 664-65 (E.D.Tex.1990), cited in Manual for Complex Litigation § 33.27-.28 (3d Ed.1995). One other recent case, affirmed in a split verdict of the Ninth Circuit, also used a bellwether technique. Hilao v. Estate of Marcos, 103 F.3d 767 (9th Cir.1996). These are not necessarily the only examples of bellwether trials, but they appear to be most unusual.

The use of statistical sampling as a means to identify and resolve common issues in tort litigation has, however, been severely criticized. See In re: Fibreboard Corp., supra; Hilao, supra at 787-88 (Rymer, Judge, concurring in part and dissenting in part). Among other things, the technique may deprive nonparties of their Seventh Amendment jury trial right. In Matter of Rhone-Poulenc Rorer Inc., Judge Posner observed that bifurcating liability and causation questions may require the same issue to be reexamined by different juries. That is, even if the bellwether jury found liability on the part of Chevron, later juries could be called upon to reassess that decision when faced with questions of comparative causation or comparative negligence. That all the plaintiffs are here represented by a single set of attorneys does not, in my view, alleviate Seventh Amendment concerns; to the contrary, it compounds them with potential ethical problems. Additionally, as Judge Higginbotham cautioned in In re Fibreboard Corp., there is a fine line between deriving results from trials based on statistical sampling and pure legislation. Judges must be sensitive to stay within our proper bounds of adjudicating individual disputes. We are not authorized by the Constitution or statutes to legislate solutions to cases in pursuit of efficiency and expeditiousness. Essential to due process for litigants, including both the plaintiffs and Chevron in this non-class action context, is their right to the opportunity for an individual assessment of liability and damages in each case. Nowhere did the district court explain how it was authorized to make the results of this bellwether trial unitary for any purposes concerning the 2,970 other plaintiffs' cases pending before him. In sum, I simply do not share Judge Parker's confidence that bellwether trials can be used to resolve mass tort controversies.

On the narrow basis that the court's adoption of non-bellwether methods for conducting a bellwether trial is uniquely harmful and unauthorized, I concur with the majority's award of mandamus relief.