Triumphs of Experience: The Men of the Harvard Grant Study

by Vaillant, George E.

  TROUBLE IN PARADISE

  Writing to Bock in 1944, Grant chastised him for managing “a cooperative study with a good deal more distrust, disdain and independence of outsiders than I have ever seen productive of satisfactory results.”24 He would be willing to contribute to the Study in the future, he said, only if Bock also secured additional funds from the Harvard Department of Hygiene. He promised one more payment of $30,000 ($300,000 in 2009 dollars) to fund the project until he and the trustees of his foundation could feel confident that the Study was organized effectively. In hindsight, it appears that no one in the Study had fully understood the constraints of longitudinal studies, or the patience that they require. And while Clark Heath was a model of clinical expertise and personal warmth, administrative organization was not his forte. Ironically, there were gifted future administrators at the Grant Study, and one, at least, had an intuitive understanding of the powerful but demanding nature of this kind of research. Donald Hastings, once a Grant Study psychiatric interviewer, carried out and harvested one of the first well-designed long-term studies of neurotic outpatients when he was chairman of psychiatry at the University of Minnesota Medical School in 1958.25 Hastings’s study was one of the first prospective longitudinal studies ever published in the American Journal of Psychiatry, and the first that I ever clipped out of a journal as a medical student to keep in my own files. It was Hastings who showed me how prospective longitudinal study could bring order to muddled psychiatric thought, and I admired his work for years before I ever heard of the Grant Study.

  Bock grumbled to Earl Bond, Grant’s closest scientific advisor and future Grant Foundation trustee, “I had to tell Mr. Grant that the trustees of the foundation could not expect to run the Grant Study.” After this, not surprisingly, Grant Foundation largesse dried up for almost twenty years. The embryonic Grant Foundation had no medical researchers or clinicians among its trustees, who did not yet understand either the limitations or the potential of the Grant Study’s longitudinal design. An orchard like the Grant Study takes time to bear its fruit.

  Even as publications began to emerge more regularly (Appendix F), the Study authors clung to the bounds of their respective disciplines, selecting the kinds of narrow topics that yield academic journal papers. They were not yet exploiting the prospective, longitudinal potential of the Study. For example, they could have tested the somato-type data more critically; certainly they had the necessary material to examine in 1946 whether this information did or didn’t predict anything about military rank at discharge. But they didn’t. By 1970, even after fifty publications, the Grant Study remained little known and little cited. It had also been underfunded for fifteen years, as I will describe shortly.

  On the other hand, the seeds that Grant, Bock, and Heath had planted really did promise a rich harvest decades into the future. For example, as we’ll see in Chapter 10, Woods’s intuitively derived personality traits permitted us to predict political choice reliably over the next half-century of the men’s voting lives. And forty years later, Monks’s careful debriefing of the men after the war regarding the details of their combat experiences led to one of the world’s few prospective studies of the antecedents of posttraumatic stress disorder.26

  THE WAR AND ITS AFTERMATH

  After Pearl Harbor, most of the Grant Study men went into the armed services directly from Harvard, and most of them performed unusually well there. But they took their first real steps into adult life in a world reeling from the stupendous social and economic sequelae of the war, which made rigorous demands on their adaptive capacities.

  THE NEXT TWENTY-FIVE YEARS: 1945–1970

  The end of the war saw a sea change in Western social science as well as in Western society. Hooton’s ferocity notwithstanding, genetics and biological determinism were out, partly out of emotional overreaction to the Nazi misuse of eugenics in pursuit of a racist ideology. Relativism, Skinner, and cultural anthropology were in, and scientists were looking beyond constitutional endowment for the conditions of life that shape individual outcomes. In the United States, psychoanalysts were becoming chairmen of major departments of psychiatry.

  Censorship issues during the war years limited communication between the Grant Study and the men to the occasional V-mail. But once the war was over, the Study began following the College cohort with annual questionnaires. (That practice continued through 1955; it’s been every two years since then, with only a few exceptions.) The questionnaires are long, and designed to take advantage of the men’s high verbal skills. They inquire about employment, health, habits (vacations, sports, alcohol, smoking), political views, family, and, especially, quality of marriage.27 They have varied somewhat from decade to decade and from director to director, but inevitably they reflect changes in intellectual climate. Only in 1955, for instance, came the first inquiries about college roommates and girlfriends. A perfect prospective study, of course, would forbid retrospective data like this.

  By 1948, the Grant Foundation had made good on its threat to withdraw funding. But Alan Gregg of the Rockefeller Foundation stepped in with the salary for a cultural anthropologist. After discussion with Gregg and his colleagues, Bock and the Study staff decided to rename the project the Harvard Study of Adult Development. (Since 1970, that name has covered both the College cohort of the Grant Study and the Inner City cohort of the Glueck Study.) At this point the Study was essentially supported by Harvard Health Service funds. But to avoid confusion, letters and questionnaires to the men continued to address them as “Grant Study” members, and the Study has continued to be known that way informally to this day.

  Margaret Lantis was the new cultural anthropologist, trained to look for, respect, and study cultural differences. She had a doctorate in her field, and was a vastly more sophisticated interviewer, if a less cozy one, than Lewise Gregory, whose training had taken place around Virginia tea tables. In 1950, Lantis interviewed 205 Study men and their wives in their homes. In addition, in the ten years after the war, 171 men dropped by the Study to visit, and the staff was fortuitously able to glean some current information from them. Clark Heath and Gregory usually did those debriefings.

  In collaboration with Charles McArthur, a social psychologist who would soon become director of the Study, Lantis administered the Thematic Apperception Test to many of the men. The TAT is a projective test designed to study affect and relationships; this was a departure from the cognitive focus of Frederic Wells. In those days the TAT was a clinical guide, not a research instrument. Twenty years later (in the 1970s), Charles Ducey, a psychologist who was skilled in projective testing but blind to other Study results, reviewed the men’s TATs and their earlier Rorschachs. While the projective tests provided good portraits and corroborated many of the men’s known personality quirks, they did not help the Study make accurate predictions about the men’s mental health or flourishing at age forty-seven. (That’s an age that keeps coming up because it was the approximate age of the men in their twenty-fifth reunion years; more on this shortly. So was seventy-two, the year they celebrated their fiftieth reunions.) In 1981, however, Dan McAdams, soon to become a distinguished researcher of Intimacy and Generativity (see Chapters 4 and 5), successfully used those same TATs to predict intimate relationships in a manner that an academic journal found worthy of publication. His construct of intimacy motivation could significantly predict success in the two fundamental criteria of the good life so often ascribed to Freud—Arbeiten und lieben (to be able to work and to love).28

  In 1953, the President and Fellows of Harvard College declared that unless additional support could be garnered from private sources, Harvard would no longer continue to provide for the Study as it had over the five lean years past. And so on July 12, 1954, Drs. Bock and Heath wrote a sad letter: “To all Grant Study Participants, We regret to inform you that financial support of the Study ceased July 1, 1954. . . . Your cooperation has been a source of great satisfaction and pleasure to us. . . .”29

  Final fa
rewells were averted, however. In December of that year, the desperate Study submitted a formal grant to the Tobacco Industry Research Committee for $15,880 (roughly $150,000 in 2009 dollars), suggesting that something might be learned about the “positive reasons” that people smoke. And for more than a decade tobacco became the Study’s main source of support, although the documentation is sketchy.

  It’s worth noting that in 1954, when this deus ex machina materialized, the Study was already sixteen years old. It had achieved the fifteen years first envisioned for it and more, yet there was never any thought of letting it go without a struggle. Longitudinal studies were becoming better known and better understood; there was a lot of motivation to keep it going as the stakes of lifetime studies began to come into sharp focus.

  Also in 1954, Dana Farnsworth, M.D., a psychiatrist committed to the mental health of students, replaced the retiring Arlie Bock as director of the Department of Hygiene, now renamed Harvard University Health Services. Frederic Wells had retired. Margaret Lantis had left in 1952 without replacement. In 1955, Clark Heath left the Study due to its financial uncertainty, and Farnsworth appointed Charles McArthur director of the Grant Study.

  THE CHANGING OF THE GUARD

  With the appointment of a social psychologist as director, the first great phase of the Grant Study came to an end. The early researchers had theorized that development was driven by intellectual ability, physical constitution, and personality traits that invariably resulted from them. This model was now abandoned in favor of a model of relationship-driven development. Years later, when I first became director, Arlie Bock, who had not even included psychiatrists on the full-time staff of his Department of Hygiene, was surprised that I had discovered among his “normal” Grant Study boys alcoholics, personality disorders, and manic depressives. With tongue in cheek he complained to me in the hallway, “They must have been spoiled by you psychiatrists. They didn’t have problems like that when I was running the Study!”

  Dana Farnsworth had attended a one-room school in West Virginia, and graduated from Harvard Medical School in 1933. He had headed the health services at Williams College and M.I.T. He was widely considered the nation’s foremost expert on student mental health, and was far more sympathetic to psychiatry than Bock had ever been. As soon as he took the helm he hired five full-time and two part-time psychiatrists, and established relationships between them and each of Harvard’s residential houses and graduate schools. While Farnsworth gave little direct financial help to the Study, he did provide free office space, and was personally supportive of Charles McArthur and then of me.

  McArthur, a brilliant young Health Services psychologist, was completing his Ph.D. in social relations at Harvard when Farnsworth appointed him director of the virtually penniless Study. He held that position from 1955 until I, a research psychiatrist, succeeded him in 1972. When he began, he was the only member of the Study staff, and Harvard was paying him to be a clinician to the students, not to do psychological research. Between 1955 and 1967 he did manage to achieve continuity of funding by asking questions about smoking habits—“If you never smoked, why didn’t you?”—a nod to another tobacco-related patron, Philip Morris. And even though he sometimes had to supplement such outside funds with his own staff psychologist’s salary, he managed to carry on the work of maintaining addresses, keeping contact with the men, and sending out questionnaires. Between 1960 and 1970, however, there were only four publications, frequency of questionnaires dropped to every three years, and seventeen Grant Study men were noted to have been AWOL for more than a decade.

  McArthur’s research foci were smoking habits and social class differences among the Grant Study men. He was also interested in longterm validation of the Strong Vocational Test—at the time the best predictor of career choice and satisfaction available—and published several papers on his findings (see Appendix F). This was the first time that the Grant Study really capitalized on its prospective longitudinal design.30

  In 1967 the Grant Foundation awarded McArthur a three-year grant of $100,000 ($555,000 in 2009 dollars) to bring Lewise Gregory Davies, now married, back to the Study, and to obtain secretarial help in mailing out questionnaires. Davies’s return worked wonders. The seventeen nonresponders came back to the fold. In the forty-five years since then, only seven active members have withdrawn from the Study. No member has been entirely lost to follow-up, for publicly available records and alumni reports have allowed us to characterize the occupational success, marital status, and general life adjustment of those seven dropouts and the twelve who left before 1950. Only four of the nineteen dropouts are still alive; the Study has death certificates for thirteen of the other fifteen.

  LATER DATA ANALYSIS CONTINUES: THE TWENTY-FIFTH REUNION YEAR

  I was in my first years with the Study when the twenty-fifth reunion years of the College cohort began to roll around. These were the occasion for a major round of data analysis in which we considered the men’s adaptation to post-college life and above all to life after the Second World War. The reunions were also an opportunity to compare the College men with their classmates. In 1969 I devised a questionnaire that the Reunion Committee of the Harvard Class of 1944 distributed at the time of its twenty-fifth reunion.

  We already knew from freshman physicals that the Grant Study men did not differ significantly from their classmates in such attributes as height, visual acuity, eye color, hay fever, or history of rheumatic fever. Twenty-five years later, it appeared that they did not differ significantly in future occupation, either. A quarter of each class became lawyers or doctors; 15 percent became teachers, mostly at the college level; 20 percent went into business. The remaining 40 percent were distributed through such other fields as architecture, accounting, advertising, banking, insurance, government, and engineering. There were a few artists of varying persuasions. The same proportions held true of the Study men.

  But by the reunion years, critical differences had appeared, and some of them are summarized in Table 3.3. The comparison isn’t perfect. Alumni who have not enjoyed relatively smooth sailing in life, love, and career often decline to complete reunion questionnaires. While 92 percent of the Grant Study subjects completed this one, in the table they are being compared to a 70 percent sample of their classmates, probably self-selected for health and success. Even with this selection bias, however, the Study men seemed to have striven harder for success in college and in life than their classmates. Their mean IQ of 135 was only minimally higher than the 130 of their classmates, but a far higher percentage of them graduated with honors. They tended to take fewer sick days. A point of interest: this 1969 analysis was the first time a computer was used to analyze Grant Study data. At that time, the Harvard computer (quite a concept now!) filled a whole building. Data had to be submitted on eighty-column punch cards, often carried half a mile through the snow to the computing center, and data turnaround could take several days.

  Table 3.3 Responses to a Questionnaire Distributed at the 25th Reunion of the Harvard Class of 1944

  Very significant = p<.001; Significant = p<. 01; NS = Not significant.

  Given their apparent ambition and greater striving, it wasn’t surprising that the Grant Study men were more likely than their classmates to have exceeded the success of their fathers by the time they were in their late forties. However, the selection of the Grant Study men favored conventional definitions of success. Stoics outnumbered Dionysians in the Study; the achievement bias in the selection process probably worked against stable youngsters who were more interested in private contentments than in chasing brass rings, and artists whose development takes longer and is frequently less remunerative. Furthermore, capacity for intimacy had been valued less highly in the selection process than the capacity to grin and bear it. One staff member defined a “healthy” person as “someone who would never create problems for himself or anyone else,” and many of the Grant men lived up to that ideal. One boasted that what he enjoyed most in life was “be
ing beholden to no one and helping others.”

  Even taking this into account, the extent of the men’s successes as they entered middle age was striking. Four members of the College cohort ran for the U.S. Senate. One served in a presidential cabinet; one was a governor, and another was president. There was a best-selling novelist (no, it was not Norman Mailer, even though he was Harvard class of ’43), an assistant secretary of state, and a Fortune 500 CEO. And while the average man in the College cohort had the income and social standing of a successful businessman or physician, he displayed the political outlook, intellectual tastes, and lifestyle of a college professor. At age forty-five, the Grant Study men’s average income was about $180,000 (in 2009 dollars) a year, but fewer than 5 percent of them drove sports cars or expensive sedans. Despite their economic success, they voted for Democrats more often than for Republicans, and 71 percent viewed themselves as “liberal.”

  I JOIN THE GRANT STUDY AND RE-INTERVIEWS BEGIN

  I had arrived at the Grant Study in 1966, supported by an NIMH Research Scientist Development Award. McArthur allowed me to design the next two regular questionnaires. Perhaps the most important change I instituted at the time was to begin collecting physical exam data every five years, including chest x-rays, EKGs, and standard blood and lab work. The exams began as the men turned forty-five (1965–1967) and have continued since then, through age ninety (2010–2012). This prospective record of objective physical health distinguishes the Grant Study from the other great longitudinal studies of personality development.

  My approach differed from Charles McArthur’s. I wasn’t a social psychologist, but an M.D., and in psychoanalytic training. My early studies of recovery from heroin addiction and acute psychosis had left me impressed by the involuntary adaptive coping mechanisms that resilient men employed. My youthful preference for blacks and whites and either/ors had given way to an appreciation of—and a serious interest in—the reality and power of incremental adaptation. Now, at the Grant Study, I could investigate these mysteries to my heart’s content.