Northeastern University researchers follow up on 80-year-old study of crime prevention program that made a shocking discovery

New research from Brandon Welsh, a professor of criminology and criminal justice at Northeastern, follows up on the long-term effects of a program launched 70 years ago to prevent crime among young people living in urban areas. Photo by Adam Glanzman/Northeastern University

Forty years ago, a researcher made a shocking discovery about a Massachusetts program designed to prevent crime among young people who lived in urban areas. It had achieved the opposite result.

Those who received counseling, mentoring, and other services intended to steer them away from a life of crime were found, years later, to have committed more crimes, led less healthy lives, and died younger than people who received none of these services.

The findings, in 1978, of a study started in 1939, sent shockwaves through the burgeoning field of criminology, and disturbed advocates of social welfare programs, according to Brandon Welsh, professor of criminology and criminal justice at Northeastern.

Now, a Northeastern research team led by Welsh has followed up on the longer-term effects of the crime prevention program, named the Cambridge-Somerville Youth Study, after the Boston-area cities where it was conducted. It turned out that people who received counseling and mentoring had lived just as long, and were no more likely to die of disease, drug or alcohol abuse, accidents, or homicide. This was true not only for people who died in their 70s and 80s, but also for those who died in their 40s.

“There were no significant differences in mortality and cause of mortality between the two groups,” says Welsh, who noted that his research, which comes 80 years after the original study began, is the longest evaluation of a randomized controlled experiment in the world.

Welsh and his colleagues’ research was published last month in JAMA Network Open, a peer-reviewed, open-access medical journal of the American Medical Association.

Welsh acknowledged that his study observed only the age and cause of death, and not how often those who participated in the Cambridge-Somerville Youth Study committed crimes throughout their lives. This, he said, is what he hopes to explore next.

A study 80 years in the making

The Cambridge-Somerville Youth Study ran from June 1939 to December 1945. It was the first randomized clinical trial—which refers to a study in which participants are randomly allocated to different groups—to investigate the effects of a preventive intervention on children’s criminal behavior.

The study involved 506 boys, ages 5 to 13, all of whom were selected because they had been caught stealing, been frequently absent from school, failed to come home after school, or identified as displaying other antisocial behavior.

Half of the youth in the study, the treatment group, received various degrees of counseling and mentoring. They remained in the program, on average, for five-and-a-half years. Many received academic tutoring, attended summer camps, and were encouraged to join community youth programs. The other half, the control group, received no such services.

Four follow-up evaluations have been conducted over the past 70 years to assess the impact of the program. The first two, in 1948 and 1956, found no differences in criminal behavior between the treatment and control groups.

The third follow-up, however, conducted in the 1970s, found that those who received counseling and other services were significantly more likely to commit two or more crimes, fall victim to alcoholism and mental illness, have high blood pressure and heart trouble, and die before they turned 35.

Welsh initiated the fourth follow-up study when he took over as the director of Cambridge-Somerville Youth Study about five years ago. He sought to discover whether the negative effects on mortality observed in middle age had persisted, or changed, in old age. The research team also includes Northeastern criminology and criminal justice professor Gregory Zimmerman and doctoral student Alexis Yohros, as well as former Northeastern doctoral student Steven Zane, now an assistant professor of criminology at Florida State University.

What the new research found

By early 2018, Welsh’s team had tracked down records for 96.4 percent of the original study’s participants, and 88.1 percent were confirmed deceased. About 91 percent of those in the crime prevention program, the treatment group, had died. Eighty-five percent in the control group had died.

Welsh said that it’s unclear why his research did not find that the crime prevention program had negative effects on the study’s participants, as the earlier study had found. He said the earlier findings on premature mortality, specifically, “might have been a singular event.”

Welsh said that when the findings of the third follow-up were publicized, there was great concern that they would lead legislators to curb funding and other support for social welfare and mentoring programs. However, he said, as time has passed, the findings have underscored the importance of and need for follow-up studies that examine the long-term effects of crime prevention programs.

For media inquiries, please contact