Download data and study materials from OSF
Paul H. Robinson
University of Pennsylvania
Geoffrey P. Goodwin
University of Pennsylvania
Sample size: 300
Field period: 1/15/2009-9/3/2009
To what extent do the punishments assigned by the United States' criminal justice system match the moral intuitions of its citizens? Answering this question is of importance to debates concerning the legitimate aims of the criminal justice system. On one side of this debate are theorists who propose that the ultimate aim of a criminal justice system should be what is referred to as "crime control": the reduction and ideally prevention of future crime through the deterrence (or incapacitation) of both actual and would-be offenders. On the other side of this debate are theorists propose that the criminal justice system should primarily aim to administer retribution (i.e., "just deserts") for crimes that have been committed, such that offenders are punished in proportion to the moral severity of their crimes. This debate is important, because the punishment decisions that either side of the debate recommend do not always converge.
The present proposal aims to investigate this notion of "empirical desert" by examining whether a series of modern legislative provisions accurately reflect community intuitions of justice. There is reason to suspect that a good deal of modern "crime control" legislation produces sentencing decisions that are considerably harsher than would be found acceptable by the community. This kind of discrepancy potentially arises with respect to at least seven different kinds of "crime control" doctrines currently in operation in criminal justice systems throughout the United States: three strikes rules, drug offense penalties, adult prosecution of juveniles, abolition or narrowing of the insanity defense, strict liability laws, felony murder laws, and the criminalization of regulatory violations.
The goal of this study was to compare subjects' treatment of cases involving the crime-control doctrines described above to the law's treatment of those cases. Subjects were asked to rank and then assign a specific punishment to the cases, and these rankings and assigned punishments were then compared to the rankings and punishments that the law gives the cases.
We also examined subjects' treatment of a set of "milestone" cases, which did not engage the various crime control doctrines of interest, and we compared subjects' responses to these cases with their responses to the crime-control cases.
Our main prediction was that for the crime-control cases, the subjects' sentences would be much less punitive than the law's. In contrast, while we expected that there might be some discrepancies between subjects' sentences and the law's for the milestone cases, these were predicted to be much smaller and less systematic than the discrepancies for the crime control cases.
We conducted a within-subjects experiment that compared participants' punishment decisions about a set of cases fitting seven crime control doctrines ("crime control" cases), with their punishment decisions about a set of control cases. We were chiefly interested in whether participants' decisions converged with the sentencing guidelines of the American criminal justice system – we predicted that such convergence will occur for the control cases, but not for the test cases. Participants were asked to rank the severity of, and apply sentences to, descriptions of offenses that match the facts of specific, real-world cases. They were presented with 12 milestone offenses and 12 "crime control" offenses, each of which briefly describe an actual case. The 12 milestone offenses differed widely in severity, and each represented different legal offenses. They have been shown to produce an extremely high degree of concordance in people's rankings of their relative blameworthiness (see Robinson & Kurzban, 2007). The 12 crime control cases correspond to one of the seven crime control doctrines mentioned above, and also differ widely in severity. The entire set of 24 cases were presented one at a time, in a new random order for each participant. Participants' first task was to rank each case in terms of its relative blameworthiness. They had to decide on an exact position for each case – no ties were allowed.
Our primary outcome variable was subjects' rank ordering of the different cases they encountered, and how this compared with the law's treatment of these cases -- considering the "crime control" cases on the one hand, and the "milestone" cases on the other.
We also tested for the degree of concordance in subjects' rankings of these crimes, as well as the relation between their responses and particular demographic variables.
Consistent with previous research, subjects tended to agree very strongly on the rank ordering of the "milestone" cases (Kendall's W = .86, p < .001). There was also moderate agreement on the rank ordering of the "crime-control" cases (Kendall's W = .52, p < .001).
As predicted, subjects' sentences for the "crime-control" cases were systematically (and quite substantially) lower than the law's sentences for these cases. In contrast, subjects' sentences for the "milestone" cases were much closer to the law's. If anything, subjects' sentences were higher than the law's for these cases.
These results provide strong confirmation for our hypothesis that there exists a substantial discrepancy between the law and lay individuals' intuitions regarding the blameworthiness of offenses that fall under several important crime control doctrines.
Robinson, Paul H., Geoffrey P. Goodwin, and Michael Reisig. 2010. The Disutility of Injustice. New York University Law Review, 85:1940