The first post in this series addressed general principles for identifying and analyzing possible regulatory actions that either could have prevented Sandy Hook or prevent future such events. Subsequent posts discussed, in general terms, various regulatory options including changes in physical security at elementary schools, changes in personal security at elementary schools, and more stringent gun control.
Today we consider regulations that would change mental health screening and treatment — that is, regulating certain people rather than places (such as schools) or articles (such as guns). The reason this alternative may be appealing is that mass murders have been committed predominantly by persons with verified or suspected mental illness. It is plausible that some combination of improved diagnosis, screening for a predilection toward violence, and treatment could reduce the numbers of future mass murders.
The focus here is on a pragmatic question (would regulation be effective?) and a technical one (would regulation be efficient — i.e., less costly than other alternatives of equal effectiveness?). Debating legal and constitutional matters is left to others, but such debates should be irrelevant for any regulatory alternative that would not be effective. Regulations that are ineffective have no social benefits, and it is therefore impossible for them to ever be efficient.
Among the first disclosures about Adam Lanza, the Sandy Hook shooter, is that he apparently had been diagnosed with Asperger’s Disorder. Asperger’s is one of several conditions often grouped together under the umbrella of autism. The key characteristic of persons with Asperger’s is difficulty with social interaction, as indicated by the diagnostic criteria in DSM-IV 299.80. However, a predisposition to violence is not one of these criteria. Others have said that Lanza was a “nerd” and honor student, attributes that also are irrelevant. Thus, as a tool for identifying persons who might be predisposed toward violence in general, much less mass murder, a diagnosis of Asperger’s Disorder (or being a “nerd” or honors student) is essentially useless, and thus it has no potential benefit. Meanwhile, the public disclosure of such information has real costs by stigmatizing persons with Asperger’s, or who are “nerds,” or who are honors students.
When the Sandy Hook investigation is completed, it may be revealed that Adam Lanza also had been diagnosed with other conditions that suggest a propensity toward violence. This could be helpful, but it begs three questions:
- How great a propensity toward violence is sufficient to justify public intervention, such as involuntary commitment and treatment? Setting a high bar reduces the number of true positives (a bad thing) and the number of false positives (a good thing).
- How accurate is the diagnosis and the information on which it is based? Experts may disagree about the diagnosis; they may place different weights on different facts; and some of the facts on which they rely may be false. If opinion is highly variable across experts, we can be confident that there is actually a lot of uncertainty. (If opinion is not variable, it could mean that uncertainty is low or that experts’ opinions are not independent.)
- Will regulation unintentionally deter those who need mental health treatment from seeking and obtaining it? The prospect of involuntarily losing one’s liberty as a result of seeking treatment will not encourage those who genuinely need treatment from coming forward. A regulation that empowers experts to violate normal confidentiality rules would deter some persons from seeking treatment. A regulation that requires them to violate confidentiality would deter even more.
Choosing a decision rule is very hard. Deciding which persons (whether as adults or children) to designate as potential risks entails a number of costs that are usually ignored. As the decision rule widens to capture an increasing number of people, the number of false positives rises, as does the cost of classification error. We know in principle where to draw the line — where the sum of costs from false positives (classifying people as risky when they are not) and false negatives (failing to classify people as risky when they are). But we do not know, at a practical level, how to estimate these costs.
That doesn’t mean we do not classify, however imprecisely. For example, the Transportation Security Administration maintains a “no-fly list” that is analogous to the classification scheme described qualitatively above. According to an FBI official quoted on NPR’s Morning Edition, in January 2011 the number of persons on the no-fly list was about 10,000, with between 500 and 1,000 of them US citizens. For some, the no-fly list is controversial on civil liberties grounds:
Ben Wizner, an attorney for the American Civil Liberties Union, says he thinks that traveling to Yemen, as Gulet Mohamed did, is a red flag that could get you on the list.
“I think if the class of people is going to be defined that broadly, it really underscores the need for a fair process to sort out who belongs on it and who doesn’t,” Wizner says.
Reasonable people may disagree about whether traveling to Yemen should earn a red flag, but flagging travelers to Yemen for greater scrutiny is not analytically unreasonable. There also would be arguments about about when to flag persons for a predilection toward violence and either be denied a civil right (e.g., the right to own a firearm) or be compelled to take some involuntary action (e.g., submit to commitment and treatment).
Note that Connecticut gun laws already prohibit certain persons from purchasing or possessing firearms, including persons who have been found not guilty of a crime by reason of mental illness or who have been recently confined in a mental hospital. According to the Law Center to Prevent Gun Violence:
Thirty-four states and the District of Columbia have laws that restrict access to firearms by persons who are mentally ill. While most states use definitions of mental illness similar to the federal Brady Act [18 U.S.C. § 922] and its implementing regulations, several states have broadened the category of mentally ill persons who are prohibited from purchasing or possessing firearms.
These laws and regulations are inherently reactive, of course. They deal only with firearms, not other weapons such as knives, and not tools that have been used as weapons. Moreover, they only deal with persons who have already made contact with law enforcement or mental health practitioners and systems. There are no published reports that Adam Lanza had a criminal record or had been previously committed involuntarily, and there are several reports that say he had not.
We do know, however, that if Lanza was a person proscribed from gun possession by Connecticut law and regulation, the law and regulation were ineffective. As we’ve noted previously, regulations that are ineffective yield no benefits.