by Bryan Caplan
Since researchers bundle together direct and indirect effects, their results are a lot stronger than they look. Suppose an adoption study finds that nurture doesn’t influence smoking. A parent could not reasonably object, “Right, but since peers influence smoking, and parents influence peers, I still matter.” The flaw in this argument is that a normal adoption study includes any and all indirect effects. If the “parents influence peers, peers influence smoking” channel worked, then an adoption study would report that parents matter. The same goes for neighborhoods. If parents choose their neighborhood, and neighborhoods affect kids’ drinking or college attendance, then a twin study would conclude that nurture matters for drinking and college. So remember: Researchers’ failure to detect an effect doesn’t just undermine simple stories of parental influence; it throws cold water on whatever stories you might propose.
MEASURING NATURE AND NURTURE EFFECTS
It’s easy to understand what it means for nature or nurture to have zero effect. What does it mean, though, for nature or nurture effects to be “small” or “large”? The clearest measures come from a thought experiment I call “Switched at Birth”:Imagine you have an identical twin, but there’s a mix-up at the hospital: A nurse accidentally switches your twin with another family’s baby. You and the strangers’ baby grow up with your biological parents. Your twin grows up with the strangers. Decades later, the hospital discovers its mistake and arranges a meeting between you, your identical twin, and your accidentally adopted sibling.
To measure the effect of nature, just answer this question:
Suppose you’re higher on some trait—height, intelligence, income, conservatism, you name it—than 80 percent of your peers. How high on this trait should you expect your TWIN from Switched at Birth to be?
To measure the effect of nurture, just answer this question:
Suppose you’re higher on some trait—height, intelligence, income, conservatism, you name it—than 80 percent of your peers. How high on this trait should you expect your ADOPTED SIBLING from Switched at Birth to be?
Most twin and adoption studies report enough information to measure both effects; if you’re curious about the math, see the appendix at the end of this chapter on “Where the Nature-Nurture Effect Sizes Come From.” Intuitively, if nature didn’t matter at all, you would expect your separated twin to be average—in the 50th percentile. If nature were destiny, you would expect your separated twin to match you in the 80th percentile; you’d both be higher in the trait than four out of five peers. Similarly, if nurture didn’t matter at all, you would expect your adopted sibling to be average; if nurture were destiny, you would expect your adopted sibling to be in the 80th percentile, just like you.
I’m going to use Switched at Birth’s measures of nature and nurture over and over. Once we get to specific twin and adoption studies, I’ll usually tell you their precise predictions about your separated twin or adopted sibling. As shorthand, if your counterpart in the thought experiment is in the 51st through 55th percentiles, I’ll call that a “small” effect. If he’s in the 56th through 65th percentiles, I’ll call that a “moderate” effect. Anything more I’ll call “large.”
THE PARENTAL WISH LIST: CAN PARENTS MAKE THEIR DREAMS COME TRUE?
The children must be molded, shaped and taught /
That life’s a looming battle to be faced and fought!
—Mary Poppins
Parents are supposed to love their children unconditionally, but that doesn’t stop them from having an extensive wish list for the next generation. Some parents map out their kids’ future in great detail: You will go to Harvard, become a cardiologist, marry a nice Jewish girl, and live in the house across the street. You’re probably much more accepting, but you’d still prefer your kids to become healthy, smart, happy, successful, virtuous adults who share your values and appreciate what you’ve done for them. And if things don’t pan out, even the most accepting parents wonder, “Where did I go wrong?”
The underlying assumption is that parents have the power to grant their own wishes. Do they? How much influence do parents really have over their children’s traits? To answer these questions, I now climb upon the shoulders of giants—the hundreds of researchers, past and present, who have used twin and adoption methods to unlock the secrets of nature and nurture. Researchers have carefully studied every entry on the typical parent’s wish list. It’s time to confront their evidence, wish by wish.
WISH #1: HEALTH
My mom is a longtime antismoking crusader. When I was growing up in Los Angeles, she always requested the nonsmoking section—even if she knew quite well that the restaurant didn’t have one. My mother was so opposed to smoking that she offered her kids a cash prize—payable on our twenty-first birthdays—if we stayed away from tobacco. I never smoked, and my mom paid up. Yet to be honest, the prize was overkill: Smoking disgusted me from an early age. While my mom tried to shape my behavior, and got the result she wanted, that hardly shows that her bribe worked.
The same goes for a long list of health habits that parents struggle to instill. Parents pressure their kids to eat right, get fresh air and exercise, brush their teeth, and stay away from tobacco and drugs. Part of parents’ motivation is to protect their kids’ current health, but the typical kid is too full of life to expect much immediate payoff. The main rationale for health-related nagging is to instill healthy habits that last a lifetime. My parents put strict limits on sweets, even though I was a skinny kid with perfect teeth. However harmless and tasty candy might seem, they weren’t going to let me develop bad eating habits that would hurt me when I grew up.
If you take parents’ rhetoric literally, there are strong implications. For example, if health nagging really works, then parents should affect how long their children live. If you influence your child’s decision to smoke, and smoking affects life span, it follows that your child’s life span is partly under your control.
If parents want to convince themselves that their efforts pay off, it won’t be hard: Long life runs in families. But there are two explanations for family resemblance in health, just as there are two explanations for family resemblance in anything: upbringing is one, heredity is the other. When medical researchers use twin studies to weigh the importance of these competing explanations, nurture turns out to be far less important than parents think.
Parents don’t affect life expectancy. Major twin studies find no influence of family environment on life span. One looked at almost 3,000 pairs of Danish twins born between 1870 and 1900. By the time the study was conducted in 1994, virtually all of the twins were dead, so researchers could easily compare the longevity of identical versus fraternal twins. They found moderate genetic effects: If you lived longer than 80 percent of the population, you could expect your separated identical twin to live longer than 58 percent of us. Yet the Danish twin study found “no evidence for an impact of shared (family) environment.” Growing up in the same home, being nagged by the same people about the same things, does not make your life expectancy more alike.
Another study looked at the mortality of about 9,000 Swedish twins born between 1886 and 1925. Since the Swedish sample was younger, some of the subjects were still alive; the goal of the study was to study twins’ probability of surviving to a given age. The Swedish twins study found strong genetic effects, especially for males—and zero effect of upbringing.
Parents have little or no effect on overall health. There’s more to health than staying alive. Do parents make their children’s years healthier even if they don’t make them longer? Probably not. Researchers have examined twins’ objective health (what particular health problems they’ve had) and subjective health (how healthy they say they feel). A study of over 3,000 elderly Danish twins found moderate effects of heredity on hospitalizations and self-reported health, but no effects of family environment. Another team of researchers looked at about 2,500 Swedish twins and found moderate genetic effects on self-rated health, but small or none
xistent nurture effects. A smaller study of older female Finnish twins found moderate effects of upbringing on self-rated health, but no effect of upbringing on health ratings based on doctors’ examinations.
The most notable exception comes from the Swedish Adoption/ Twin Study of Aging. Researchers studied over 700 twin pairs, about half raised apart. For respondents seventy and older, they found a moderate-to-large nurture effect on objective health. If you were in the 80th percentile of objective health, you could expect your adopted sibling to be in the 65th percentile. For all younger groups, however, the nurture effect was zero.
Parents don’t affect height, weight, or teeth. We all want our kids to grow up to be tall and fit, with sparkling white teeth. Most of us try to make these wishes come true. We push healthy foods, limit sweets, encourage exercise, and smell our kids’ breath to make sure they use toothpaste. Parents probably recognize that genes play a role, but we clearly think we make a difference. Otherwise, what’s the point of telling your kids how big and strong they’ll be if they eat their vegetables?
In reality, genes strongly influence both height and weight, while upbringing influences neither. In the Swedish Adoption/Twin Study of Aging, twins raised together were about as similar in height and weight as twins raised apart, and identical twins were about twice as similar as fraternal twins—both strong signs that nature matters and nurture doesn’t. Suppose you’re six inches taller than average, and twenty pounds overweight. Expect your separated identical twin to be five inches taller than average, and fifteen pounds overweight—and your adopted sibling to be perfectly average.
A major survey article on the genetics of obesity confirms the power of nature and the impotence of nurture. Parents may not even affect the weight of young children, though one study finds a small family effect for girls in their tweens. If it seems hard to believe that parents are this powerless, remember the old saying that you can lead a horse to water, but you can’t make him drink. It’s a big pain to make a picky eater clean his plate—or keep South Park’s Eric Cartman on a low-carb diet. And if parents refuse their kids the food they crave, they’ll find it somewhere else.
What about teeth? Surely parents’ commitment to brushing, flossing, and regular checkups has to matter. A major study of adult Swedish twins looked at gingivitis, periodontal disease, and complete tooth loss. All three conditions showed moderate genetic influence. Family environment, in contrast, mattered only for the rare problem of losing all your teeth.
Parents might have a small effect on smoking, drinking, and drug problems. I understand why my mom paid me not to smoke. I don’t want my kids to smoke, either. I don’t care whether drugs are legal or not; I don’t want my kids to use them. When relatives give my seven-year-olds a little beer, it irks me. Still, the question remains: “How much can I as a parent do about it?” When I bite my tongue and let my kids take a sip of alcohol, do I risk their future?
The answers from twin and adoption research aren’t completely one-sided. Some conclude that nature fully explains why smoking, drinking, and drug use run in families. A study of over 3,000 male twins from the Vietnam Era Twin Registry found almost no family influence on nicotine and alcohol dependence; another study of the same group of twins found that nurture mattered for regular use of marijuana, but not amphetamines, cocaine, or sedatives. A study of Australian twins found that family environment made almost no difference for alcoholism. Researchers using the Virginia 30,000 sample—a massive study of Virginian twins and their families—concluded that being raised by smokers makes you a little less likely to share their vice. Maybe cigarettes are a lame way to rebel against parents who smoke.
But other researchers find that nurture plays a role as well. One team studied the tobacco, alcohol, and drug use of about 1,000 seventeen- and eighteen-year-olds from the Minnesota Twin Family Study. They found moderate family effects for boys, and large family effects for girls. A couple of other studies conclude that parents don’t affect drinking, but siblings do. One looked at about 1,000 adoptees and their families, and found that adoptees drank like their adopted siblings but not like their adopted parents. Another study of over 1,000 Koreans adopted by American families found that children raised by mothers who never drank alcohol were almost 20 percentage points more likely to never drink themselves.
What’s a parent to think? You should put more stock in the larger studies that find small or nonexistent nurture effects. If you have paternalism to spare, though, you might as well try to keep your kids off tobacco, alcohol, and drugs. You probably won’t change your child, but there’s an outside chance that you’ll make a big difference.
WISH #2: INTELLIGENCE
New parents get excited by the faintest signs of their babies’ intelligence: “She could crawl, use a spoon, and say no before she was one.” By the time they’re toddlers, we’re furtively comparing our kids’ brain power to their playmates’—and struggling to give our offspring an edge with books, educational videos, museum trips, and tutors. If our child does poorly in school, we comfort ourselves with the thought that “he just needs to apply himself.” We care so much about our kids’ intelligence that we try to increase it prenatally. One misreported study about “the Mozart effect” led moms to play the classics for babies and fetuses alike.
However, a large scientific literature finds that parents have little or no long-run effect on their children’s intelligence. Separated twin studies, regular twin studies, and adoption studies all point in the same direction.
The Minnesota Study of Twins Reared Apart reunited almost 100 separated identical twins and triplets and gave them two standard IQ tests. It found large effects of genes on adult intelligence. If you did better than 80 percent of the population on both tests, you should expect your separated identical twin to do better than 72 percent on one test, and 74 percent on the other. In contrast, when researchers compared the reared-apart twins to a control group of reared-together twins, the effect of nurture was barely detectable. If you did better than 80 percent of the population on both tests, you should expect your adopted sibling from Switched at Birth to do better than 56 percent on one test, and just 50 percent on the other.
The Swedish Adoption/Twin Study of Aging studied the intelligence of about 150 pairs of middle-age and elderly twins. Some of these twins were identical, some fraternal; about half were raised apart, half together. Researchers individually administered a fourhour test of spatial and verbal intelligence, memory, and processing speed to each twin. They found an even larger effect of genes on intelligence than the Minnesota study and confirmed the irrelevance of upbringing for adult intelligence: “Growing up in the same family does not contribute to similarity in cognitive abilities later in life.”
Another team of researchers tested about 1,600 reared-together adult twins from the Dutch Twin Registry. Identical twins got extremely similar scores on a standard IQ test, and the similarity between fraternal twins was about half as strong—implying a big effect of genes, and no effect of upbringing. In the Switched at Birth experiment, if you were smarter than 80 percent of the population, you should expect your separated twin to be smarter than 76 percent but your adopted sibling to be perfectly average.
In 1975, the Colorado Adoption Project began studying 245 adopted babies, their birth mothers, and their adoptive parents. It also set up a control group of 245 comparable babies being raised by their biological parents. By the age of twelve, adoptees raised in high-IQ homes were no smarter than those raised in average homes. The researchers double-checked this result four years later and found the same thing. The Texas Adoption Project, which spent decades studying over 300 adoptees and the families that adopted them, also found no effect of upbringing on the IQs of late adolescents.
From a parent’s point of view, these are strong results. Today’s Typical Parents strive to mentally stimulate their children and struggle to protect their brains from being turned to mush by television and video games. Yet by adulthood, the fruit of
parents’ labor is practically invisible. Children who grew up in enriched homes are no smarter than they would have been if they’d grown up in average homes.
WISH #3: HAPPINESS
Good parents want to make their children happy. In the short run, they often succeed. If I want to put big grins on my sons’ faces, I buy them ice cream. Works every time. But most parents are more ambitious. They believe that if they raise their children in the right way, they’ll grow up to be happy adults. Maybe the recipe is unconditional love; maybe it’s daily sermons about gratitude and “how much harder life was when I was a kid.” Parents differ on the best approach but agree that parenting makes a big difference for the happiness of the next generation.
Parents grossly overestimate their influence. The Minnesota Twin Registry gave personality tests to over 1,300 pairs of adult twins raised together. Fraternal twins’ happiness resemblance was barely perceptible: If you’re happier than 80 percent of people, your fraternal twin will typically be happier than 53 percent. If all of this were due to nurture, it wouldn’t be much. On closer look, though, the credit belongs to nature, not nurture. Identical twins were far more similar in their happiness than fraternal twins, leaving no room for nurture effects.
One of the main problems with happiness tests is that the subject might be having a bad day or a bad year. To deal with this concern, the Minnesota Twin Study waited about a decade, then retested. As expected, the research team found that happiness is fairly stable over time; humans have a “happiness set point” to which they gravitate. The researchers were amazed, however, to discover that people are as similar to their identical twin a decade ago as they are to themselves a decade ago. The implication: “Nearly 100 percent of the variation in the happiness set point seems to be due to individual differences in genetic makeup.”