Triumphs of Experience: The Men of the Harvard Grant Study (39 page)

It is of great interest, then, that maturity of defensive style did not seem to be affected by socioeconomic status, intellectual ability, or gender. It is true that the College sample looked a little better than the Inner City men in this regard (11 percent and 25 percent respectively used predominantly immature defenses). But that can be explained by the original selection process; the College men were selected for mental health, while the Inner City men were not; in fact they had been deliberately matched with delinquents. Mental health and defenses are closely correlated.

A less distorted view of the effect of privilege on defensive style can be achieved through within-group comparison—that is, by comparing the members of one group to each other. In this way initial selection bias is circumvented.
Table 8.5
examines the effect of social class, IQ, and education upon differences in defensive maturity within
three
different groups. The associations are insignificant. Even the relationship of a warm childhood environment to maturity of defense is less than one might expect.

Table
8.5
Correlations Between Maturity of Defenses and Biopsychosocial Antecedents

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

*
Thirty men with IQs less than 80 were excluded.

The effect of cultural diversity was tested in the following manner. Sixty-one percent of the Inner City men had parents who had been born in a foreign country, but the men themselves had all grown up in Boston, were fluent in English, and had been sampled and studied in the same way. Thus it was possible to vary ethnicity and culture of rearing while holding other demographic variables constant. In some facets of adult life, parental ethnic differences among the Inner City men were seen to exert a profound effect. For example, as I’ll discuss in
Chapter 9
, men of white Anglo-Saxon Protestant (WASP) and Irish extractions had rates of alcohol abuse five times those of men of Italian extraction. But there was little cultural difference in defensive style. Dissociation was the only defensive style that appeared to be used significantly more by the WASPs than by Italians.
13
(Dissociation was also the defense with the lowest rater reliability.)

But if culture appears to have little effect on defensive style, this is not true of biology. The central nervous systems of some of the Inner City men had been impaired by chronic alcoholism. (By this I do not mean acute intoxication; most men were quite sober when interviewed.) In addition, some men had possible early cognitive impairment, as suggested by IQs less than 80. Both these groups exhibited significantly less mature defensive styles than the rest of the Inner City men. All immature defenses were two to four times as common in these two compromised groups as in the unimpaired sample.

CONCLUSION

Empirical investigation provided clear answers to the Study’s three major questions about involuntary coping mechanisms. First, maturity
of
defenses can be rated reliably. Second, maturity of defenses demonstrated predictive validity toward future mental health. Third, maturity of defenses is independent of social class and gender, but is affected by biology.

Defense mechanisms are not just one more dogma of the psychoanalytic religion. On the contrary, the brain’s mechanisms of involuntary adaptation are a fit subject for serious study by social and neurological scientists. But, as with all matters of lifetime development, long and deep access is needed to study them. It was the unusual longitudinal and naturalistic nature of the Grant Study that permitted the conclusions in this chapter. Fledgling efforts are now under way to image what the brain does while deploying defenses, but scientists will likely have to wait upon advances in brain imaging technology for further confirmation.
14

9

ALCOHOLISM

Remember that we deal with alcohol—cunning, baffling, powerful!

—ALCOHOLICS ANONYMOUS

ALCOHOLISM IS A DISORDER
of great destructive power. Depending on how we define it, it afflicts between 6 and 20 percent of all Americans at some time in their lives. In the United States, alcoholism is involved in a quarter of all admissions to general hospitals, and it plays a major role in the four most common causes of death in twenty-to forty-year-old men: suicide, accidents, homicide, and cirrhosis of the liver.
1
The damage it causes falls not only on alcoholics themselves but on their families and friends as well—and this damage touches one American family in three. Life is not a cog railway that we step onto at birth and off at death, secure in the knowledge that we are safe from accidental derailments and the tug of gravity. No matter how blessed by good fortune we start out or how blighted by its lack, our circumstances can always change, and so can the conditions under which we meet them. This is the Study’s sixth lesson, and much of what we learned about it came from our prospective longitudinal investigation of alcohol use and abuse.

The Grant Study’s involvement in alcohol research was one of the silver linings of our perpetual anxiety about funding, and true sterling. Without it, I would not have been forced into this last ten years’ reexamination of marriage, divorce, and the development of intimacy. Yet that reexamination called into question not only some cherished assumptions
of my own, but also assumptions that predated my tenure at the Study, and the received wisdom of several generations. Lifetime studies are bread cast upon the waters. You can’t know in advance everything you should be finding out. But on the other hand, some of what you find out and have no idea what to do with may turn out to be invaluable unforeseen years later. The work with alcohol was like that.

It isn’t easy to identify who is and is not an alcoholic. Until now, most major longitudinal studies of health (for example, the Framing-ham Study in Massachusetts and the Alameda County Study in California) have taken into account only alcohol consumption, not alcohol abuse.
2
Unfortunately, as I’ve said before, reported alcohol consumption identifies alcohol abuse almost as poorly as reported food consumption reflects obesity. In contrast, the Grant Study has always focused on alcohol-related
problems.
Where alcohol is concerned, it is what people do, not what they say, that is important.

Our study of the College and Inner City cohorts is the longest and most thorough study of alcohol abuse in the world. It has established answers to seven major questions:

1. Is alcoholism a symptom or a disease?

2.
Is alcoholism environmental or genetic?

3. Are alcoholics premorbidly different from nonalcoholics?

4. Should the goal of alcohol treatment be abstinence?

5. Can “real” alcoholics ever drink safely again?

6. How can relapse be prevented?

7. Is recovery through AA the exception or the rule?

In
several instances, the Study’s longitudinal findings differ from those of well-respected cross-sectional studies.

METHODS

The Study’s unique structure gave it three advantages as we made our usual effort to replace opinion with science. First, the men were followed for their entire lifetimes—a rarity but a necessity, because alcoholism is a relapsing and evolving disease. Second, the Study quantified alcoholism not by reports of quantity or frequency of drinking, but by objective numbers of alcohol-related problems. Third, over its own lifespan the Study has had between thirty and fifty contacts with each of its members, which greatly facilitated the collection of this data.

A few notes on how the men were studied. The questionnaires they received every two years asked if they, their friends, their families, or their physicians had expressed concern about their drinking, and whether and for how long they had ever
stopped
drinking (evidence not of control, but of loss of control). At interview, alcohol abuse or its absence was always specifically recorded. When the men reached forty-seven, 87 percent of them participated in a two-hour semi-structured interview with a detailed twenty-three-item section on lifetime problem drinking.
3
Since forty-seven, they have undergone physical examinations every five years. Any man not previously classified as an alcohol abuser who answered yes twice in a row to two or more of the four concern questions, or who through interview or telephone contact acknowledged alcohol abuse, or whose physical exam revealed evidence of alcohol abuse, was classified an alcohol abuser. The age that each participant first met DSM-III criteria for alcohol abuse was estimated from all available data: questionnaires, any relevant court records, social service data, family interviews, etc.

(In
1962, before the Inner City cohort joined the Harvard Study of Adult Development, arrest records and records of psychiatric hospitalization, if any, had been searched for more than 95 percent of the men, who were at high risk for alcoholism, and for the preceding two generations of their families.
4
These are data that are well-nigh irreplaceable, as recent privacy legislation now precludes such searches.)

By mining interviews, clinical data, objective documents, and self-reports obtained from the participants by clinicians experienced in treating alcoholism, we were able to establish both a categorical and a dimensional scale of alcohol use for the men of both cohorts. The categorical scale was derived from DSM-III, the third edition of the
Diagnostic and Statistical Manual of the American Psychiatric Association;
that was the version current in 1977–1980, when this analysis was taking place.
5
It distinguished three types of alcohol use:
social drinking
(that is, no chronic problems related to alcohol use),
alcohol abuse
(chronic problems but no physiological dependence), and
alcohol dependence
(presence of withdrawal symptoms or hospitalization for detoxification). In this chapter, I will use the term
alcoholism
to refer to both of the latter two categories.

The dimensional scale, the Problem Drinking Scale, or PDS, assessed problem drinking on a continuum of severity by means of sixteen equally weighted questions (similar to those of the Michigan Alcoholism Screening Test).
6
The PDS inquired about social, legal, medical, and job problems caused by alcohol abuse. It also asked about blackouts, going on the wagon, seeking treatment, withdrawal symptoms, and problems with control. Scores of 4–7 on the PDS usually met the DSM-III criteria for “alcohol abuse,” and scores of 8–12 usually met the criteria for “alcohol dependence.” Men with fewer than 4 lifetime problem points on the dimensional scale were usually classified as social drinkers.

Observant readers will note that the numbers in this chapter vary
sometimes
from those in older reports. This is because in those earlier analyses we included all the Study men who had ever met DSM-III criteria for alcohol abuse (153 out of the 456 original Inner City men and 56 of the original 268 College men). In refining the original analysis for this book, however, men whose problem drinking scores were borderline (3 or 4) and who abused alcohol for less than five years and who returned to social drinking for the rest of their lives (13 Inner City and 2 College men) were reclassified as social drinkers.

We have excellent and inclusive death data, including death certificates for all Study members, including those who withdrew, except for two who died abroad. Survival or mortality has been ascertained and documented through the National Death Index or credit agencies, whichever was applicable, and death certificates obtained. Data from death certificates, and recent physical examinations from participating Study members, were used to infer major causes of death.

We assessed the alcohol status of all cooperating Study members, whether or not they had been alcoholic in the past, every year between the ages of twenty and seventy, using the biennial questionnaires and triangulating them with other material. (Since the Glueck men had not been personally followed between thirty-two and forty-five, for those years we had to depend on history and public records, including records of arrests.) We categorized the alcoholics as follows:
Abstinent:
less than one drink (0.5 ounces of ethanol) a month for a year.
Return-to-controlled-drinking:
a former alcohol abuser consuming more than one drink a month for at least three years with no reported problems.
Continued alcohol abuse:
Clear past history of sustained alcohol abuse and one or more acknowledged problems caused by drinking in the past three years. When data was missing for three years, the yearly status was rated as unknown. Data on alcohol abuse for every man was obtained between twenty and forty times over sixty (on average) years of observation. Nonresponders for two consecutive ques
tionnaires
were interviewed in person or by telephone. Men were classified as
Dropped
if they asked to withdraw or ignored questionnaires and follow-up telephone calls for ten years.