ON THE SECOND FLOOR OF AN OFFICE at 151 Merrimac Street in Boston stand 25 five-drawer file cabinets crammed with hundreds of thousands of handwritten records. Some date from as early as 1939 and are in danger of crumbling to illegibility. But each of 724 files documents the unfolding of a man’s life—the particulars of his childhood, his marital joys and conflicts, his tendencies toward neuroses, the trajectory of his career and, at the nuts and bolts level of health, his cholesterol and blood pressure levels.

There was nothing average about these men. The core participants – those in the Grant Study – were recruited when they were sophomores at Harvard College, and they were predominantly white, well-off and destined to succeed. President John F. Kennedy, from the class of 1940, was among them, and so was Kennedy’s classmate, Washington Post editor Ben Bradlee. Others included a future governor, an assistant secretary of state, a best-selling novelist and a Fortune 500 CEO. Decades later, in 1970, they were joined by a very different group of subjects, a separate cohort of 456 male participants who had grown up in Boston’s inner city.

All of them were part of a study intended to last perhaps 15 or 20 years. But it’s now 75 and counting, and it has mountains of new data still being collected and old data being sorted through. Among the lofty goals of researchers have been to determine what constitutes optimal physical and mental aging, what kinds of men have the best marriages and what personality traits suggest that someone’s health might fail at age 55 or could help him survive to 90 or beyond. The study has generated hundreds of articles and has been cited by thousands of researchers.

“This kind of study will never be done again; it’s serendipitous that it has lasted this long,” says psychiatrist Robert Waldinger, director of the Center for Psychodynamic Therapy and Research at Massachusetts General Hospital and director of the Harvard Study of Adult Development, comprising the Grant Study and the inner city group, the Glueck Study. “Most longitudinal studies [which seek to follow study subjects over extended periods] end before 10 years because too many participants drop out or funding gets cut,” says Waldinger.

While the Harvard Study has had its own issues with on-again, off-again financial support, it has managed to persevere. And while there are questions about what a study of educated white men born long ago can have to say about health and happiness today, the study’s wealth of data continues to be mined. Now researchers can employ tools such as DNA analysis and neuroimaging that hadn’t been invented in its early days, when the Harvard men submitted to Rorschach tests and handwriting analysis. And a recent grant expected to total $4 million over five years from the National Institute on Aging is enabling researchers to study more than two thousand male and female offspring of the Harvard Study men. “The study won’t die when the last of the men do,” says Waldinger, who intends to explore what may be learned from the children and grandchildren of the original participants—will patterns of health and happiness be repeated over the generations?—for as long as the funding holds out. Meanwhile, every page of the study’s records is being scanned and will be available online to researchers who agree to keep confidential the identities of the men who are so richly described.

Yet not everyone is a fan. Marcia Angell expressed her misgivings in a 2013 critique of Triumphs of Experience: The Men of the Harvard Grant Study, by psychiatrist George Vaillant, who directed the study for 38 years. Writing in The New York Review of Books, Angell, former editor-in-chief of The New England Journal of Medicine and a senior lecturer in social medicine at Harvard Medical School, suggested that while the study, at a cost of more than $20 million, is undoubtedly a “brave attempt,” its lack of any real hypothesis meant that “a mountain of information was amassed to no specific end…. Given the mass of data, you could more or less pull out what you were looking for, and that was partly a matter of which theories were in vogue at any given time.”

But it is precisely that serendipity and the passing of time that has revealed the study’s richest findings, argues Vaillant. It took 68 years, for example, for accumulated data to reveal the terrible toll that alcoholism took on the marriages of the men in the Harvard Study. It also took the many decades of the study for researchers to conclude that childhood environment did not predict whether someone would have a stable marriage, and that good marriages could be made despite loveless childhoods. “To understand lives takes a lifetime, and the more time goes by, the more interesting the data become,” he says.

The all-male sophomore classes of Harvard College in 1939 through 1944 provided Arlen V. Bock (known as Arlie), chief of the student health services, the 268 participants he needed to launch a study on optimal health. Bock’s friend, William T. Grant, who owned a chain of discount stores, provided initial funding of $60,000 (almost $1,000,000 in 2014 dollars) with the hope that the research would reveal what social and emotional qualities made a good store manager. The first staff included an internist, a psychologist, an anthropologist, a psychiatrist, a physiologist, a case worker and two secretaries. Students who agreed to participate had to undergo psychiatric evaluations, hours of interviews, physical exams, lab tests, EEGs and, much later, neuroimaging and cognitive testing. They weren’t paid.

In 1948, nine years into the study, Grant was frustrated by the slow pace of the research and by what he considered the disorganization of the study. He withdrew funding, which he had provided through the W. T. Grant Foundation, but he would later reinstate it—and end it and reinstate it and end it again—during the next several decades. Looking for other support, the study turned to the Harvard University Health Services; the Tobacco Industry Research Committee (which hoped to find evidence that smoking had positive effects); the National Institute on Alcohol Abuse and Alcoholism; the National Institute of Mental Health; and the National Institute on Aging.

Each new funding source tried to put its own slant on the structure of the study and shaped the questions it sought to answer. Then, in 1970, came a further complication—the addition of 456 men whose backgrounds were starkly different than those of the Harvard men. This new cohort was actually the control group of another longitudinal project—the Glueck Study of Juvenile Delinquency, started in 1940 by Harvard criminologists Sheldon and Eleanor Glueck. And although these men had fared better than the others in the Glueck Study, managing to stay out of reformatories as children and adolescents, they, too, had grown up poor in Boston’s roughest neighborhoods. Grant Study researchers thought the inner city men—many of them high-school dropouts and with an average IQ of 95—would provide a perfect contrast to the Harvard men, whose IQs were an average of 30 points higher.

There have been other decades-long longevity studies of mental and physical health, each with its own strengths and weaknesses. But what sets the Harvard Study apart is its attempt to understand the power of human relationships—how the subjective experience of connectedness or lack of connectedness with others is a huge factor in health and happiness. “We have found that relationship quality is a better predictor of a long life than some hard-science biomarkers such as cholesterol levels,” says Waldinger.

The study’s data suggest that a man’s capacity for intimate relationships is all-important in predicting future material success. The Harvard men who had been close to their siblings and their parents (especially their mothers) when they were young earned an average of about $265,000 a year (in 2014 dollars) between ages 55 and 60, compared with $111,000 a year for the men who’d had the worst childhood relationships. Strong childhood bonds were also associated with positive coping mechanisms, such as humor and patience, later in life. Those who had those skills earned $401,000 a year compared with those who had the least mature coping skills, who averaged only $173,000 a year.

Even for the inner city men, whose early lives were often plagued with unemployment and poverty, warm relationships in their youth had more to do with future financial stability than did any other factor, including intelligence.

Whether someone’s childhood had been loveless or idyllic also contributed heavily to a man’s mental health, according to Harvard Study findings. By the time they reached their 70s, those who’d had bad childhoods were eight times more likely than those with happy early years to have experienced a major depression and five times more likely to have an anxiety disorder.  They also tended to be heavy smokers and drinkers and to take lots of tranquilizers. In contrast, those who had happy memories of their youth also tended to have many more friends at 70 compared with those who had lacked loving relationships as a child. “Children who fail to learn basic love and trust at home are handicapped later in mastering the assertiveness, initiative and autonomy that are the foundation of successful adulthood,” Valliant writes in Triumphs of Experience.

The connection between childhood experiences and mental health in adulthood was so striking in both groups of men that researchers are now examining their offsprings’ childhoods to see whether the link also exists for that later generation. “We’re hoping to identify the precise ways in which bad childhoods contribute to poor health once you’re an adult,” says Waldinger. He notes that one lingering effect is an adult coping mechanism called negative urgency—reacting to stress by overeating, drinking or having an emotional outburst. “Maybe we can learn what negative effects of childhood can be reversed through interventions such as psychotherapy,” says Waldinger.

Yet while the quality of early relationships was crucial, happiness in another important connection—to a spouse—was more likely to develop later in life. In their middle years, only 18% of Harvard Study men reported being in a happy marriage of at least 20 years. But by age 75, half of the surviving men said they were happily married, and by age 85, more than three-quarters described their marriage as good. And the happier the marriage, the more satisfied older people were with their lives—even when aging brought physical disability and illness. “Being in a happy marriage kept people’s moods up despite not feeling well,” says Waldinger.

Aging to 70, 80 and 90 “is not the nightmare I thought it would be when I was 50,” says Valliant, who is now 80. “Growing old is a lot more fun than people worry that it will be. If you control for major depressive disorder and Alzheimer’s disease, older people tend to be less depressed than younger people.”

That may be true of the Harvard men, whose financial success paved the way for a comfortable retirement. But old age is no picnic if you’re also worried about paying medical bills or hiring caregivers to help you age at home, counters Angell. “As the founders of the Grant Study intended, these were the most fortunate of men, and it is wrong to assume that others will age in the same gentle way,” she writes.

Yet as the Harvard Study itself has aged, it has revealed surprising depths, offering new clues about the factors that underlie mental and physical well-being. Additional data on marriage, for example, has emerged since after 2003, when Waldinger’s team of researchers at long last began focusing on the wives of the Harvard and inner city men. One recent study looked at 81 couples from the Harvard Study, investigating how marital conflict affects cognitive decline late in life. “The brain is more fragile as we get older,” says Waldinger. “People who are lonelier lose their cognitive functions sooner than people who are well connected.” The study found that marital stress took a large toll on the women who felt they couldn’t emotionally lean on their husbands during rough times. Those women, whose mean age was 75.7 years, had greater memory problems 2.5 years after initial testing than did wives who experienced marital conflict but who felt more securely attached to their mates.

Other research has taken advantage of the length of the Harvard Study to explore the connections between mental health and physical illness at various ages. It has long been known that unhappy people have higher levels of stress and inflammation and weaker immune systems, impairing their ability to fight infections and heal wounds, and setting them up to develop disease. The men in the Harvard Study who had poor mental health—as evidenced by substance abuse, troubled relationships and depression—were also those who tended to be physically ill during mid-life. Among those with the worst childhoods, more than a third didn’t live to age 35 or were chronically ill by then compared with only 11% of those with the warmest childhoods, for example. But by age 80, it was a man’s genetic and environmental risk factors, rather than the quality of his childhood, that tended to determine physical health. More than three-quarters of the men who had risk factors at age 40 that set them up for strokes or heart attacks—smoking, alcohol abuse, hypertension, obesity, diabetes—had died or were chronically ill at 80. But only 44% of men with no cardiovascular risk factors in midlife had died or were sick at age 80. Men free of cardiovascular risk factors lived to an average age of 86, while those men with three or more cardiovascular risk factors in midlife lived an average of 68 years. “It appears that poor mental health causes its damage earlier in life,” says Vaillant.

The inner city men tended to be sicker than the Harvard men. Overall, those who’d grown up poor had twice as much lung cancer, emphysema and cirrhosis, and three times as much type 2 diabetes as did their Harvard counterparts. The inner city group became disabled 10 years earlier on average than the college men and died sooner, with an average longevity of 70 years compared with the college men’s 79. Yet those disadvantages didn’t exist for the 25 inner city men who graduated from college.

People can and do change, however. “One of the most important results of this study is that the maturation of the brain and adult development continues long after adolescence,” says Vaillant. “Childhood trauma doesn’t always have to blight our lives, and getting divorced at a young age, for example, says nothing about what your second marriage may be like at age 80,” he says. Maturing defenses and personality traits, and a new restorative marriage can create the resilience and growth that makes the second half of a person’s life much happier—and healthier—than the first.  “Researchers who say things don’t change may be studying people for 10 or 15 years,” says Vaillant. “But the world looks different in 50-year chunks.”

At the same time, the Harvard Study found that poor mental health at an early age is a reliable predictor of psychological problems later in life, while good mental health remains relatively constant throughout life. Current research suggests that the interaction between genes and certain life experiences, such as harsh childhoods, is responsible for the state of someone’s mental health. “A gene for emotional processing, such as the serotonin transporter gene, might set you up for depression, but it’s the traumatic life experience that clinches the deal and creates a bad outcome,” says Waldinger. The challenge, then, is to protect genetically vulnerable people from environmental risks that the Harvard Study and other research are identifying.

Such work will continue to spring from this 75-year-old font of data and life stories. But is it possible now, based on so many decades of analysis, to identify the secret of a life well lived? “What keeps coming up over and over again are how profoundly important relationships are for mental and physical health—way more important than we used to think,” says Waldinger. “Tending to relationships is essential for well-being, and people who don’t do that are at great peril.” Vaillant boils it down even further. While acknowledging that the study has focused on two divergent populations—the very privileged and those who are much less so—he sees a clear lesson. “The most important contributor to joy and success in adult life,” he says, “is love.”