Friday, December 12, 2003

Numerators and Denominators

Math has broken out in the recently revived debate over the McDonald's coffee lawsuit. (The Curmudgeonly Clerk has links to everybody's contributions here.)

We start with CC's observation that McDonald's knew of 700 previous instances in which customers complained of coffee burns. Ted Frank responds that this figure is really just a numerator in search of a denominator. The denominator, says Frank, is in the tens of billions (of cups of McDonald's coffee served), and a risk of 1 in 24 million, says Frank, doesn't render a product unreasonably dangerous. CC responds that the point of the 700 prior incidents is not so much to calculate risk coefficients for McDonald's coffee, but to show that the plaintiff's experience in the infamous lawsuit was not unique, and that McDonald's was on notice of the problem.

It might be interesting for all sides to apply comparable reasoning internally, within the tort reform debate. Let's give it a try.

After trolling the internet for litigation horror stories, Newsweek was able to come up with some smallish number of examples that it was willing to consider egregious. Call this number N.

To Frank, N should again be a numerator in search of some denominator (call the denominator D). What might be a suitable one? That depends on the point to be made. If the point is that the American tort system handles (say) personal injury claims too permissively, then seemingly D should equal the number of p.i. cases brought in the United States over whatever time period N represents. Frank might protest that Newsweek collected only a partial sample, and that point would be valid, but our own point does not depend on Newsweek's sampling techniques, and would apply even if N were taken to be the total number of unduly plaintiff-friendly outcomes in all p.i. cases, whether Newsweek found them or not. (We are supposing, of course, that everyone has agreed on some metric for what should count as "unduly plaintiff-friendly outcomes"). Call such outcomes "false positives" (because the the judicial system has permitted a factfinder to diagnose a tortiously caused injury where none is supposed to have existed). If our concern is the risk of injustice faced by defendants in the p.i. system, then the figure of interest should arguably be the rate of false positives, or N/D. (Of course, so far, that's just one side of the picture. If we are concerned to gauge the overall justness of the p.i. system for all parties, and not just defendants, we should care about false negatives too -- as Frank would presumably agree. We should care, that is, about the frequency with which deserving plaintiffs are thwarted in seeking recovery.)

Meanwhile, however, perhaps N might be cited not so much for the purpose of calculating rates of injustice, but more to show that one particular episode of injustice is not unique, and that further investigation, and even remedial measures, may be warranted. Certainly that seems to be the spirit in which advocates of tort "reform" offer their examples. It's worth noticing, in this vein, that the Curmudgeonly Clerk has responded in just such a spirit, by attempting to evaluate whether cases like the McDonald's coffee litigation should be counted in N at all (much as one might imagine McDonald's responding to reports of recurrent coffee burns by determining whether it could or should have taken measures to prevent them).

There is an interesting parallelism, here, to recurring debates in the toxic tort context. Toxic tort defendants, and their experts, commonly want to emphasize the quantitatively low absolute risks allegedly associated with the relevant hazardous substance. Sometimes, they go so far as to claim that those quantitative risks are the only pertinent consideration. Toxic tort plaintiffs, by contrast, and their experts, may tend to feel that even seemingly low risks are unacceptable, where they eventuate in harm. And sometimes, they go so far as to claim that the quantitative risk should be ignored, in favor of their anecdotal evidence.

Are we wrong, or do the parties to the tort reform debate sometimes tend to invert their respective styles of toxic tort reasoning, when evaluations of the civil justice system come into play? Perhaps that claim is unfair, because the truth may be that both styles are often pertinent, in both contexts. We are, or should be, concerned with the numerical prevalence of good and bad outcomes. We don't want a high ratio of bad verdicts or hospitalized McDonald's customers. But we should also be concerned to do our best in the particular case. The goal for the tort system, confessedly unrealizable in practice, should be zero bad outcomes. The goal for McDonald's, equally unachievable in reality, should be zero scalded patrons. When we fall short of those goals, we should consider doing something about it -- in both instances.

Our wisdom does not extend to prescribing perfect remedies for the American tort system. But we do have a modest proposal for the fast-food problem. McDonald's should drop the pretense of serving "coffee," and stick to soft drinks.

Update: Our modest proposal doesn't seem unrealistic, in light of news that McDonald's Corporation now plans to give up on pizza and Mexican food.
Fed. R. Evid. 702: If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.