How long will you live? Ask your friends.

When actor James Gandolfini died in the summer of 2013, at age 51, a prominent cardiologist described him as “a heart attack waiting to happen.” The award-winning Sopranos star was overweight and inactive, and on the evening he died, he had indulged himself in a diet of rum, beer and fatty foods. In short, he didn’t take care of himself, and this lack of discipline no doubt contributed to his untimely death.

Scientists have long known that personality is a good indicator of future health and mortality. In fact, personality traits are better predictors of lifespan than either intelligence or socioeconomic status. But it’s not easy to measure personality in a reliable way, since self-reports are notoriously biased and misleading.

That’s why psychological scientist Joshua Jackson, of Washington University in St. Louis, turned to friends instead. He and his colleagues decided that, instead of asking subjects about their own temperaments, they would combine the assessments of close friends—to see if these peer estimates of personality were a better predictor of mortality.

They used an existing data base called the Kelly/Connolly Longitudinal Study. Back in the 1930s, 600 individuals in their mid-20s—300 engaged couples—volunteered for a study of personality and newly formed marriages, and this study had included personality ratings by five close friends—most of them members of the wedding party. It also included self-ratings of personality. Last year—75 years after the original study began—the scientists tracked down most of the original volunteers—or at least their obituaries. So they were able to compare friends’ views with the volunteers’ views of themselves—and see which was a better predictor of life and death.

Friends did much better at identifying the personality traits linked to mortality, as described in a forthcoming issue of the journal Psychological Science. Specifically, young men who were viewed by their groomsmen as conscientious and open to experience—these men went on to live the longest lives. Conscientiousness and openness are two of the so-called Big Five personality traits. Conscientious people are dutiful and disciplined and organized and dependable, while open people are intellectually curious and inventive. Men with these two traits lived longer—or alternatively, those lacking these traits died earlier.

For women the picture was different. High levels of agreeableness and emotional stability—as identified by their bridesmaids and other close friends—were most protective over the lifespan. Emotional stability is the opposite of neuroticism—the tendency toward anger and anxiety and depression—while agreeableness encompasses cooperation and compassion. These findings must be seen in historical context, the scientists caution: The subjects entered adulthood in the 1930s, when these positive traits were indicative of a supportive and easy-going wife, the emotional leader of the family.

The men’s self-reports were also pretty good at predicting life and death, though not as good as the friends’ assessments. The women’s self-reports did not predict mortality. For all the young men and women coming of age 75 years ago, their friends picked up on something in their personalities that was related to health and well-being—insights that could prove relevant to public health policy today.

Follow Wray Herbert’s reporting on psychological science in The Huffington Post and on Twitter at @wrayherbert.