In a study of crimes committed by people with serious mental disorders, only 7.5 percent were directly related to symptoms of mental illness, according to new research published by the American Psychological Association.

Researchers analyzed 429 crimes committed by 143 offenders with three major types of mental illness and found that 3 percent of their crimes were directly related to symptoms of major depression, 4 percent to symptoms of schizophrenia disorders and 10 percent to symptoms of bipolar disorder.

"When we hear about crimes committed by people with mental illness, they tend to be big headline-making crimes so they get stuck in people's heads," said lead researcher Jillian Peterson, PhD. "The vast majority of people with mental illness are not violent, not criminal and not dangerous."

The study was conducted with former defendants of a mental health court in Minneapolis. The participants completed a two-hour interview about their criminal history and mental health symptoms, covering an average of 15 years. The study, published online in the APA journal Law and Human Behavior, may be the first to analyze the connection between crime and mental illness symptoms for offenders over an extended period of their lives, said Peterson, a psychology professor at Normandale Community College in Bloomington, Minn.

The study didn't find any predictable patterns linking criminal conduct and mental illness symptoms over time. Two-thirds of the offenders who had committed crimes directly related to their mental illness symptoms also had committed unrelated crimes for other reasons, such as poverty, unemployment, homelessness and substance abuse, according to the research. "Is there a small group of people with mental illness committing crimes again and again because of their symptoms? We didn't find that in this study," Peterson said.

In the United States, more than 1.2 million people with mental illness are incarcerated in jails or prisons, according to the federal Bureau of Justice Statistics. People with mental illnesses also are on probation or parole at two to four times the rate for the general population.

In addition to interviews with offenders, the researchers reviewed criminal history and social worker files to help rate crimes based on their association with symptoms of schizophrenia disorders (hallucinations and delusions), bipolar disorder (impulsivity and risk-taking behavior) or major depression (hopelessness and suicidal thoughts). The ratings were: no relationship between mental illness symptoms and the crime, mostly unrelated, mostly related or directly related.

A crime could be rated as mostly unrelated or mostly related to mental illness symptoms if those symptoms contributed to the cause of the crime but weren't solely responsible for it. For example, an offender with schizophrenia who was agitated because he heard voices earlier in the day later got into a bar fight, but he wasn't hearing voices at the time of the altercation, so the crime was categorized as mostly related.

When the directly related and mostly related categories were combined, the percentage of crimes attributed to mental illness symptoms increased from 7.5 percent to 18 percent, or less than one in five of the crimes analyzed in the study. Of crimes committed by participants with bipolar disorder, 62 percent were directly or mostly related to symptoms, compared with 23 percent for schizophrenia and 15 percent for depression. Some participants may have described their mood as "manic" during a crime even though they could have just been angry or abusing drugs or alcohol, so the percentage of crimes attributed to bipolar disorder may be inflated, Peterson said.

Almost two-thirds of the study participants were male, with an average age of 40. They were evenly divided between white and black offenders (42 percent each, 16 percent other races), and 85 percent had substance abuse disorders. The study did not include offenders with serious violent offenses because the mental health court did not adjudicate those crimes, but the participants did describe other violent crimes they had committed. The study also did not examine how substance abuse interacted with mental illness to influence criminal behavior.

The researchers said programs designed to reduce recidivism for mentally ill offenders should be expanded beyond mental health treatment to include cognitive-behavioral treatment about criminal thinking, anger management and other behavioral issues. Programs to address basic needs also are essential to reduce recidivism for all offenders after incarceration, including drug treatment and housing and employment support, Peterson said.