Heinz professors will study violence, psych disorders

The nightly news sometimes tells tragic stories: grade school students lose control and open fire on fellow students. A man shoots a store clerk over a minor altercation.

These seemingly irrational acts of violence are perplexing. Experts speculate as to what might have caused them, but there has been little evidence to support those experts’ theories.

Now, two researchers in the Heinz School of Public Policy and Management think that statistical methods can offer some insight into patterns of violence and psychiatric disorders.

Daniel Nagin, a Teresa and H. John Heinz III Professor of Public Policy, and Jeffrey Dominitz, an assistant professor of economics and public policy, have been granted a $1.1 million award from the National Institute of Mental Health to continue their work in quantifying the developmental pattern of psychiatric disorders and chronic violence. The three-year grant will support their work in identifying relationships between abnormal behavior and events in the subject’s past that may have been related to the behavior.

Nagin and Dominitz are looking at behaviors such as chronic depression, schizophrenia, conduct disorder, and violent actions that resulted in criminal convictions.

“These kinds of behavior tend very rarely to be isolated events in [people’s] lives,” said Nagin, who has been studying statistical trajectories of behaviors for the past ten years. “[Violent] behaviors can often be traced back to very early in life.”

Trajectory models generally plot the number of incidents over time across different subject groups. For instance, they allow one to see the differences between chronically violent people and people who were only violent as adolescents.

According to Nagin, common factors linked to violent behavior are low IQ, parents with a criminal record or poor child-rearing practices, and risk-taking behaviors.

However, marriage helps to decrease men’s violent behavior, said Nagin. “The woman almost becomes a full-time parole supervisor,” he said.

“The propensity to be physically aggressive is built into our species. It’s something that most of us are born with,” said Nagin. He believes that everyone is born with violent tendencies, but that they are usually suppressed as a child develops the control necessary to conform to society.

“If you focus on the act, and not the consequences, then the most physically aggressive people on earth are two-year-olds,” said Nagin. According to Nagin, the people who do not learn to control early aggressive impulses as children are much more likely to become adults who commit violent acts.

In 2001, Nagin and Richard Tremblay, a professor of pediatrics, psychiatry, and psychology at the University of Montréal, reported on their study of physical aggression in boys in Canada. Nagin and Tremblay compared boys who were continually aggressive from ages six to fifteen with boys who were aggressive at age six, but later stopped.

They found that the only two differences between the groups were that the perpetually aggressive boys were more likely to have mothers who began childbearing as teenagers and to have mothers who were poorly educated. Their findings motivated Quebec to establish a $70 million program targeted at helping high-risk mothers improve their parenting.

In their current work, Nagin and Dominitz are also trying to address the problem of separating cause from effect in the non-experimental data they use. Since violent and psychiatric behaviors cannot be manipulated in an experimental manner, it is difficult to determine whether events in a person’s past caused violent behavior, or whether it is only a correlation.

The data used for their statistical models comes from a number of different countries, and according to Nagin, the datasets range in sample size from 400 to 3000, with most sets containing measurements from about 1000 people.

While Nagin and Dominitz are primarily developing models for psychiatric disorders and violent behavior, their work could also be used in other areas.

“The statistical methodologies I’m developing can be applied to many other domains,” said Nagin. For example, his trajectory models can also be used to predict obesity and testosterone levels.