The scope-severity paradox

jpg law justice 003 Do criminals deserve a less severe punishment if they harmed more people ?

Most people would almost certainly answer "no". Of course: punishment should be sensitive to the severity of the crime. That's what we usually think.

Yet in a compelling paper published in Social Psychological and Personality Science in August 2010, Loran F. Nordgren and Mary-Hunter Morris McDonnell found that increasing the number of people victimized by a crime actually decreases the perceived severity of that crime and leads people to recommend less punishment.

The scope-severity paradox presented in the article is indeed astonishing. The paper is also exemplary in how beautifully it combines lab experiments and analysis of real-world data.

The authors conducted experiments to test the effect. 60 participants were asked about the severity of a crime. A financial advisor has betrayed his clients, how severe is his crime, how many years of jail should he get? The astonishing (and robust) result is that the crime is seen as significantly less serious when there are 30 victims rather than 3. When asked about how participants imagine one of the victims, participants gave an average of 3 traits less in the case of 30 victims than in the case of 3 victims, showing that their representation is much more vivid in the case of fewer victims. According to the authors, this difference in the vividness of how the victims are represented accounts for the paradox.

The authors then turned to archival data to demonstrate that the scope-severity paradox is a robust, real-world effect. They collected archival data of actual jury verdicts concerning 136 cases of poisoning in the U.S. spanning over a 10-year period from 2000 to 2009. They found that juries required defendants to pay higher punitive damages when their negligent behavior harmed fewer people.

The authors then returned to the lab and tried to find a way to reduce the effect. It appears that by making victims more individually identifiable, the effect is reduced. With identifiable victims people stop punishing less criminals who have harmed more people. They do not, however, punish them more: they just give the same punishment, regardless of the number of victims.

I suggest that there may be a strong connection with Susan Carey's work on the acquisition of mathematics: there is a bootstrapping process going on when children learn to deal with 3, then with 4, then with 5. It may well be that the bootstrapping process is never entirely complete, and we don't have, even as adults, a very clear concept of 30, or at least it is far less clear than our concept of 3. Imagining 100 tables appears much more difficult to me than imagining 5. Imagining 100 real people probably goes beyond my cognitive capacities.

Can this effect at least partially explain why the world, or at least the media pay so much more attention to single criminals than to institutions who fraud and harm hundreds or thousands of people?

Comments (1)
Connecting numerical cognition and moral reasoning
Brent Strickland
Sunday, 06 November 2011 15:41

This is an interesting article, and I like your idea about connecting the finding to numerical cognition. Some people believe that there are two systems for representing number. The "subitization" system allows for very precise representation of small numbers (1-3) and the approximate number system allows for much coarser representation of large numbers. For each of their examples, Nodgren and McDonnel chose number pairings that straddled the divide between the subitization and approximate systems. Study 1 compared 3 to 30 victims and Study 2 compared 2 to 20 victims. So I wonder if this is an effect of which number system is being activated? An easy way to test this would be to have a single scenario where there are either 30, 20, 10, 3, 2 or victims. I would predict no differences between 30, 20 and 10 and no differences between 3, 2 and 1, but that there would be differences between any pairing in those two groups (30 vs. 3, 20 vs. 3, etc...)

Add your comment

Your name:
Subject:
Comment (you may use HTML tags here):

Additional information