Owens-Corning Fiberglas Corporation (“OCF”), a fabricator of asbestos pipe-covering products, appeals from jury verdicts in favor of John Consorti, a pipe insulation worker who developed mesothelioma as the result of exposure to asbestos dust, and his wife Frances for loss of consortium. The trial in the Southern District of New York before Robert W. Sweet, Judge, which consumed 25 trial days, consolidated the claims of four plaintiff couples against numerous manufacturers of asbestos products. The jury rendered verdicts for the plaintiffs total-ling in excess of $47 million. OCF’s main claims on appeal are: (1) that the four cases should not have been consolidated; (2) that the $12 million award for John Consorti’s pain and suffering was excessive; and (3) that Frances Consorti had no claim under New York law for loss of consortium because John’s noxious exposure occurred prior to their marriage. OCF also challenges the sufficiency of evidence on various points, and numerous rulings made at trial.
We certified the question of Mrs. Consor-ti’s entitlement to sue for loss of consortium to the New York Court of Appeals. Consorti *1005v. Owens-Corning Fiberglas Corp., 45 F.3d 48, 49 (2d Cir.1995). On October 24, 1995, the Court of Appeals ruled that the facts proved do not give rise to a cause of action for loss of consortium under New York law. Consorti v. Owens-Corning Fiberglas Corp., 86 N.Y.2d 449, 634 N.Y.S.2d 18, 657 N.E.2d 1301 (1995). Accordingly, we vacate the award for loss of consortium. We affirm on all other issues except the size of John Con-sorti’s award for pain and suffering, which we believe was beyond what New York law permits.
BACKGROUND
A detailed account of the case is set forth in the district court’s opinion in In re New York Asbestos Litig., 847 F.Supp. 1086 (S.D.N.Y.1994) (“Consorti et al.”). We include here only the facts necessary for the discussion below.
John Consorti (“Consorti”) was a 40% owner of Veteran Pipe Covering, a family insulator business. He worked for Veteran as a pipe covering insulator from 1960 to 1963, and from 1970 to 1978. From 1963 to 1970, he worked as an insulator for another family business, State Pipe Covering. In 1978, he became Vice-President of Veteran, a position in which he remained until 1992. While working at these businesses, and at least through the mid-1970s, Consorti was exposed to asbestos products, including OCF’s product.
Consorti began to suffer back problems in August 1991. In February 1992, he was diagnosed with pleural mesothelioma, an incurable cancer of the lining of the lung. He died after trial, at the age of 51, in November 1993.
Consorti initiated his asbestos personal injury action in August 1992 against numerous defendants in the Southern District of New York. His suit was one of the many thousands of asbestos cases reassigned by the Multi-District Litigation Panel to the Eastern District of Pennsylvania for discovery and pre-trial proceedings. Because of his rapidly advancing illness, his case was deemed an emergency “hardship” and was remanded to the Southern District of New York for expedited trial. It was consolidated for trial in 1993 with three other cases of mesothelioma due to asbestos exposure. In re New York Asbestos Litig., 145 F.R.D. 644 (S.D.N.Y.1993) (“New York Asbestos I”); In re New York Asbestos Litig., 149 F.R.D. 490 (S.D.N.Y.1993) (“New York Asbestos II”). OCF was a defendant only in the Consorti action.
OCF moved to sever the Consortis’ case from the other plaintiffs, and the court denied the motion. New York Asbestos II, 149 F.R.D. at 491.
The trial of the consolidated action began on June 21, 1993. At trial, Consorti proved that his incurable fatal disease was caused by his ingestion and respiration of asbestos fibers. He demonstrated that he had endured enormous suffering from the mesothelioma and was likely to die within a few more months..
On July 22, 1993, after a five-week trial, the jury returned its first special verdict, awarding another plaintiff, Vincent Tabolt, over $13 million in damages, including $7.5 million for approximately 18 months of pain and suffering. After the jury announced the award, Judge Sweet, seeking to encourage the panel after its long days of hard work, said:
You are probably the best jury that has ever been put together in the United States of America. You paid very close attention, you worked very hard.... Thank you very much for the good work you have done.
Outside the presence of the jury, OCF objected to the court’s praise of the jury, arguing that the judge’s remarks would be taken as approval of the enormous verdict and would influence the verdicts yet to be rendered. The following day, the court denied a motion for a mistrial, but gave a curative instruction: ■
I wanted to be clear that since you have entered a verdict in one case that I was referring, of course, not to the verdict in any way, about which I have no control, that is entirely in your hands, but, rather, in the way in which you conducted yourselves .... It’s the way in which you have conducted yourself, not the verdict itself.
*1006The jury then deliberated on the Consorti ease. In summation, plaintiffs’ counsel had suggested awards of $8 million for John Con-sorti’s past pain and suffering, plus $4 million for his future suffering. The jury returned a verdict that awarded exactly those amounts. The jury found John’s remaining life expectancy to be nine months. OCF was found 7% responsible for Consorti’s injuries.
OCF moved for judgment as a matter of law, a new trial, or a remittitur, all of which were denied. However, the court did conclude that the jury’s calculation of Frances Consorti’s future non-economic loss was against the weight of the evidence and must be set aside. Consorti et al., 847 F.Supp. at 1105. In lieu of a new trial, the parties stipulated to an award for non-economic consortium of $332,000. The district court also reduced the verdict on economic damages for Frances’s loss of future consortium to $750,-000. Id. at 1106.
After judgment molding, the district court entered judgment against OCF for more than $11.5 million. Id. at 1086. This appeal followed.
DISCUSSION
OCF appeals numerous aspects of the case below and asks that the judgment be set aside or that remittitur be ordered.
A. Consolidation
Relying on a flawed understanding of our holding in Malcolm v. Nat’l Gypsum Co., 995 F.2d 346 (2d Cir.1993), OCF contends that Judge Sweet erred in consolidating the cases of several plaintiffs into a single trial. OCF treats Malcolm as establishing a strong anti-consolidation bias. We take pains to emphasize that we have made no such suggestion. The reversal in Malcolm was attributable to factors unique to that ease. We stressed that “[w]e do not wish to be understood as condemning all consolidations of asbestos cases.” 995 F.2d at 354. In fact, after observing that the federal courts were “swamped with asbestos suits,” 995 F.2d at 348, and that the “ ‘heyday of individual adjudication of asbestos mass tort lawsuits has long passed,’ ” id. (quoting In re Asbestos Prods. Liab. Litig. (No. VI), 771 F.Supp. 415, 419 (J.P.M.L.1991)), we indicated that another consolidation of a substantially larger number of asbestos cases coming from the Brooklyn Navy Yard (which were not appealed on that ground) would have passed muster. 995 F.2d at 353.
Consolidation is a valuable and important tool of judicial administration. This is especially true when the courts are overwhelmed with huge numbers of eases which involve substantially the same questions of fact, as happens when large numbers of plaintiffs allege that they have developed similar illnesses in reaction to a particular toxic substance. See Fed.R.Civ.P. 42(a). In such circumstances, consolidation permits the federal court to furnish trials in hundreds, even thousands of cases it might otherwise not reach for many years. If carefully and properly administered, as it was by Judge Sweet below, consolidation is also capable of producing, with efficiency and greatly reduced expense for all parties, a fairer, more rational and evenhanded delivery of justice.
We noted in Malcolm that at that time more than 200,000 asbestos cases had been filed in state and federal courts, with as many as 250,000 additional cases expected to come. 995 F.2d at 348. It requires little imagination to recognize that without consolidation the courts are simply incapable of handling litigation of such volume. The waste of time and expense involved in empaneling separate juries to decide the same sorts of questions over and over again is staggering. This is all the more true when one recognizes that each successive jury must be educated by expert witnesses to understand the toxicity of asbestos fibers, the etiology of asbestos-induced diseases, the state of the art regarding the industry’s knowledge of these dangers through the years, and the economic issues involving loss of services and future income that recur so frequently in these cases.
Needless to say, efficiency cannot be permitted to prevail at the expense of justice. Id. at 350. The obligation of the courts to deliver justice is paramount, and it may not be scrapped for the benefit of cheaper and more rapid dispositions. Johnson v. Celotex *1007Corp., 899 F.2d 1281, 1285 (2d Cir.), cert. denied, 498 U.S. 920, 111 S.Ct. 297, 112 L.Ed.2d 250 (1990). On the other hand, the judiciary, like every other institution, must be open to discarding habits that have outlived their usefulness, and must bend under the pressures of modern life to find greater efficiency in accomplishing its mission. When enormous savings of expense and gains of efficiency can be accomplished without sacrifice of justice, courts must be ready to alter their procedures. See American Law Institute, Complex Litigation: Statutory Recommendations and Analysis 16-18 (1994).
This is all the more true when increases in efficiency also improve the quality of the justice rendered. When cases are properly chosen for consolidation and well administered, this will be the case. One of the most persistent and troublesome problems in the administration of justice in our civil jury system is the unpredictable relationship between different juries’ awards, particularly for intangibles such as pain and suffering. It of course should be the goal of the overall administration of such litigation that more seriously injured plaintiffs receive higher compensation than those less seriously injured. However, when each case is tried before a different jury, the relationship between the size of one judgment for intangibles and another will be largely happenstance. See generally James F. Blumstein et al., Beyond Tort Reform: Developing Better Tools for Assessing Damages for Personal Injury, 8 Yale J. on Reg. 171, 177 (1991). When numerous claims are tried before a single jury, that jury will recognize that an important part of its chore is to scale the relative seriousness of the various plaintiffs’ injuries and to see to it that their respective awards are consistent with that sealing.
Different juries also can produce inconsistent verdicts on liability and the extent of responsibility. This may be encouraged by counsel’s litigation strategy. When only a small segment of a much larger picture is before the jury, counsel can present a deceptive view of the relative responsibility of various actors — especially entities that are absent from the trial. This becomes far more difficult when the jury is considering a broader spectrum of information and more of the relevant actors are present defending their positions. See Thomas D. Rowe, Jr. & Kenneth D. Sibley, Beyond Diversity: Federal Multiparty, Multiforum Jurisdiction, 135 U.Pa.L.Rev. 7, 15 (1986) (“Beyond the sheer economy of not having to litigate the same matters twice, consolidation ... can reduce such problems as inconsistent outcomes, whip-sawing (from the ability of defendants in separate litigations to point to a nonparty as the one truly hable), and uncoordinated scrambles for the assets of a limited fund.”) (footnotes omitted).
Furthermore, no logic supports the proposition that the incremental addition of similar cases will reduce the jury’s ability to understand and resolve the issues placed before it. Without doubt, consideration of a single toxic tort case is challenging, requiring jurors to grapple with complicated issues of chemistry and medicine. But it does not follow that the jury will be less able to deal with those issues if the same questions are repeatedly put to it over a substantially longer period of time. Quite to the contrary, if a jury spends many weeks, or many months, considering numerous eases of asbestos disease, and repetitively hears the disputes of experts and the arguments of counsel on case after case, that jury is likely to develop a far deeper understanding of the issues than a jury whose exposure to those complicated questions is brief, and requires answering only a single set of questions.
Malcolm suggests nothing to the contrary. The majority in Malcolm, over Judge Walker’s dissent, found various case-specific problems that led it to the conclusion that justice had been sacrificed. A principal problem identified in Malcolm was that the very basis of consolidation was found to be inapplicable to the case. The consolidation was premised on the supposed fact that each plaintiff had spent more than 50% of his work history at one or more of 40-odd power plant construction sites in New York State. Malcolm, 995 F.2d at 351. At trial, however, it emerged that several plaintiffs had spent only small fractions of their exposure history at power plant sites, and the jury in fact had before it *1008evidence of over 250 work sites with varying product mixes. Id. Another problem was that 13 days before trial, over 200 companies had been impleaded as third-party defendants, and some of them in turn impleaded fourth-party defendants. Id. at 348. The court expressly queried whether these late-added parties had adequate opportunity to prepare for trial. Id. at 352. The most critical problem undermining the Malcolm consolidated verdict was the majority’s perception that, instead of apportioning responsibility on the basis of the evidence, the jury had simply “throw[n] up its hands,” id., and apportioned “an equal 9% liability to each defendant,” id., in spite of the fact that “the evidence regarding [plaintiffs’] exposure to [the defendant’s] products was vague, minimal, and heavily circumstantial when compared to the extensive evidence regarding the products of [another defendant].... ” Id.
OCF argues that some of the factors we listed in Johnson, 899 F.2d at 1285, and Malcolm, 995 F.2d at 350-51, as helpful in analyzing the propriety of consolidation, weigh against consolidation here. Prior to trial, Judge Sweet entertained a reconsideration of consolidation in light of our newly-issued Malcolm decision. He reviewed those factors carefully and adhered to his decision that consolidation was proper. New York Asbestos II, 149 F.R.D. at 499. It is important to remember that factors noted by a court as helpful in analyzing a particular question are no more than that; they are not a substitute for answering the question itself. When the issue for an appellate court is whether the trial court’s decision to consolidate exceeded its discretion, the question remains whether the consolidation caused such confusion or prejudice as to render the jury incapable of finding the facts on the basis of the evidence. It is clear the consolidation here had no such consequence.
Furthermore, we have noted repeatedly that a district court can greatly assist a jury in comprehending complex evidence through the use of intelligent management devices. See, e.g., Johnson, 899 F.2d at 1285; Malcolm, 995 F.2d at 349, 352-53. Such management devices include organizing evidence by topic, using charts and visual aids, allowing note-taking by jurors, furnishing the jury with notebooks and albums of pertinent exhibits structured in a manner to help it master complex materials, interim explanations by the judge on issues of law and fact and on the limited use of evidence, interim addresses to the jury by counsel,1 and questionnaires and special verdict forms to help the jury approach deliberations in a well-organized fashion.
Judge Sweet took numerous such steps to help the jury segregate the issues before it and distinguish among the four sets of plaintiffs and the various defendants. The jury was provided specialized notebooks with a photograph of each plaintiff, accompanied by undisputed biographical information. Jurors were encouraged to take extensive notes during trial. During opening statements and summations, counsel used charts to help the jury distinguish among the plaintiffs’ exposure histories, together with time lines to help the jury differentiate between relevant periods in the state of the art. Counsel also provided citations to the trial record to help guide the jury to arguments regarding specific evidence. Consorti, et al., 847 F.Supp. at 1093. Throughout the trial, Judge Sweet gave numerous cautionary and limiting instructions to assist the jury in its understanding of the bearing of the evidence on the various cases. The verdict forms which Judge Sweet prepared guided the jury step by step through the various issues it needed to consider and resolve. The jury was encouraged to consider each case separately.
By implementing such measures, Judge Sweet insured that plaintiffs and defendants received a trial by jury that fairly addressed the individual claims while effectively managing the resources of the court and giving the parties the benefit of an efficient and economical trial. We have no reason to believe that the consolidation prevented the jury from rendering verdicts based on the evidence as it related to each independent claim. OCF’s contention of improper consolidation is without merit.
*1009B. Excessiveness of Award for Pain and Suffering
OCF claims that the award to John Con-sorti of $12 million for his pain and suffering was excessive and should be set aside. We agree.
1. The Role of Judge and Jury
We note at the start, our finding in no way deprecates the enormous suffering that John Consorti endured. He developed a tumor which gradually enveloped his spine. The tumor pressed against his vocal cords, causing him to lose his voice and choke. It interfered with eating, swallowing, even breathing. His circulatory system was impaired, causing painful and disfiguring swelling of his head and neck. And, as his disease progressed, it became increasingly difficult (later impossible) for him to walk or to care for himself. His pain grew worse as time passed, and was, of course, deepened by the certainty of imminent death. We take it as a given that reasonable people of his age, in good mental and physical health, would not have traded one-quarter of his suffering for a hundred million dollars, much less twelve.
It does not follow that courts should permit a verdict of a hundred million dollars, or twelve million, to stand. While the law seeks by reasonable compensation to make a plaintiff whole, we must recognize that compensation for suffering can be accomplished only in a symbolic and arbitrary fashion. There are at least two serious shortcomings to the endeavor. First, money awards do not make one whole; they do not alleviate pain. Second, there is no rational scale that justifies the award of any particular amount, as opposed to some very different amount, in compensation for a particular quantum of pain. See Gibbs v. United States, 599 F.2d 36, 39 (2d Cir.1979) (“measuring pain and suffering in dollars is inescapably subjective”); 2 American Law Institute, Reporters’ Study-Enterprise Responsibility for Personal Injury 201-02 (1991).2
We acknowledge furthermore that a judge’s training in law gives the judge no greater ability than a jury to determine the dollar amount that appropriately compensates any particular level of suffering. See 2 American Law Institute, supra, at 202. This recognition can lead judges to abdicate responsibility for review of jury verdicts. As we have no greater knowledge than jurors of the amount of money that suitably compensates for suffering, the question arises why our assessment should take precedence over theirs?
Nonetheless, there are important reasons why courts cannot properly leave it to juries to set the limits of compensation for such injuries. Even where the law is incapable of furnishing a rational answer, it seeks at least to be evenhanded, fair, and predictable. It should be our goal that persons who endure a similar degree of suffering can expect to receive a roughly similar award of compensation, see Nairn v. Nat’l R.R. Passenger Corp., 837 F.2d 565, 568 (2d Cir.1988) (vacating a damages award as excessive in light of awards for similar injuries in other eases), and that similarly situated defendants be burdened by similar judgments. If each jury is given unbridled authority to set the level of damages, awards will vary widely and unpredietably. Great discrepancies in awards destroy the fairness of the judicial system, as well as the predictability of litigation. This vigilance must be directed not only against awards that are too high, but also against those that are too low. See, e.g., In re Brooklyn Navy Yard Asbestos Litig., 971 F.2d 831, 853-54 (2d Cir.1992) (ordering new trial on issue of pain and suffering where award of $25,000 was so low, given evidence of lengthy and intense suffering, as to shock the conscience). Such goals cannot be achieved unless courts exercise responsibility to keep jury awards within limits. Cf. Restatement (Second) of Torts § 901 (1977) (note) (describing rules for determining damage awards); Fowler V. Harper et al., The Law of Torts 250-51 (2d ed. 1985) (describing judge’s role in fashioning tort damages).
*1010The unpredictability of jury awards is pernicious not only because it is unfair. Widely varying jury verdicts make it difficult for risk bearers to structure their behavior to efficiently manage risk. While it is an aim of tort law to deter wrongful conduct, see Restatement (Second) of Torts § 901 (1977), at the same time tort law seeks to avoid over-deterrence that would stifle socially valuable enterprise.
Where liability costs are relatively predictable, they can be avoided (where it is efficient to do so) or ‘built in’ to the costs of goods and services.... But errors in valuation may cause over-deterrence — the taking of too many costly precautions, or withdrawal from risky activity altogether. For example, during the liability insurance crisis of the 1980s, many obstetricians reportedly stopped delivering babies, and some manufacturers ceased development or production of certain drugs and goods.
Randall R. Bovbjerg et al., Valuing Life and Limb in Tort: Scheduling “Pain and Suffering,” 83 Nw.U.L.Rev. 908, 925 (1989) (citations omitted). See also Peter H. Sehuck, Agent Orange on Trial: Mass Toxic Disasters in the Courts 289 (1987) (arguing that “individual juries acting in isolated tort cases ... seem especially prone to overdeter some risky activities,” thereby “discourag[ing] the taking of risks from which society might well benefit”); Peter H. Sehuck, Introduction to Tort Law and the Public Interest 31-36 (Peter H. Sehuck ed. 1991) (describing empirical research); 2 American Law Institute, supra, at 202-03; Oscar G. Chase, Helping Jurors Determine Pain and Suffering Awards, 23 Hofstra L.Rev. 763, 769 (1995). When courts fail to exercise the responsibility to curb excessive Verdicts, the effects are uncertainty and an upward spiral. One excessive verdict, permitted to stand, becomes precedent for another still larger one.3 Unbridled, spiraling, excessive judgments predictably impose huge costs on society.
A failure by courts to impose limits on jury verdicts would cause serious social dislocation. At the first level, unchecked costs attributable to tort liability, and resultant increases in insurance premiums, would inevitably raise the price of goods and services to the public. More serious exaggeration in unchecked jury awards can cause bankruptcies in productive enterprises, with consequent disappearance of jobs, and even bankruptcies among insurers, leaving segments of society unprotected. Finally, in the circumstance represented by asbestos litigation, where it appears virtually certain that the resources of major defendants will eventually be completely consumed by their liabilities, allowing excessive awards to stand in the early stages of the litigation of the many thousands of cases they face will mean that over-compensation of the early plaintiffs will leave insufficient resources for the equally deserving plaintiffs whose cases are heard later.4
For all these reasons, even if courts have no better ability than jurors to measure the dollar value of pain, they must accept the responsibility of controlling the limits of jury awards. See generally David W. Leebron, Final Moments: Damages for Pain and Suffering Prior to Death, 64 N.Y.U.L.Rev. 256 (1989) (advocating increased judicial supervision of jury awards for pain and suffering prior to death).5
2. The Award for Consorti’s Pain and Suffering
The first step in deciding whether the award of $12 million for John Consorti’s pain and suffering is excessive must be to ascertain whether federal or state rules apply. *1011The question has significance in this case because the New York rule, discussed below, is less deferential to the jury’s verdict than the federal standard of whether the award was so excessive as to shock the conscience of the court. Scala v. Moore McCormack Lines, Inc., 985 F.2d 680, 683 & n. 7 (2d Cir.1993); see, e.g., Nairn v. National R.R. Passenger Corp., 837 F.2d 565, 567 (2d Cir.1988) (applying federal standard in FELA case). We held in Martell v. Boardwalk Enters., Inc., 748 F.2d 740, 750 (2d Cir.1984), that in a federal court diversity suit state law governs the issue of excessiveness of a jury verdict on a motion for remittitur. Plaintiffs contend that decision was inconsistent with the Supreme Court’s ruling in Donovan v. Penn Shipping Co., 429 U.S. 648, 97 S.Ct. 835, 51 L.Ed.2d 112 (1977). Accordingly, we review the question.
Since the landmark decision of Erie R.R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), it has been a hallmark of federalism that a federal court sitting in diversity, while using its own rules to govern procedural matters, will apply the substantive law of the forum state. The application of this standard to questions of remittitur has caused considerable confusion. See In re Air Crash Disaster Near New Orleans, 767 F.2d 1151, 1155 & n. 4 (5th Cir.1985) (noting “apparent conflict in the cases concerning whether the excessiveness of a verdict is a matter governed by state or federal law”); 19 Charles A. Wright, Arthur R. Miller and Edward H. Cooper, Federal Practice and Procedure § 4511 n. 23 (1982 & Supp.1995).6 This is perhaps because courts have not always recognized that a question of remittitur arising in a particular case can involve either procedural or substantive issues, or both.
The Supreme Court’s decision in Donovan involved a procedural question—whether a plaintiff who has accepted a district court’s offer of remittitur may nonetheless appeal its determination that the jury’s verdict was excessive. The Supreme Court unsurprisingly ruled that such a question arising in a ease heard in the federal diversity jurisdiction would be governed by federal law. In delivering its ruling, however, the Court made an extraordinarily broad statement that “[t]he proper role of the trial and appellate courts in the federal system in'reviewing the size of jury verdicts is ... a matter of federal law_” Donovan, 429 U.S. at 649, 97 S.Ct. at 837.7 Citing this language, plaintiffs contend that all remittitur questions are to be governed by federal rules. We do not agree.
The question we face is how much money John Consorti may be awarded in damages by reason of his pain and suffering. This is an issue of substantive rights under the laws of New York. If, for example, New York’s statutes provided that no compensation was to be awarded for nonpecuniary loss, or scheduled the awardable amounts in the manner of workers’ compensation statutes—i.e., $X for injury to an arm, $Y for injury to a leg, $Z maximum for pain and suffering—federal courts ruling in diversity would unquestionably be bound by those substantive rules. See Blasky v. Wheatley Trucking, Inc., 482 F.2d 497, 498-99 (6th Cir.1973) (Ohio statute prohibiting award of nonpecuniary damages is controlling substantive state law). And if a jury’s award exceeded those standards, the federal court would be required to reduce the verdict to bring it into compliance with the substantive law of New York. Here, New York law provides that a jury verdict may not exceed that amount which would “deviate[ ] materially from ... reasonable compensation.” N.Y.Civ.Prac.L. & R. § 5501(c) (McKinney Supp.1995). This is the substantive rule provided by New York law. To disregard New York law in this case, in favor of the federal “shocks the conscience” standard, would violate the fundamental constitutional doctrine *1012of Erie, 304 U.S. at 78, 58 S.Ct. at 822.8
This is substantially clarified by a more recent decision of the Supreme Court. Browning-Ferris Indus. v. Kelco Disposal, Inc., 492 U.S. 257, 109 S.Ct. 2909, 106 L.Ed.2d 219 (1989), involved a claim that a federal district court sitting in diversity had improperly failed to limit the size of a punitive damages award through remittitur. In that case, the Court reaffirmed the importance of applying federal procedural standards, but then made clear that “the propriety of an awárd of punitive damages ... [is a] question[ ] of state law.” Id. at 278, 109 S.Ct. at 2922. The Court expressly refused to “craft some common-law standard of ex-eessiveness” since “these are matters of state, and not federal, common law.” Id. at 279, 109 S.Ct. at 2922. As here, “[a]dopting a rule along the lines petitioners suggest would require us to ignore the distinction between the state-law and federal-law issues.” Id. Hence, the Browning-Ferns Court approved the district court judge’s application of “the proper state-law standard in considering whether the verdict returned was excessive.” Id. at 280, 109 S.Ct. at 2922.
For purposes of deciding whether state or federal law is applicable, the question whether an award of compensatory damages exceeds what is permitted by law is not materially different from the question whether an award of punitive damages exceeds what is permitted by law. Browning-Ferris made clear that the excessiveness.of punitive damages is governed by state, and not federal, law. The state’s rule must also determine the excessiveness of compensatory damages.9 Accordingly, we decided in Martell, 748 F.2d at 750, that, when conducting remittitur review as a court sitting in diversity, we would seek the governing standard in state court decisions. We therefore reject the plaintiffs’ argument and look to the law of New York to determine whether the award was excessive. Like a New York court, we will “look ... to other jury awards condoned by the courts of’ New York, recognizing that “New York appellate courts regard prior awards as not binding but instructive. E.g., Senko v. Fonda, 53 A.D.2d 638, 639, 384 N.Y.S.2d 849, 851 (2d Dep’t 1976) (prior awards ‘may guide and enlighten the court[s]’ ‘in the exercise of their discretion.’).” Martell, 748 F.2d at 750.
The standard for determining excessiveness and the appropriateness of remittitur in New York is somewhat ambiguous. Prior to 1986, New York law employed the same standard as the federal courts, see Matthews v. CTI Container Transport Int’l Inc., 871 F.2d 270, 278 (2d Cir.1989), which authorized re-mittitur only if the jury’s verdict was so excessive that it “shocked the conscience of the court.” See, e.g., Neal v. Rainbow House Fruits, 87 A.D.2d 511, 447 N.Y.S.2d 487, 488 (1982); Juiditta v. Bethlehem Steel Corp., 75 A.D.2d 126, 428 N.Y.S.2d 535, 543 (1980); Petosa v. City of New York, Dep’t of Sanitation, 63 A.D.2d 1016, 406 N.Y.S.2d 354, 355 *1013(1978). In 1986, New York State legislated a new standard that required the Appellate Division of the Supreme Court to exercise closer control over jury awards. The new CPLR § 5501(c) provides,
In reviewing a money judgment in an action in which an itemized verdict is required ... in which it is -contended that the award is excessive or inadequate and that a new trial should have been granted unless a stipulation is entered to a different award, the appellate division shall determine that an award is excessive or inadequate if it deviates materially from what would be reasonable compensation.
N.Y.Civ.Prae.L. & R. § 5501(c) (McKinney Supp.1995) (emphasis added). This standard requires the reviewing court to determine the range it regards as reasonable, and to determine whether the particular jury award deviates materially from that range, taking corrective action if it does. Material deviation from reasonableness is less than that deviation required to find an award so excessive as to “shock the conscience.”10
Because this statute makes explicit mention only of the Appellate Division, courts have expressed some confusion as to whether its standard applies also at the trial level. See, e.g., In re Joint E. & S. Dists. Asbestos Litig., 798 F.Supp. 925, 936-37 (E. & S.D.N.Y.1992) (“McPadden et al.”), rev’d on other grounds, 995 F.2d 346 (2d Cir.), and rev’d on other grounds sub nom. Malcolm v. Nat'l Gypsum Co., 995 F.2d 346 (2d Cir.1993). See also David D. Siegel, Supplementary Practice Commentary, C5501:10, reprinted in McKinney’s Consolidated Laws of New York Annotated, Vol. 7B at 5 (1995 Supp.). We need not resolve whether the “deviates materially” standard applies to trial courts as well as to the Appellate Division. Unquestionably, it provides the standard for appellate review and must therefore be applied by this court, regardless whether it also governed the district court sitting as a New York trial court.
Because the standard of excessiveness is furnished by state law, the best guide for a federal court, if available, would be decisions of the state’s courts in comparable cases indicating at what point awards become excessive. See Martell, 748 F.2d at 750. Although a number of pertinent New York State judgments on damages for pain caused by mesothelioma were available, the district court relied instead on McPadden, a recent federal trial court decision. Consorti et al., 847 F.Supp. at 1096-97. In McPadden, the district court upheld a jury award of $4.5 million for pain and suffering due to mesothelioma. McPadden et al., 798 F.Supp. at 937-38.
In discussing below the question of excessiveness of the jury’s award to Tabolt of $7.5 million for pain and suffering, Judge Sweet converted the $4.5 million McPadden award to a monthly figure of $409,090.91 for McPadden’s 11 months of suffering. Multiplication of that figure by' Tabolt’s 18 months of suffering produced a figure of $7,363,-636.36, which was not substantially lower than the jury’s award to Tabolt. Consorti et al., 847 F.Supp. at 1097. Although it was not expressly discussed, the district court presumably used the same conversion of the McPadden award to a monthly amount to justify the jury’s $12 million award to Consorti, whose disease the jury predicted to run 32 months from onset to death. Twelve million expressed as a monthly figure over 32 months is $375,000, which is below the $409,-090.91 monthly figure Judge Sweet extrapolated from the McPadden judgment.
We disagree with the district court’s use of McPadden to justify the jury’s award to Consorti for two reasons. First, rigid application of a monthly multiplier for diseases of significantly different durations is an oversimplification, as well as inconsistent with New York precedent.11 Unquestionably, the *1014duration of suffering is a relevant factor. But it cannot be automatically assumed that suffering which lasts 30 months, from onset of illness to death, should command a verdict three times that of a case where ten months pass from onset to death. In many cases, the degree of average daily suffering will be greater in a ease where the disease progresses rapidly. The fact that McPadden’s disease progressed three times more rapidly than Consorti’s does not necessarily justify an award for Consorti nearly three times the size of McPadden’s. Moreover, we believe the court relied excessively on federal district court precedent to justify the jury verdict, giving insufficient regard to a large number of recent New York State court decisions which strongly indicated that the Consorti award was far beyond what New York law permits.
Justice Helen Freedman is the trial judge to whom all New York asbestos litigation has been assigned by administrative order. In the period shortly prior to the Consorti trial, Justice Freedman rendered numerous judgments reducing awards in comparable eases of mesothelioma to levels far below what the jury awarded to Consorti. In Didner v. Keene Corp., Justice Freedman reduced a $3.25 million award to $2.3 million; the plaintiff had suffered 26-27 months of pain and suffering due to mesothelioma. Didner v. Keene Corp., N.Y.L.J., Jan. 4, 1991, at 22 (N.Y.Sup.Ct. Dec. 17, 1990), aff'd, 188 A.D.2d 15, 593 N.Y.S.2d 238 (App.Div.1993) (no discussion of remittitur), modified, 82 N.Y.2d 342, 604 N.Y.S.2d 884, 624 N.E.2d 979 (1993). The court concluded that while Didner’s pain and suffering was severe, “courts in this jurisdiction have not countenanced awards of this size for two plus years of pain and suffering.” Id.
Four months later, in the consolidated Brooklyn Navy Yard litigation, Justice Freedman reduced fourteen pain and suffering awards in mesothelioma death cases. In re New York City Asbestos Litig., Index No. 40,000/88, slip. op. at 3-5 (N.Y.Sup.Ct. May 13,1991), aff'd in part, rev’d in part on other grounds, 188 A.D.2d 214, 593 N.Y.S.2d 43 (App.Div.1993) (no discussion of remittitur), aff'd, 82 N.Y.2d 821, 605 N.Y.S.2d 3, 625 N.E.2d 588 (N.Y.1993).12 The jury awards for pain and suffering from mesothelioma that Justice Freedman considered ranged from a low of $1 million, for a course of disease that took only three months, to a high of $5 million, in a ease of 34 months of suffering. The jury awards thus ranged from 8% to 42% of this jury’s award to Consorti. Justice Freedman found all those awards excessive under New York law. Those most comparable in duration to Con-sorti’s were two awards of $5 million in cases where the suffering lasted 34 and 35 months 0Cohen and Lopez). Justice Freedman reduced these to $3 million. And as to three awards ranging from $3.2 to $4.25 million, for suffering periods of 26 to 29 months, they were reduced to $2.2 to $2.4 million (Chierico, Savella, and Silver).
While it is difficult, especially without a study of the complete record, to evaluate the pertinence of one judgment for pain and suffering for another, the strong pattern of Justice Freedman’s rulings is helpful in evaluating the Consorti awards. In the face of so substantial a pattern of New York judgments communicating that New York courts find excessiveness in such cases at far lower levels, we believe it was error for the district court to conclude that the jury’s $12 million award was within the range accepted by New York law.13 We note furthermore that Justice Freedman employed the “shockingly excessive” standard, while we on appeal must use the standard of CPLR § 5501(c) — “deviates materially from ... reasonable eompen-*1015sation” — a standard less deferential to the jury’s appraisal.14
Although we recognize that John Consor-ti’s suffering was very great, Justice Freedman’s eases also all involved deaths from mesothelioma. It is not suggested to us that the suffering of the plaintiffs she considered was any less horrendous.15 We recognize that the New York decisions in question were at the trial court level, and thus less compelling as authority than a decision of New York’s Court of Appeals. Nonetheless, especially given the uniformity of the large number of decisions, they provide a better indication of New York law than any other authorities cited to us.16
Given the consistent pattern, over a considerable number of cases of death by meso-thelioma, of reduction of jury verdicts for pain and suffering by New York courts to figures ranging from approximately $1 million to a maximum of $3 million, we conclude an award exceeding $3.5 million would deviate materially from what is deemed reasonable under New York law. Accordingly, we direct entry of a conditional order of a new trial on the issue of damages for pain and *1016suffering, unless plaintiffs accept remittitur to that amount.
C.Counsel’s Recommendation of an Award
OCF claims that error occurred during summation, when plaintiffs’ counsel made specific dollar recommendations to the jury as to appropriate awards for pain and suffering. He suggested $8 million for John Con-sortes pain and suffering to date and $4 million for future pain and suffering. He also suggested a specific figure for Frances Consorti’s award. After the jury left the room, defense counsel objected to plaintiffs’ counsel’s request.
OCF’s protest is undermined by the fact that its own counsel was the first to suggest a value for Consortes pain and suffering. Earlier in the summations, counsel for OCF told the jury:
And I feel that to make a suggestion to you that he’s not entitled to any compensation would not be fair.... [if you decide] that he was entitled to maybe $750,000 to maybe even a million dollars, I don’t think that that would be an unfair figure.
While this court has not adopted a per se rule about the propriety of suggested damage amounts, see McPadden et al., 798 F.Supp. at 936, we wish to emphasize that specifying target amounts for the jury to award is disfavored. Such suggestions anchor the jurors’ expectations of a fair award at a place set by counsel, rather than by the evidence. See Mileski v. Long Island R.R. Co., 499 F.2d 1169, 1172 (2d Cir.1974) (“A jury with little or no experience in such matters, rather than rely upon its own estimates and reasoning, may give undue weight to the figures advanced by plaintiffs counsel -”). A jury is likely to infer that counsel’s choice of a particular number is backed by some authority or legal precedent. Specific proposals have a real potential to sway the jury unduly. While under the circumstances present here we do not yet reach the point of holding that it is error to permit such recommendations, it is not a desirable practice. We encourage trial judges to bar such recommendations.
D. Comment by Judge
As outlined above, after the jury returned its first verdict, the judge commented, among other things, that “you are probably the best jury that has ever been put together in the United States of America.” After counsel objected, the judge gave a curative instruction. He told the jury that he had intended to praise “the way in which you have conducted yourself, not the verdict itself,” and that he had no stake in their verdict.
OCF argues that the judge’s comments impermissibly communicated to the jury his approval of their verdict. Judge Sweet immediately recognized that, coming on the heels of the generous verdict for Tabolt, his encouragement of the jury might be misunderstood as endorsement of the generosity of its verdict. His prompt corrective instruction cured the problem. There is no basis for a mistrial.
E. Other Claims
OCF asserts a number of other claims. OCF argues that the district court erred in failing to order a new trial due to the jury’s receipt of extra-judicial information. A month after the jury returned its verdicts, the court received a letter from juror Carl D’Andrilli that said that during deliberations, another juror had mentioned “that Johns-Manville, a non-party defendant, possesses a one billion dollar fund to pay lawyer’s fees to fight asbestos litigation cases.” The district court ruled that any well-informed citizen could possess general knowledge about Johns-Manville and that the report from the juror was not prejudicial to the defendants. This ruling was certainly within the trial court’s discretion.
The district court was also within its discretion in excluding the summary evidence proffered by OCF. OCF attempted to use summary evidence — analyzing Veteran and State invoices using a computer list of asbestos-containing products — to demonstrate that the Consorti family companies had purchased few asbestos-containing products from OCF. OCF did not introduce the computerized list or make it available in court or to counsel. *1017The district court ruled that: “In the absence of the list and any testimony concerning its preparation, the summary based upon it became inadmissible.” We agree. The records may not have comprised the universe of Veteran documents, and the reason for the exclusion of certain invoices from the summaries was unclear.
We reject OCF’s claims that the evidence was insufficient to support the jury’s findings that OCF acted recklessly and in concert with other defendants. The record contains sufficient evidence to support the jury’s determination.
CONCLUSION
We have considered OCF’s other claims and find them to be without merit. Based on the above, the verdict for John Consorti’s pain and suffering will be set aside and a new trial ordered unless plaintiff-appellee agrees to accept an award of $3.5 million. The award for loss of consortium is vacated. The judgment of the district court is affirmed in all other respects.
. See Pierre N. Leval, From the Bench: West-moreland v. CBS, Litigation, Fall 1985, at 7, 66.
. Juries may be especially inclined to make substantial awards when they believe that the defendant’s liability is covered by insurance. While insurance frequently does cover the liability, jurors may neglect to consider the broader effect of their award on insurance premiums. See W. Page Keeton et al., Prosser & Keeton on the Law of Torts 591 (5th ed. 1984).
. See discussion below at n. 13.
. For example, Keene Corporation informed the trial court that it had already spent over $430 million to resolve 95,000 claims, with 98,000 others still pending; it had no further insurance coverage, and would "run out of funds long before all of the present claimants have been even partially compensated." Keene advised the court, "[T]here is no hope that those who filed late claims or who file claims in the future will receive any compensation." New York Asbestos II, 149 F.R.D. at 499-500.
.We note that in the punitive damages context, the Supreme Court has held the availabilily of judicial review over jury awards to be constitutionally required. See Honda Motor Co., Ltd. v. Oberg, - U.S. -, -, 114 S.Ct. 2331, 2340-41, 129 L.Ed.2d 336 (1994).
. Our court has twice found it unnecessary to resolve this question in the context of the New York remittitur statute. Tyler v. Bethlehem Steel Corp., 958 F.2d 1176, 1190 (2d Cir.), cert. denied, 506 U.S. 826, 113 S.Ct. 82, 121 L.Ed.2d 46 (1992); Gardner v. Federated Dep’t Stores, Inc., 907 F.2d 1348, 1353 n. 1 (2d Cir.1990).
. This court, too, has made confusingly broad categorical statements on this subject in the past. See West v. Jutras, 456 F.2d 1222, 1225 n. 6 (2d Cir.1972); Karlson v. 305 East 43rd Street Corp., 370 F.2d 467, 472 n. 1 (2d Cir.), cert. denied, 387 U.S. 905, 87 S.Ct. 1690, 18 L.Ed.2d 625 (1967).
. We recognize that, in cases where state and federal substantive standards differ, some courts have chosen to follow federal law. See, e.g., Nodak Oil Co. v. Mobil Oil Corp., 533 F.2d 401, 410 (8th Cir.1976). For the reasons stated in the text, we decline to follow this approach. No doubt it will sometimes be difficult to distinguish between procedural and substantive remittitur rules. Consider, for example, American Business Interiors, Inc. v. Haworth, Inc., 798 F.2d 1135, 1146 (8th Cir.1986), which applied federal remit-titur rules in a diversity case where the forum state had abolished remittitur. It is difficult in the case of a state law doing away with remittitur to determine whether the rule is procedural or substantive. Such a rule presents itself as a procedural bar, but it has significant substantive ramifications. In contrast, New York's “deviates materially” standard is quite clearly a substantive rule governing the allowable size of jury verdicts.
. We note that, notwithstanding Browning-Ferris, a few courts have continued to apply federal standards of excessiveness in reviewing punitive damages. See Mayer v. Gary Partners & Co. Ltd., 29 F.3d 330, 334 (7th Cir.1994); Mattison v. Dallas Carrier Corp., 947 F.2d 95, 99 (4th Cir.1991); The Post Office v. Portec, Inc., 913 F.2d 802, 809-10 (10th Cir.1990), vacated and remanded, 499 U.S. 915, 111 S.Ct. 1299, 113 L.Ed.2d 235 (1991), appeal dismissed per stipulation, 935 F.2d 1105 (10th Cir.1991). Some have avoided the issue by referring to both standards. See Latham v. Nickerson American Plant Breeders, 978 F.2d 1493, 1499 (8th Cir.1992), cert. denied, - U.S. -, 113 S.Ct. 3037, 125 L.Ed.2d 724 (1993). Other courts have applied state standards of excessiveness. See Hiltgen v. Sumrall, 47 F.3d 695, 704 (5th Cir.1995); American Employers Ins. Co. v. Southern Seeding Servs., Inc., 931 F.2d 1453, 1458 (11th Cir.1991). We do likewise.
. Decisions of New York State courts show that the two standards are different. See Harvey v. Mazal Amer. Partners, 79 N.Y.2d 218, 581 N.Y.S.2d 639, 643, 590 N.E.2d 224, 228 (1992) (remanding to Appellate Division for reconsideration of remittitur using proper “deviates materially” standard, where Appellate Division had referred to both standards); Shurgan v. Tedesco, 179 A.D.2d 805, 578 N.Y.S.2d 658, 659 (1992) (affirming trial court’s additur where court set aside jury award as materially deviating from what would be reasonable compensation).
. See discussion below at n. 15.
. Justice Freedman's published opinion at 151 Misc.2d 1, 572 N.Y.S.2d 1006, omits discussion of the specific remittiturs cited in the text.
. In McPadden, the district court observed that in the state court’s Brooklyn Navy Yard consolidation, "Justice Freedman allowed several multimillion dollar awards for pain and suffering....” McPadden et al., 798 F.Supp. at 938. Those awards, reducing $5 million verdicts to $3 million, initially served as authority justifying an award of $4.5 million to McPadden, which in turn, converted to a monthly average figure, was used below to justify this award of $12 million.
. Although we apply the New York “deviatés materially” standard, we note that Consorti’s award was also so excessive as to "shock the conscience.” Hence, we would remit this $12 million award for pain and suffering even if we were to follow federal law, although application of this standard might allow for a slightly higher award than the $3.5 million figure we have allowed.
. We note that one of the difficulties of remitti-tur review is the limited data available about the particular cases used for comparison. See David C. Baldus et al., Improving Judicial Oversight of Jury Damages Assessments, 80 Iowa L.Rev. 825 (1995); Blumstein et al., supra. But even if we approach the New York precedents on a monthly basis — the way Judge Sweet used the McPadden precedent — they offer no better support for this award. Probably in recognition that the course of the disease will involve varying degrees of suffering in different periods, and that a short course is likely to cause greater average daily suffering than one which is slow to develop, Justice Freedman's judgments, stated on a monthly basis, vary roughly inversely to duration. The plaintiffs who died (or were expected to) within a year of the disease's onset received monthly sums of $110,000 to $200,000, while those whose illness lingered from one and a half to three years received between $80,000 and $90,000 per month. This provides no authority for Consorti's monthly figure of $375,000.
. In their petition for rehearing, plaintiffs cite two new authorities in support of the jury's verdict. In Manning v. Georgia Pacific Corp., No. 102757/94 (N.Y.Sup.Ct. Feb. 3, 1995), Justice Freedman remitted a $5 million verdict for pain and suffering caused by mesothelioma to $4 million. The Manning decision does not alter our view of the problem. First, the reduction to $4 million does not differ significantly from our reduction to $3.5 million. Second, Justice Freedman, in raising the level of damages she considered acceptable, deferred to the very district court judgments (in McPadden and in this case) which we here criticize for insufficient deference to her earlier rulings.
Plaintiffs also point to a recent decision of New York's Appellate Division, 1st Department, In re New York County DES Litigation, 211 A.D.2d 500, 621 N.Y.S.2d 332 (1995). For several reasons, plaintiffs’ submissions do not convince us to change our views of the present case. First, the Appellate Division's ruling gives no information about the verdict being upheld. It simply states, "As to damages, we do not find that there has been a material deviation from what would be reasonable compensation ... for the unique, far-ranging and severe physical and psychological injuries suffered by these plaintiffs.” Id. 621 N.Y.S.2d ht 333 (citations omitted). The decision tells neither the amounts of the verdicts, nor what they were awarded for. Plaintiffs submit a newspaper article reporting the decision, which asserts that among the judgments affirmed were judgments for $12 million and $10 million. DES Jury Verdicts Affirmed on Appeal, N.Y.L.J., Jan. 19, 1995, at 1. Even this source, however, does not tell us how those judgments were composed or what elements they included. What we do know from the opinion of the First Department is that the reviewing court considered the injuries to involve "far-ranging and severe physical and psychological injuries.” New York County DES Litigation, 621 N.Y.S.2d at 333. Material from the -briefs in the case tells us that these awards were made to women of childbearing age who— as a result of DES exposure — contracted clear cell adenocarcinoma of the vagina and other diseases. Their illnesses required that they receive radical surgery, which eliminated their reproductive capacity. The elements of suffering submitted to the jury included severe sexual dysfunction, aborted pregnancy, fear and sadness relating to the fate of unborn children, disappointment of plans to conceive a family, as well as many years of past and future suffering from cancer. The First Department's opinion states that it considers those injuries to be “unique.”
In short, we have little guidance as to the exact nature of the verdicts upheld in the DES litigation, and the guidance we have tells us they involved quite different emotional considerations from those arising here. See Nairn, 837 F.2d at 568 (evaluating excessiveness by reference to other awards for similar injuries).