Nancy Segal, Born Together—Reared Apart: The Landmark Minnesota Twin Study, Harvard University Press, 2012, 410 pp. $49.95.
The Minnesota Study of Twins Reared Apart (MISTRA) was one of the most important psychological studies of the last 50 years. It began in 1979, at a time when it was widely believed that intelligence and personality were almost infinitely malleable by the environment. By the time the study ended 20 years later, it had played a key role in overthrowing this dogma. It established beyond any doubt that genes are crucial to who we are.
Born Together—Reared Apart is a detailed account of how MISTRA did its work and what was learned, and Nancy Segal is surely the best person to have written it. Herself a fraternal twin, she worked full time on the project from 1982 to 1991, and worked closely with it until it ended. She is a prominent twin researcher in her own right, and continues to make discoveries in the field.
MISTRA, carried out under the leadership of Professor Thomas J. Bouchard of the University of Minnesota, was based on the unique nature of twins. Monozygotic or MZ twins are genetically identical, while dizygotic or DZ twins share approximately half their genetic material, just like ordinary siblings. MZ twins occur when a single fertilized egg splits in two and becomes two fetuses in the same womb. DZ twins occur when there is double ovulation, and two eggs are fertilized and become fetuses.
Long before MISTRA, it was well known that MZ twins acted very similarly, but environmental theory held that this was because they had grown up in the same environment and had been treated similarly. Sometimes, however, twins have been separated at birth and reared in different households. If separated MZ twins have similar abilities and personalities, this cannot be because they had the same family environment; it must be because of their genes.
The basic design of MISTRA was to find monozygotic twins reared apart (MZA) and to see how similar they were. MISTRA also did the same for dizygotic twins reared apart (DZA), which they used as controls, for a total of 81 MZA and 56 DZA pairs. DZA twins were used as controls so that a shared gestation environment could be ruled out as the reason MZA twins are so similar.
The criteria for selection were that the twins had been separated before age four and spent their formative years apart. At the time of testing the 137 twin pairs had spent an average of 95 percent of their lives apart, and in some cases had grown up in different countries.
Most of the twins had been separated because they were illegitimate. Their mothers gave them up for adoption and they were taken into different homes. In many cases they never knew they had a twin, and each story of reunion was different. Some twins found each other because one started looking for his birth parents and learned he had a twin. Thirteen percent of the pairs learned of each other because of mistaken identity; someone who knew one twin bumped into the other and thought it was the same person. This happened almost exclusively with MZA twins; only one pair of DZA twins was so similar they caused this kind of confusion.
MISTRA learned about reunited twins mainly through news reports, and invited the twins to come to Minnesota for testing. After MSITRA became better known, some twin pairs who had discovered each other contacted the study team directly. In a few cases, MISTRA found separated twin pairs by looking through adoption records.
Thus, there was considerable variation in the amount of time twin pairs had been reunited before MISTRA tested them, and critics complained that this gave them time to study each other and to become more similar. In fact, this variable made it possible to test whether time spent together before testing really did make twins more similar. It did not.
Almost invariably, the reunion of identical twins was a joyful event. Some spoke of the “ecstatic shock” of discovering someone so similar. After just a short time together, most MZAs felt closer to their twin than to adoptive siblings they had known all their lives. When MISTRA itself brought MZAs together for the first time, staff filmed the reunions to capture this unique encounter. Reunions of DZA twins were not nearly so life-changing.
The twins were brought to the University of Minnesota and put through a full week of testing. First they were given medical tests to determine whether they were MZ or DZ. Even people who work with twins cannot always tell identicals from fraternals. Prof. Segal writes that one of the best ways is to examine ear folds; if they are the same, twins are identical.
During the evaluation week, twins completed about 15,000 paper-and-pencil test items, and were examined for everything from gum disease and tooth formation to heart function and blood composition. There have been other studies of twins reared apart, but none that gathered so much information. MISTRA data are still being analyzed for research papers.
MZA twins were so remarkably similar that MISTRA attracted a lot of media attention, but this did not lead to generous funding. It cost an average of $7,000 to fly a pair of twins to Minnesota for a week of tests, and MISTRA often ran out of money. There was intense resistance to research that contradicted the orthodox belief that in the right environment we can all be made happily equal and equally happy. One grant-application reviewer for the National Science Foundation wrote that MISTRA would “fan the controversy regarding heritibility [sic] of intelligence . . . rejection is the only intellectually defensible course for NSF.”
In the end, the Pioneer Fund, which is known for supporting such scholars as Arthur Jensen, Philippe Rushton, and Richard Lynn, provided $1.4 million of the total $2.3 million that MISTRA needed. Critics complained that Pioneer money was “tainted,” but Harry Wehyer, the New York lawyer who ran the fund, took no part in research design or interpretation of data. As Prof. Bouchard noted, “If not for Pioneer we would have folded long ago.” The critics of the Pioneer Fund were really critics of the research itself.
Although Prof. Segal does not put it in these terms, MISTRA yielded what amount to two different kinds of findings: quantitative and impressionistic. The former come from personality, intelligence, medical, and other testing, whereas the latter include the almost eerie, unmeasurable ways in which MZA twins are alike.
The first twin pair MISTRA evaluated was particularly striking. The two men met when they were 39, and found that both had been in law enforcement but were now working as firemen. Both had loved math in school and hated spelling. Both did woodworking as a hobby, and their favorite vacation spot was Pas Grille Beach in Florida. One had named his son James Alan and the other had named his James Allan. They looked very much alike, had the same smoking habits, and always held a beer can with a pinky under the can. Both had put on 10 pounds at the same age for no apparent reason.
Not all twins were so alike, but this book is full of astonishing similarities. In one MZA pair, one twin was reared in Germany and the other in Trinidad, and they had never met before they came to Minnesota for testing. When they arrived at the airport each was wearing a light blue shirt with epaulettes, and wire-rimmed glasses. They both collected rubber bands, which they wore around their wrists, and washed their hands both before and after using the bathroom. Both liked to startle people by sneezing loudly in elevators.
One pair of MZA women both wet the bed until age 12 or 13. When they were teenagers they started having nightmares about the same things: fishhooks and doorknobs. Both had problems with nightmares for more than ten years.
One pair of MZA men had been overweight until middle school and then became quite thin. They had speech problems for which they received therapy in kindergarten or grade school. Both were diagnosed as hyperactive at about the same age, and both were actively and openly homosexual.
A pair of female MZA twins from Australia found each other because of a case of mistaken identity. They both worked as fashion buyers for competing department stores, and a customer accused one of moonlight for the competition. They were both very elegant, dressed with the same style and the same kind of jewelry, smoked the same cigarettes, and had the same hairstyle, posture, tastes, and speaking voice. One MZA pair of male twins were both fitness fanatics who ran their own body-building gyms. MZA twins generally have the same posture and arrange their hands and legs in the same way while DT twins do not.
Prof. Bouchard, who ran MISTRA, once had occasion to meet a man who had run a smaller-scale MZA study in Denmark in the 1960s, and asked him if he had found such astonishing similarities. The man replied that he had, but he did not report them because was no way to measure such similarities—and he was afraid no one would believe him.
Prof. Segal writes that it was “thrilling” to get to know MZAs and discover how similar they were, but she, too, was frustrated because it was not possible to measure or assess similarities in complex behavior. She notes that when she interviewed MISTRA people to write this book, many looked back with nostalgia on the excitement of their discoveries. One researcher who administered intelligence tests to the twins wished that he had filmed them taking the tests. As he wrote:
I sat quietly behind them. The strategies [for answering test questions] were so different between pairs but within the MZA pairs they were so similar. Both twins vocalized or turned around or stared at the screen or solved the problems quickly. It was amazing. I smiled to myself when I saw these things, thinking no one would believe me.
Of course, there were many findings that could be quantified, the most obvious being intelligence. There is no better way to measure the heritability of intelligence than to study MZA twins. (For a detailed discussion of the concept of heritability, please see this video.) Because their environments are completely different—though not so different as to include malnourishment or physical abuse—similarities in IQ can have only genetic causes. Test results of such twins are often so similar that it is like testing the same person twice, and MISTRA yielded a heritability of 0.70 for IQ. This is higher than figures in the 0.50 to 0.60 range that have come from other studies, but the difference is no doubt due to having studied adults rather than children. IQs of children can be affected by their family environments, but by adolescence, the effect of genes and non-shared environment (see below) dominate.
The test results that certainly caused the most surprise were measures of personality. At the time, it was common to assume that personality was formed almost exclusively by family influence. It is not; it is formed in about equal parts by genes and by what is called “non-shared environment,” or the micro-environment each person makes for himself. Parents think they have a lot of influence over how their children turn out, but they flatter themselves.
MZT twins (identical twins reared together) have very similar—but not identical—personalities. People always assumed the similarities came from growing up in the same environment. But MZA twins also have very similar—but not identical—personalities, and there is no detectable difference in the degree of similarity between twins who grew up together and twins who grew up in different families—sometimes in different countries. The household, or the “shared environment,” has very little effect on personality, at least by the time people are adults.
Likewise, when biologically unrelated children are adopted and reared in the same home, they may resemble each other slightly when they are small, but as they grow up they become as different as complete strangers. It is well known that shared environment can have an early effect on IQ as well. “Virtual twins,” or unrelated children of the same age who grow up together, have a correlation of 0.3 for IQ at age five, which declines to 0.11 at age 11, and to essentially zero by adolescence.
And yet the personalities of MZA twins are not identical. Similarities are due to genes, and the degree of similarity is a direct measure of heritability. Dissimilarities must be the result of environment. But remember: The similarities and dissimilarities are about the same for MZA and MZT twins—that is to say, growing up in the same household has little effect on personality—so what makes twins dissimilar is the non-shared environment. This is different for all people and not entirely understood, but is thought to consist of peer groups, teachers, extracurricular activities etc.
It is not possible to distinguish the effects of shared and non-shared environment with only MZT and DZT twin pairs. However, it can be done with data from studies of twins reared apart, and by assessing other kin relationships, such as parents and ordinary siblings. By means of a technique called biometrical modeling, it is possible to calculate the effect of genes, shared environment, and non-shared environment on personality.
The table below, from page 103 of the book, gives results for a standard personality assessment tool called the Multidimensional Personality Questionnaire. Non-specialists can ignore the “C” column, which is the additivity-epistatic component of genetic contribution. The numbers in the other three columns add up to very close to 1.00, and indicate the estimated percentage of individual variation in a trait that is accounted for by genes and environment. To take the first personality component, “Well-being,” 48 percent is under genetic control, 13 percent is due to the shared environment, and 40 percent due to nonshared environment. The genetic and environmental contributions vary somewhat by trait, but the shared environment never contributes more than the 19 percent found for “Social closeness.” The warmth of a child’s family seems to have some effect on his own levels of warmth and closeness.
Tests of the “big five” personality traits (openness to new experience, conscientiousness, extraversion, agreeableness, and neuroticism) as well as results from the MMPI (Minnesota Multiphasic Personality Inventory) also showed similar levels of heritability.
Needless to say, results of this kind infuriated egalitarians. The great liberal goal is to invent “programs” that will make every child grow up smart, law-abiding, hardworking, and happy. If intelligence and even personality are under substantial genetic control, and environmental influences come from micro-environments children make for themselves, it does not leave much room for uplift. Harvard psychologist Leon Kamin, who has devoted his life to trying to discredit the concept of race and the idea that anything other than such things as height and eye-color can be heritable, practically accused MISTRA researchers of fabricating data.
Other MISTRA evaluations found a very high heritability for drug and alcohol use—0.78—a figure that does not bode well for “programs” either. Traditionalism appears to have a heritability of 0.53. For measures of conservatism, there was a sex difference: 0.65 for men and 0.45 for women. The political views of women may therefore be more open to persuasion.
Most people adopt the religion of their parents, so this is under strong environmental control, but intensity of religious conviction seems to be about 0.50 heritable. Vocational interests are about 0.50 heritable, with very little family influence.
MISTRA also found that genes largely determine spelling ability and whether someone is a “morning person” or a “night person.” How well someone’s hand coordination improves with practice is also heritable, as is the frequency of headaches.
MISTRA contributed substantially to the “set point” theory of happiness. This is the view that we all have a more or less consistent level of happiness (or unhappiness). Life’s events make us temporarily happier or unhappier, but we tend revert to the base level, which appears to be under substantial genetic control. Education, achievement, marital status, income, religious faith, etc. have very little correlation with happiness. For MZA twins, the happiness level of one twin was a far better predictor of the happiness level of the other than were life circumstances.
Whether a man is homosexual or heterosexual appears to have a heritability of about 0.50, whereas the sexual orientation of women is under more environmental influence. Prof. Segal reports that none of the female MZA twins was concordant for homosexuality, and that the twin who was homosexual had had delayed puberty and was heavier. The author does not say so, but this suggests women may establish relations with other women if they are not so attractive to men.
Children who grow up together do not find each other sexually attractive—this anti-incest mechanism is known as the Westermarck Effect. Twins reared apart are not subject to this effect, and MISTRA found that DZA different-sex twins were often flirtatious when they were reunited. This is not surprising; genetic similarity theory predicts it. Biological fathers/daughters and mothers/sons can be sexually attracted if they have not lived together. In the case of reunited male MZA homosexual twins, attraction can lead to sexual relations.
MISTRA found great physiological similarities between MZA twins. The immune system is an example: “In spite of different environmental exposures to antigens [because of being reared apart], the predominant factor(s) determining total immunoglobulin and isotypic antibody levels in these twins was genetic rather than environmental.” Even when their diets were different, the dental histories for MZA twins were very similar, and MISTRA found a substantial genetic contribution to the likelihood of gum disease.
The electrical conductance of the skin changes according to the activity of sweat glands, which in turn, changes with emotions such as fear or anger (this is the principle of the lie detector). MZA twins were very similar to each other in this respect, but men were more similar than women. Male MZA twins also had closer body weights than female MZA twins, and there appears to be a fairly high shared-environment effect on body weight.
The genetic contribution to what we eat and how often is about 0.33, and individual differences in metabolism influence weight gain. Prof. Segal reports an interesting experiment that was not part of MISTRA. Twelve MZT male twin pairs were put on a diet for 84 days that was 1,000 calories more than their usual intake. They were not allowed to exercise. All subjects gained weight, but the amount varied from 9.5 to 29.3 pounds. The amount of weight gain of one twin was much closer to that of his co-twin than to the other subjects, which suggests that metabolism rates are strongly heritable. Other studies have found only a weak link between food intake and body size, which also suggests differences in metabolism have a big influence on weight gain.
MZ twins are about 50 percent more likely than the general population to be left-handed: 15 percent as opposed to 10 percent. About 75 percent of both MZA and MZT twin pairs had the same handedness, which suggested there is no shared environment effect on handedness. What caused the other 25 percent not to have the same handedness is not fully understood.
A Scientific Legacy
MISTRA has left an enduring scientific legacy. Prof. Segal reminds us of a tragic event from 1965 that would surely never happen today. That year, a boy’s penis was destroyed during a bungled circumcision. Psychologist John Money convinced the parents that if their son were surgically given a vagina, treated with estrogen, and raised as a girl, he would grow into a normal woman. The project was a failure from the start, though Dr. Money claimed for years it had been a success. The subject of this sequence of misfortunes, then living openly as a man named David Reimer, killed himself in 2004.
It would be hard to find an example of greater faith in the power of environment, but such thinking was common 40 years ago. Today, thanks to studies such as MISTRA, no psychologist would attempt such a thing.
Unfortunately, a scientific legacy is not enough. The people who make laws, run schools, and write editorials still act as though MISTRA—and a hundred studies like it—had never been done.
Perhaps the most obvious example of this is the 2001 No Child Left Behind Act. Passed by huge majorities in both houses, the law requires that all normal students—those not in “special education”—read and do math at a level of “proficient” or better by 2014. In 2002, the year after the law passed, only 36 percent of 12th graders were “proficient” according to the National Assessment of Educational Progress exam (NAEP), the most commonly used standard. Not surprisingly, by July 2012, 24 states had already received federal waivers from the law’s requirements, and 13 more states had applied. Anyone with the slightest understanding of the heritability of intelligence knows that the goal of 100 percent “proficiency” was fantasy.
At the official level, the United States is still mired in 1960s illusions about universal human perfectibility. No one knows how long it will take to shed those illusions, but at least among scientists, MISTRA played a huge role in sweeping out the cobwebs.