She Has Her Mother's Laugh: The Powers, Perversions, and Potential of Heredity (40 page)

These abilities are not a random, disconnected collection of skills, tests show. When scientists give people tests on different abilities, their scores are correlated. If people are very good at recalling information from stories, for example, they also tend to do well at recalling words from lists. Different tests for logical reasoning correlate with each other as well. In turn, these broad abilities—such as reasoning, memory, spatial ability, processing speed, and vocabulary—correlate with each other. Psychologists can
measure this underlying correlation with a single factor known as
g
, short for general intelligence.

It may seem strange that the speed at which people correctly hit a button can roughly predict
whether they can recognize a word like
defenestrate.
Yet the deep connections revealed by intelligence researchers are among the
best-replicated findings in all of psychology.

Intelligence is also a surprisingly durable trait. On June 1, 1932,
the government of Scotland tested almost every eleven-year-old in the country—87,498 all told—with a seventy-one-question exam. The students decoded ciphers, made analogies, did arithmetic. The Scottish Council for Research in Education scored the tests and analyzed the results to get an objective picture of the intelligence of Scottish children. Scotland carried out only one more nationwide exam, in 1947. Over the next couple of decades, the council analyzed the data and published monographs before their work slipped away into oblivion.

In 1997, an expert on intelligence named Ian Deary stumbled across a mention of the Scottish Mental Survey in a book. Given his line of work—and given that he worked at the University of Edinburgh—he was startled that he had never heard of it before. The book Deary was reading only mentioned the survey in passing, but it was enough to inspire him to find out more about it. Intelligence testing was so time-consuming that researchers typically could manage to examine small groups of people. Here was a test of an entire population. And all the eleven-year-old test takers who were still alive would now be seventy-six. At the time, psychologists were still debating how much intelligence tests taken in childhood said about people's later lives. If Deary could find some of them, he could give them the test once more and get an unprecedented measurement of this influence.

Deary's colleague, Lawrence Whalley, dug into the Scottish Mental Survey reports. Eventually his search led him to a basement stacked with boxes and files containing the original tests. He called Deary with the news. “
This will change our lives,” Deary replied.

Deary, Whalley, and their colleagues moved the 87,498 tests from ledgers onto computers. They then investigated what had become of the test
takers. Their ranks included soldiers who died in World War II, along with a bus driver, a tomato grower, a bottle labeler, a manager of a tropical fish shop, a member of an Antarctic expedition, a cardiologist, a restaurant owner, and an assistant in a doll hospital.

The researchers decided to track down all the surviving test takers in a single city, Aberdeen. They were slowed down by the misspelled names and erroneous birth dates. Many of the Aberdeen examinees had died by the late 1990s. Others had moved to other parts of the world. And still others were just unreachable. But on June 1, 1998, 101 elderly people assembled at the Aberdeen Music Hall, exactly sixty-six years after they had gathered there as eleven-year-olds to take the original test. Deary had just broken both his arms in a bicycling accident, but he would not miss the historic event. He rode a train 120 miles from Edinburgh to Aberdeen, up to his elbows in plaster, to witness them taking their second test.

Back in Edinburgh, Deary and his colleagues scored the tests. Deary pushed a button on his computer to calculate the correlation between their scores as children and as senior citizens. The computer spat back a result of 73 percent. In other words, the people who had gotten relatively low scores in 1932 tended to get relatively low scores in 1998, while the high-scoring children tended to score high in old age. If you had looked at the score of one of the eleven-year-olds in 1933, you'd have been able to make a pretty good prediction of their score almost seven decades later.

Deary's research prompted other scientists to look for other predictions they could make from childhood intelligence test scores. They do fairly well at predicting how long people stay in school, and how highly they will be rated at work. The US Air Force found that the variation in
g
among its pilots could predict virtually all the variation in tests of their work performance. While intelligence test scores don't predict how likely people are to take up smoking, they do predict how likely they are to quit. In a study of one million people in Sweden, scientists found that people with lower intelligence test scores were more likely to get into accidents.

The long reach of intelligence suggests that it may have some deep
biological foundations. Some scientists have proposed that, in one way or another, different intelligence tests all probe
how efficiently the brain processes information. Some of the most compelling evidence for this theory comes from a simple exam in which
a shape flashes on a computer screen. The shape is made up of two vertical lines and a horizontal one sitting on top, like a sketch of the standing stones at Stonehenge. Each time the shape appears on the screen, one of the vertical lines hangs down lower than the other. Volunteers have to indicate which one is longer.

If the shape flashes too briefly, people guess at random. But if the shape lingers for long enough on the monitor, they can give the right answer most of the time. On average, people can perceive a shape correctly if they can see it for around a tenth of a second. But that sliver of time varies a little from person to person. In one study, scientists found
some people needed just 0.02 seconds, while others needed 0.136 seconds.

Time and again, researchers have found that there's a correlation between intelligence and inspection time. People with lower intelligence scores tend to need more time in order to recognize the shape. It's not an iron law (the correlation is about 50 percent), but the link is strong enough to lead scientists to wonder if there's something in common lurking under inspection time and intelligence.

Even the simplest mental operations require our brains to fire neurons in a network of regions scattered throughout our heads. Regions of the middle and back of the brain gather perceptions and organize them. They then feed their own signals through long fibers—known as white matter tracts—to the front of the brain. There, we have regions that specialize in problem-solving and decision-making. The frontal regions then talk back to the others so that they can fine-tune their gathering of perceptions.

But Deary's research raises the possibility that the roots of intelligence dig even deeper. When he and his colleagues started examining Scottish test takers in the late 1990s, many had already died. Studying the records of 2,230 of the students, they found that the ones who had died by 1997 had on average a lower test score than the ones who were still alive. About 70 percent of the women who scored in the top quarter were still alive, while
only 45 percent of the women in the bottom quarter were. Men had a similar split.

Children who scored higher, in other words, tended to live longer. Each extra fifteen IQ points, researchers have since found, translates into a 24 percent drop in the risk of death.

In a 2017 study, Deary and his colleagues drilled further into this effect. This time around, the researchers took advantage of the second mental survey that the Scottish government carried out on eleven-year-olds, in 1947. This younger cohort was too young for World War II, so more of them were able to survive into old age. Deary and his colleagues combed over records for more than sixty-five thousand of the test takers. They noted not only who died but
how
they died.

As before, the researchers found that lower intelligence test scores raised people's risk of death. But when they broke down the deaths into the major causes, they found the same rule held true across the board. The people who scored in the top 10 percent were two-thirds less likely to have died from respiratory disease than those in the bottom 10 percent. They were
half as likely to have died from heart disease, stroke, and digestive diseases.

It's possible that intelligence test scores measure how well people can take care of themselves. As adults, they may tend to make somewhat more money, which they can spend on their health. Or they may be slightly better able to understand the information that their doctors give them. But the influence of intelligence on longevity is so broad that Deary has proposed a deeper connection. Scores on intelligence tests may gauge some broad feature of human biology, in the same way a thermometer or a blood pressure reading does. The efficiency in the brain may have something in common with how well other parts of the body run. And this “
system integrity,” as Deary calls it, may help determine how long the whole system runs before falling apart.

Early intelligence researchers were firmly convinced that heredity had overwhelming control over intelligence. “
A person can no more be trained to have it in higher degree than he can be trained to be taller,” the
English psychologist Charles Spearman once said. Yet the evidence they offered for these claims was as firm as a soft-boiled egg. Simply running through the annals of great English men could never deliver the proof Galton hoped for. The class prejudices of the early 1900s made it possible for an extravagant fiction like
The Kallikak Family
to be taken seriously for years
.

By the 1920s, however, the science of heredity had matured far enough that researchers could begin to study intelligence in a meaningful way. This was the time that twins were emerging as a tool for studying heritability, and intelligence researchers followed the example set by those studying height. Three Chicago scientists, named
Frank Freeman, Karl Holzinger, and Horatio Newman, gave intelligence tests to fifty identical and fifty fraternal twins. They found that the identical twins had closer scores to each other than the fraternal twins—suggesting that intelligence was indeed heritable.

The Chicago researchers realized there was another way they could use twins to study intelligence. Rather than compare twins that had grown up together, they could look at how strongly nature influenced twins who had been raised apart. The scientists put out advertisements for adult twins who had been separated as children, typically adopted by different families. Nineteen pairs responded.

One pair, whom the scientists
referred to only as Ed and Fred, had grown up in different states. One day, someone walked up to Ed and said, “Hello, Fred, how's tricks?” Ed had dim memories of a long-lost brother, and so he decided it was time to track down this mysterious Fred. When the twins reunited, they were shocked to find they had both dropped out of high school and become electricians. When the scientists gave the twins an IQ test, Ed scored 91, Fred 90.

When the scientists studied other separated twins, they found similarly close scores. Yet Freeman and his colleagues were cautious about the lessons they would draw from their research. “
We shall be satisfied if we have succeeded in tracing a few of the threads in the tangled web which constitutes the organism we call man,” they said at the end of their 1937 book,
Twins: A Study of Heredity and the Environment.
Despite the similarity of
Ed and Fred, they still agreed with the dictum “What heredity can do, environment can do also.”

At around the same time in London, a British psychologist named
Cyril Burt was studying the intelligence of twins as well. Burt had been set on a course into psychology as a boy. His father, a doctor, sometimes let him tag along on house calls, and on one trip he met Francis Galton. After talking with Galton, Burt bought one of his books, and his fate was sealed. Burt studied at Oxford and became a teacher, carrying out psychology research on the side. In 1912, he was appointed the London City Council's first psychologist, where he used intelligence tests to identify low-scoring children in need of special education.

Burt wanted to understand to what extent intelligence was “
a thing inborn and not acquired.” Inspired by Galton's proposal, he searched among his students for twins who separated early in life. In 1955, Burt published a study on twenty-one pairs. Their intelligence test scores were more similar than those of siblings raised in the same house.

Eleven years later, Burt published an even bigger study on fifty-three pairs of twins. The results were the same. Burt used the test scores to estimate intelligence's heritability as being 80 percent. Where Freeman and his colleagues had caviled about nature and nurture, Burt brought down his psychological gavel with a loud crack. Heredity explained most of the differences between people's intelligence test scores, he declared.

Among the people who read Burt's 1966 paper was a Princeton psychologist named Leon Kamin. “
Within ten minutes of starting to read Burt,” Kamin recalled later, “I knew in my gut that something was so fishy here that it just
had
to be fake.”

The results were too pat. They didn't seem to have anything to do with “the messy nature of the real world,” as Kamin put it. Kamin dug into Burt's research and found suggestions of fraud. In Burt's 1955 and 1966 studies, twenty of the correlations were identical. In both studies, the correlation between identical twins raised apart were 0.771. That three-digit coincidence alone would have been highly unlikely. Twenty coincidences were astronomically improbable. Elsewhere in Burt's work, Kamin found more
signs that Burt had made up much of his twin results. Burt even published papers under fake names to create the illusion that other scientists supported his findings.

In 2007, a Rutgers University psychologist named
William Tucker offered an explanation for Burt's long con: He was a eugenicist from start to finish. In 1909, Burt published a study showing that upper-class schoolboys scored higher on intelligence tests than lower-class ones. Their different upbringing could play only a small part in that difference, Burt declared. “
The superior proficiency at Intelligence tests on the part of the boys of superior parentage was inborn,” he wrote.