Why We Punish
Posted by John Darley & Pam Mueller on January 7, 2008
Previous studies have shown that when individuals are given the opportunity to punish an offender, they sentence retributively, based on the moral wrongfulness of the offender’s actions. However, those initial studies did not assess whether people based these retributive punishments on the harm the act caused or the wrongfulness of the offender’s intent.
* * *
The easiest case in criminal law involves one person intending to shoot and kill another person — who aim, fires, and hits the target. The easiest case, however, is never the most interesting. Criminal law is full of cases in which the intended assassin has poor aim, and winds up shooting the wrong person or no one at all, as well as cases in which a hunting or target-shooting bullet goes awry, making an unintentional killer out of a marksman.
Princeton social psychologists (Alter, Kernochan, and Darley) investigated situations in which the harm done and the harm intended were not the same. How much punishment — or, more specifically, how long of a prison sentence — would subjects impose on actors in these situations? Would the harmfulness of the act or the wrongfulness of the act be the primary factor in these judgments? Such sentencing preferences would tap into moral intuitions about the importance of wrongfulness (intent) and harmfulness (consequence).
Modern criminal law requires both actus reus (a wrongful act) and mens rea (a guilty mind) to coincide for a crime to have occurred. Liability is thus contingent on both harmfulness and wrongfulness. Doctrines such as “innocent agency” allow one person’s acts and another person’s intent to be consolidated so that a crime can be found to have occurred. In such a situation, a knowing actor manipulates an innocent individual into committing an act that would be criminal if mens rea were present. The person who acted wrongfully with intent in this situation is the one who is punished, not the person who acted harmfully without intent.
Prior studies (e.g. Carlsmith, Darley, & Robinson, 2002) have found that people sentence offenders based on the “just deserts” theory of punishment, in which individuals are punished in proportion to the perceived moral wrongfulness of their acts, rather than the “deterrence” theory, in which acts are punished in proportion to the costs they impose on society. These studies did not, however, separate the distinct factors of wrongfulness and harmfulness within the acts presented to participants.
The more recent set of studies used stimuli that separated harmfulness and wrongfulness. Stimuli were grouped in triads, in which a given act was manipulated to be 1) harmful, but not wrongful; 2) wrongful, but not harmful; 3) both harmful and wrongful. The first two would separate the issues that had been confounded in past experiments, while the third condition would reveal any interaction between harmfulness and wrongfulness in determining sentence length.
The research hypothesized that participants would impose higher sentences on wrongful acts than on harmful acts and that, given those tendencies, participants would also believe that a legal system that emphasized wrongfulness was fair and just, while a system of punishment rooted in harmfulness was unfair. According to those hypotheses, subjects would see a system as particularly unfair if it assigned heavy punishment to a person who caused harm with no intention to do so.
The three situations that were manipulated to create the experimental stimuli were a theft, a shooting homicide, and a simple assault. Participants separately rated the harmfulness and the wrongfulness of each of the three acts on 9-point scales. They then imposed appropriate sentences for the offenders in each of the situations, using a 13-point rating scale (ranging from no sentence to a death sentence) that has been tested in many prior experiments. Finally, after reading fabricated comments in support of a harmfulness- or a wrongfulness-based sentencing system, they rated their agreement with the statements “The legal system is unfair and unjust” and “If I were a citizen, I would feel anger toward the legal system.”
Participants assigned harsher sentences for “wrong only” cases than “harm only” cases, and assigned the harshest sentences in situations where the offenses were both wrongful and harmful. For instance, in the shooting homicide example, participants’ mean sentence was 1.07 on the 13-point scale (essentially, no liability) when there was harm without intent; with intent but no harm, the mean sentence was 7.70 (a 1-3 year prison sentence); and when there was both intent to harm and harm caused, the mean sentence was 10.64 (between 15 and 30 years in prison). Analysis showed that wrongfulness was the primary determinant of sentencing, and judgments of wrongfulness were significantly better predictors of sentence length than judgments of harmfulness.
The last dependent measure assessed participants’ attitudes toward the justice system. As expected, participants expressed greater contempt for a harm-based system than for a wrong-based system.
In the hopes of ensuring that subjects were carefully considering the implications of the sentencing judgments, a second, smaller study brought the situations closer to home for Princeton students. The experiment manipulated the harmfulness and the wrongfulness of behavior surrounding a breach of the Honor Code. Additionally, the Honor Code Board is a quasi-judicial institution meting out punishments that are more immediately relevant to Princeton students; therefore, it is a better vehicle with which to test the hypothesis that isolated unjust outcomes can cause development of contempt for the justice system.
The paradigm was similar to the first experiment: participants rated the three related situations on harmfulness and wrongfulness scales, and then were asked to make a judgment about punishment on a scale of 0 (full exoneration) to 6 (expulsion). Participants were then asked to rate their contempt for the Honor Code Board if one of two amendments were adopted by the Board, codifying either 1) punishment of offenders based on harmfulness of behavior, or 2) punishment of offenders based on wrongfulness of behavior.
As in the first study, punishments were most severe in the “harm and wrong” condition, less severe in the “wrong only” condition, and least severe in the “harm only” condition. Wrongfulness was again a reliable predictor of sentences, but in this study, harmfulness did not significantly predict sentencing. All participants expressed some contempt for the Honor Code Board, whether as a result of general negativity or, potentially, as a result of a desire for balance between evaluation of wrongfulness and harmfulness in sentencing decisions. However, as expected, expressed contempt was greater in the “harmfulness only” condition.
These findings suggest that when individuals seek to apply an appropriate sentence, they rely on the heuristic of how intuitively wrongful the offense feels. Harmfulness is not completely cast aside in the decision; it also plays a direct, albeit small, role, as evidenced by the harsher sentences for the “wrong and harm” conditions. (This role may have been further diminished by the fact the manipulations for wrongfulness were statistically stronger than the manipulations for harmfulness.)
This study provides further evidence that people prefer a “just deserts” theory of punishment to a “deterrence” theory. Thus, the justice system should continue to use wrongfulness as an important part of its punishment calculus. If an entirely harmfulness-based sentencing scheme were implemented, people would view the legal system very negatively; thus, further amendments to the sentencing guidelines should reflect a concern with wrongfulness in order to gain the most public respect.
The results also provide some insight into jurors’ thoughts and behavior in capital cases – the only juries that play a role in sentencing decisions. These jurors are likely to be affected by the same intuitive primacy of wrongfulness over harmfulness. Judges are also likely to be susceptible to these same heuristics, though their behavior will be tempered by sentencing guidelines and other legal duties; certainly, more research is necessary to reliably extend these findings to judges.
* * *
For more information about the study, click here. To link to a related law-review article co-authored by Paul Robinson & John Darley, titled “Intuitions of Justice: Implications for Criminal Law and Justice Policy,” click here (pdf).
This entry was posted on January 7, 2008 at 12:01 am and is filed under Law, Legal Theory, Public Policy, Social Psychology. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.