Courts around the world are increasingly considering whether liability should exist in various types of situations in which a plaintiff can prove that a defendant’s tortious conduct may have contributed to the plaintiff’s injury, but it is inherently impossible, given the nature of the situation, for the plaintiff to prove that the defendant’s tortious conduct actually contributed to the injury. The problematic nature of the causal issue is usually recognized when the probability of causation is not greater than 50 percent, with courts adopting different views, depending on the type of situation, on whether liability nevertheless is appropriate and, if so, whether liability should be full or only proportionate to the probability of causation. However, when the probability of causation is only slightly higher—greater than 50 percent—many courts do not view either causation or liability as being problematic. Indeed, under the commonly accepted version of the res ipsa loquitur doctrine, liability is assumed to be unproblematic even though the only indication of negligent conduct as well as causation is a mere 50+ percent ex ante statistical probability. The dramatic difference in treatment of situations that are identical except for a trivial difference in statistical probability is due to an unexamined assumption that the usual “preponderance of the evidence” or “balance of probabilities” burden of persuasion in civil cases merely requires proof of a 50+ percent statistical probability. This assumption, which is common among academics as well as courts but is rejected by courts when the statistical rather than case-specific nature of the probability is obvious, has led to inconsistent and incoherent treatment of normatively and descriptively analogous types of situations and even to erroneous denials of proof of causation and liability in some situations in which tortious causation clearly exists. The statistical probability interpretation of the burden of persuasion in civil cases is inconsistent with the traditional understanding of that burden, which instead requires the formation of a minimal degree of belief, based on evidence specific to the particular occasion, in the actual existence of the disputed fact in the particular situation. When the disputed fact is actual causation of injury, there must be a minimal belief that the causal law underlying the allegedly applicable causal generalization was fully instantiated on the particular occasion. General statistics cannot support such a belief; only concrete evidence specific to the particular situation can do so. Only when the burden of persuasion is correctly understood can many currently debated issues regarding the existence and scope of tort liability be properly understood and consistently resolved. When the various types of problematic situations are compared, it turns out that the market share liability principles adopted in Sindell v. Abbott Laboratories and, arguably, Hymowitz v. Eli Lilly & Co., which are highly controversial, are more defensible than the liability principles that are widely employed in the alternative causation cases, the medical malpractice lost-chance cases, the toxic tort cases, and, especially, in the usual formulation of the res ipsa loquitur doctrine.
Richard W. Wright,
Liability for Possible Wrongs: Causation, Statistical Probability and the Burden of Proof, in Symposium, The Frontiers of Tort Law,
Loy. L.A. L. Rev.
Available at: http://scholarship.kentlaw.iit.edu/fac_schol/702