Wikipedia Deep Dive

Twin study

16 min read

The King Who Sentenced Twins to Coffee and Tea

In 1771, King Gustav III of Sweden had a problem. He was absolutely convinced that coffee and tea were killing his subjects, and he wanted to prove it. So he did what any scientifically-minded monarch would do: he found a pair of identical twin murderers on death row and offered them a deal.

Live out your days in prison, but one of you must drink three pots of coffee every single day, and the other must drink three pots of tea. The king would assign doctors to watch and wait for the inevitable health catastrophe.

It didn't quite work out as planned.

The tea-drinking twin died first—at the ripe age of 83. Both of the doctors assigned to monitor the experiment had already died. King Gustav himself had been assassinated in 1792, decades before either twin passed. As for the coffee drinker? We don't even know when he died, only that he outlived everyone involved in the study.

This bizarre royal experiment, conducted nearly 250 years ago, was perhaps the first twin study in medical history. And while Gustav's hypothesis about the dangers of caffeine didn't pan out, he stumbled onto something far more valuable: a method for separating nature from nurture that scientists still use today.

The Perfect Natural Experiment

Here's the fundamental question that twin studies try to answer: When you look at any human trait—height, intelligence, personality, susceptibility to depression—how much of it comes from your genes, and how much from your environment?

This sounds like it should be impossible to untangle. After all, you can't run an experiment where you give one person different genes and see what happens. You can't rewind someone's life and raise them in a different family.

But nature has provided us with something almost as good.

Identical twins—scientists call them monozygotic twins, from "mono" meaning one and "zygote" meaning fertilized egg—develop from a single egg that splits early in pregnancy. They share essentially 100 percent of their DNA. They're genetic clones of each other.

Fraternal twins, by contrast, develop from two separate eggs fertilized by two separate sperm. They're called dizygotic twins ("di" meaning two). Genetically, they're just ordinary siblings who happened to share a womb—they share about 50 percent of their genes, like any brother or sister pair.

Here's where it gets clever. Both types of twins typically grow up in the same family. They share the same parents, the same house, the same socioeconomic status, the same culture, often even the same friends. In scientific terms, they share their "environment."

So if you compare how similar identical twins are on some trait versus how similar fraternal twins are, any extra similarity in the identical twins must come from their extra genetic similarity. That's the core logic of twin studies, and it's remarkably powerful.

What Twin Studies Have Actually Found

The results have been striking, sometimes controversial, and occasionally uncomfortable.

Height is one of the clearest cases. Identical twins are almost exactly the same height—far more similar than fraternal twins. The math works out to height being roughly 80-90 percent heritable. This doesn't mean 80 percent of your height comes from genes in some absolute sense; it means that 80 percent of the variation in height across a population can be attributed to genetic differences between people.

The distinction matters. If everyone ate the exact same diet and lived in identical conditions, height might be 100 percent heritable—because all the remaining differences would be genetic. If there were massive differences in nutrition and healthcare, heritability might drop, because environment would explain more of the variation.

Personality traits show intermediate heritability, typically around 40-60 percent. Whether you're introverted or extraverted, agreeable or cantankerous, conscientious or scattered—these aren't purely products of how you were raised. Identical twins raised apart often have eerily similar personalities.

Mental health conditions show substantial genetic influence too. Schizophrenia is perhaps the most dramatic example: if one identical twin develops schizophrenia, the other has about a 50 percent chance of developing it too. For fraternal twins, that drops to around 15 percent. Depression, anxiety, bipolar disorder, autism—all show significant heritability.

But here's what makes this complicated: even with a trait that's 80 percent heritable, environment still matters enormously. That remaining 20 percent represents real differences, real opportunities for intervention. And heritability isn't destiny—it's a population statistic, not a prediction for any individual.

The Missing Heritability Mystery

For decades, twin studies consistently found that most traits were at least 50 percent heritable, sometimes much more. Then something strange happened.

As genetic technology advanced, scientists began trying to find the actual genes responsible for these heritable traits. They would sequence thousands of people's genomes, identify genetic variants associated with, say, height or intelligence, and add up all their effects.

The numbers didn't match.

Twin studies said height was 80 percent heritable. Molecular studies could only find genes accounting for maybe 10-20 percent. Where was the rest? Scientists called this gap "missing heritability," and it launched a fierce debate that continues today.

Some argued that twin studies were overestimating heritability—perhaps identical twins share more similar environments than fraternal twins, inflating the apparent genetic effect. Others suggested that thousands of genes with tiny individual effects were lurking below the detection threshold of early molecular studies. Still others proposed that gene-gene interactions, epigenetics, or rare genetic variants might explain the gap.

The mystery hasn't been fully solved, but larger genetic studies have gradually closed the gap for some traits. The genes were there; we just needed bigger sample sizes to find them.

A Brief History of Looking at Twins

Humans have been fascinated by twins for as long as we've been recording our thoughts.

Hippocrates, the ancient Greek physician who lived in the fifth century BCE, noticed that twins sometimes developed different diseases. He attributed this to differences in their material circumstances—essentially an early environmental hypothesis.

A century later, the Stoic philosopher Posidonius took the opposite view. When he saw similarities between twins, he attributed them to shared astrological influences. Same birthday, same stars, same fate.

Saint Augustine, writing in the fourth century CE, used twins as an argument against astrology. If the stars determined destiny, why did twins so often live such different lives?

The scientific study of twins began in earnest with Sir Francis Galton in the late 1800s. Galton was Charles Darwin's cousin and shared his fascination with heredity. He collected data on twins, noting their similarities and differences, trying to understand how much human traits were inherited.

But Galton didn't know about the difference between identical and fraternal twins. Neither did Edward Thorndike, who conducted the first psychological twin study in 1905, testing fifty pairs of twins on mental abilities. This seems like an obvious oversight now, but the biological distinction wasn't clearly understood yet.

The breakthrough came in 1924, when German geneticist Hermann Werner Siemens developed a method for distinguishing identical from fraternal twins based on multiple physical similarities. Around the same time, Wilhelm Weinberg and colleagues figured out how to calculate the rates of each type from maternity records. The modern twin study was born.

The Equal Environments Assumption

Twin studies rest on a crucial assumption that critics have questioned for decades: that identical and fraternal twins share their environments equally.

Think about it. If parents treat identical twins more similarly—dress them alike, send them to the same activities, expect them to be similar—then any extra similarity in identical twins might not be genetic at all. It might just reflect their more similar treatment.

Researchers have tested this in various ways. One approach: find identical twins who were mistakenly believed to be fraternal (it happens), or fraternal twins mistakenly believed to be identical. If the equal environments assumption is wrong, the perceived type should matter more than the actual type. Generally, it doesn't.

Another approach: directly measure how similarly twins are treated and see if it predicts their similarity on psychological traits. The relationship is usually weak or absent.

The assumption isn't perfect, and it remains the most common criticism of twin studies. But decades of research suggest it's approximately true for most traits—similar enough that twin studies remain useful, even if the precise heritability estimates should be taken with some uncertainty.

The ACE Model: Genes, Shared Environment, and Everything Else

Modern twin researchers don't just report a single heritability number. They try to partition all the variation in a trait into three buckets, known by the acronym ACE.

A stands for additive genetic effects. This is the straightforward inheritance from your parents—the genes that add up to influence a trait. It's what we usually mean by heritability.

C stands for common or shared environment. These are the environmental factors that make twins similar: growing up in the same family, same neighborhood, same school, same socioeconomic circumstances. Shared environment explains why children from the same family resemble each other beyond just genetics.

E stands for unique or nonshared environment. This is everything that makes twins different despite their shared genes and shared family: different friends, different experiences, random events, even different positions in the womb. Identical twins aren't perfectly identical, and the E component captures why.

Here's the clever math. If you measure the correlation between identical twins on some trait—how similar they are—that gives you A plus C. They share all their genes and all their shared environment.

The correlation between fraternal twins gives you half of A (since they share half their genes) plus C.

The difference between these two correlations is due entirely to the difference in genetic sharing. So you can double that difference to estimate A. Then you can work backward to estimate C. And since identical twins still aren't perfectly correlated, whatever is left over is E.

This is called Falconer's formula, and it's the foundation of quantitative genetics. Modern researchers use much more sophisticated statistical techniques—structural equation modeling that can handle complex family structures, missing data, and multiple traits at once—but the core logic is the same.

The Surprising Power of Unique Environment

One of the most consistent and surprising findings from twin research: shared family environment matters less than almost everyone expected.

For many psychological traits, the C component is close to zero by adulthood. The family environment—all those things parents agonize over, from parenting style to neighborhood to enrichment activities—explains very little of why children turn out different from each other.

What does matter is the unique environment: the experiences that differ between siblings in the same family. But here's the puzzle—we don't really know what these experiences are. They include measurement error (the tests aren't perfect), but that's not the whole story. Random developmental noise? Peer groups? Specific teachers or mentors? Minor illnesses or injuries? We're not sure.

This finding, popularized by psychologist Judith Rich Harris in her controversial book "The Nurture Assumption," remains hotly debated. Critics argue that the statistical models might be missing something, that shared environment might operate in subtle ways that twin studies can't detect. But the basic finding has replicated across dozens of traits and hundreds of studies.

It doesn't mean parenting doesn't matter. Extreme environments—abuse, neglect, severe deprivation—clearly have lasting effects. And parenting might matter for things that twin studies don't typically measure, like the specific content of your beliefs or your relationship with your parents. But for broad personality traits and cognitive abilities, the shared family environment seems to matter less than our intuitions suggest.

Twins Raised Apart: The Ultimate Test

If you really want to separate nature from nurture, the gold standard is twins raised apart: identical twins separated at birth and raised in different families.

These cases are rare—adoption practices usually try to keep twins together—but they're scientifically precious. The Minnesota Study of Twins Reared Apart, which began in 1979, tracked down dozens of such pairs and brought them to the university for extensive testing.

The results made headlines. Identical twins raised apart were remarkably similar—almost as similar as identical twins raised together. Some of the individual cases became famous: the "Jim twins," both named Jim by their adoptive families, both married women named Linda, both divorced and remarried women named Betty, both named their sons James Allan, both had dogs named Toy, both chain-smoked Salem cigarettes, both worked as deputy sheriffs.

These anecdotes should be taken with caution—with enough pairs, you'll find coincidences, and the most dramatic stories get selected for publicity. But the systematic data told a similar story. Personality, intelligence, even vocational interests and social attitudes showed strong correlations between twins raised in completely different families.

Critics raised concerns. Maybe the adoptive families weren't that different—adoption agencies might have matched twins to similar families. Maybe the twins had been in contact before the study. Maybe meeting your identical twin as an adult creates a psychological pull toward similarity.

These are legitimate concerns, and twins-reared-apart studies have limitations. But combined with the standard twin studies, they strengthen the case for substantial genetic influence on a wide range of traits.

The NASA Twins Study: Nature's Spaceflight Experiment

In 2015, NASA conducted perhaps the most unusual twin study ever.

Astronaut Scott Kelly spent 340 days aboard the International Space Station while his identical twin brother Mark—also an astronaut—remained on Earth. Scientists compared every imaginable biological marker between them, tracking how spaceflight affected Scott's body and genes while using Mark as a genetically identical control.

The findings were fascinating. Scott's telomeres—the protective caps on the ends of chromosomes, often used as a marker of biological aging—actually lengthened in space, though they shortened again after he returned. His gene expression changed in measurable ways, with some changes persisting for months after landing. His gut microbiome shifted. His cognitive performance showed subtle declines that took time to recover.

The study wasn't really about nature versus nurture in the traditional sense—it was about the unique environment of space. But it showcased the power of the twin design for isolating environmental effects. With identical genetic backgrounds, any differences between the brothers could be attributed to their radically different environments.

Beyond Simple Heritability

Modern twin research has moved far beyond asking "how heritable is this trait?" Scientists now use twin data to answer more sophisticated questions.

Do different traits share genetic influences? By studying correlations between traits within twin pairs, researchers can estimate the degree of genetic overlap. The finding that IQ and educational achievement share much of their genetic basis—but not all of it—tells us something about how intelligence and motivation separately contribute to school success.

How do genetic influences change over time? Longitudinal twin studies track pairs across years or decades. One consistent finding: genetic influences on cognitive ability increase with age, while shared environmental influences decrease. By adulthood, the family you grew up in matters much less than it did in childhood.

Can we test causal hypotheses? If depression causes people to exercise less, we'd expect the twin who's more depressed to exercise less—even among identical twins. If the correlation between depression and exercise is entirely genetic (some genes increase both depression risk and exercise avoidance), we'd see no within-pair relationship. These "discordant twin designs" offer a powerful way to test causal claims that are otherwise confounded by genetics.

Interestingly, the depression-exercise relationship appears largely genetic. When researchers look within identical twin pairs, the more depressed twin doesn't consistently exercise less. This doesn't mean exercise therapy is useless for depression—interventions can work even when the natural correlation is genetic—but it complicates the simple story that sedentary lifestyle causes depression.

The Limits of Twin Studies

Twin studies are powerful, but they have real limitations that should temper how we interpret their results.

Twins might not be representative of the general population. The prenatal environment differs—twins compete for resources in the womb, and identical twins sometimes share a placenta (which can create unique complications). Being a twin might itself affect development, through the social experience of having a constant companion.

The equal environments assumption, while approximately true, isn't perfect. For some traits—particularly those where similar appearance might lead to similar treatment—heritability estimates might be inflated.

Gene-environment correlation complicates interpretation. Smart kids seek out intellectually stimulating environments; sociable kids attract more social interaction. When genes influence the environments people select, the neat separation of A, C, and E becomes blurry.

Heritability is population-specific. If a population has little environmental variation, heritability will be high—because genes explain most of what variation remains. In a population with massive environmental inequality, heritability might be lower. Saying a trait is "80 percent heritable" tells you about a particular population at a particular time, not some universal truth about human nature.

And crucially, heritability says nothing about malleability. Height is highly heritable, yet average height has increased by several inches over the past century due to improved nutrition. A trait can be strongly influenced by genes and still be changed by environment.

Why This Research Matters

Twin studies have shaped our understanding of human nature in profound ways. They've shown that genes influence not just physical characteristics but personality, mental health, cognitive abilities, and even political attitudes. They've revealed that the family environment, so central to how we think about child development, explains less than expected for many traits.

This research challenges both extreme views. The blank slate hypothesis—that humans are infinitely malleable, shaped entirely by environment—is clearly wrong. But so is simple genetic determinism. Genes are powerful, but they operate probabilistically, through complex interactions with environment, and always leave room for individual variation.

For the ongoing debate about missing heritability mentioned in the context, twin studies provide one of our best tools for establishing how much genetic variation should exist—the target that molecular genetic studies are trying to hit. Whether the gap reflects limitations of molecular methods or overestimation by twin studies remains an active scientific question.

Meanwhile, twin registries around the world continue collecting data, following participants across decades, combining classical twin methods with modern genomic techniques. The twins themselves—those natural experiments, those variations on a genetic theme—remain our most valuable window into the interplay of nature and nurture that makes each of us who we are.

King Gustav III would probably be pleased to know that twin research is still going strong, even if his coffee hypothesis didn't work out. The Swedish Twin Registry he indirectly inspired now contains data on over 85,000 twin pairs. His condemned murderers started something bigger than either of them—or he—could have imagined.