Moral Luck, Loss of Chance, and Legal Liability

Download data and study materials from OSF

Principal investigators:

John Darley

Princeton University



Lawrence Solan

Brooklyn Law School


Homepage: Click here

Sample size: 373

Field period: 08/17/2005-08/24/2005


This proposal requests support for a study to investigate people’s judgments about an area of law that is rather unstable: When an individual acts badly and puts another person at risk, should compensation be awarded even if the aggrieved party never actually suffers a physical injury? The law provides compensation for injury. In these cases, the only injury, if there is one, is a reduction in the probability that an individual will live a healthy life in the future. If this “loss of chance” is considered an injury, then people who do not later become ill may be overcompensated. If the loss of chance is not itself considered an injury, then two wrongdoers who misbehave in exactly the same way will be treated very differently by the law, depending on whether the victim experiences physical harm. This creates a dilemma that is difficult to resolve.
Neither legal scholars, nor judges nor lawyers are in accord about this issue (Nagareda, 1998). The proposed study, which is the first in a series of studies the investigators have designed, aims to use this puzzle to address a number of legal and policy questions that might profitably be informed by advances in social psychology and moral philosophy.


We have designed a series of studies organized around scenarios that vary systematically as to whether a victim actually suffered physical injury or whether he merely suffered an increased probability of suffering harm in the future.

Experimental Manipulations

This study asks subjects to react to a story in which fumes, which are released from a factory that manufactures glue, cause an increased risk of stroke in people taking certain medication. An individual on that medication is exposed to the fumes. Half the subjects are told that the person exposed actually has a stroke. The other half are told only that the possibility of stroke has increased. As for the increase in risk, for half the subjects it will increase to more than 50% over a five-year period, while for others it will increase to only about 25%. Some courts have used a 50% risk as a threshold for awarding damages. See Kramer v. Lewisville Memorial Hospital, 858 S.W.2d 397, 399 (Tex. 1993).
The scenarios also vary with respect to the state of mind of the person in charge of the manufacturing process. Earlier studies have shown state of mind to affect subjects’ judgment of both causation and culpability. (See Alicke, 1992; Solan and Darley 2001). The manager’s conduct may be knowing (he knew the fumes would be released, but did not stop the process), negligent (he carelessly allowed the fumes to be released, but would have stopped the process had he been aware of its failure), or innocent (the release was not his fault). Thus, the study is a 2 (stroke/no-stroke) x 2 (likelihood of stroke) x 3 (perpetrator’s state of mind) between-subject design, with subjects receiving one of twelve different versions of the questionnaire.


Dependant variables, in addition to manipulation checks, include a range of judgments about the consequences of the action, including compensation for medical bills, other pecuniary loss and non-pecuniary loss; judgments of culpability; and judgments about criminal punishment for both the company and the person directly responsible for the harm. There are slightly different versions of the questions depending on whether the subject is told that a stroke has occurred.


Darley, John M., Lawrence M. Solan, Matthew B. Kugler, and Joseph Sanders. 2007. Liability for Risk: Citizens' Perspectives on Liability for Creation of Risk and Loss of Chance.Second Annual Conference on Empirical Legal Studies. New York University Law School. 9-10 November.