Sibling differences

Why can siblings raised in the same household develop very different personalities?

Same parents, same house, same rules, same table, same schools — and by adolescence one sibling is cautious and one reckless, one sociable and one guarded. The usual explanations reach for birth order, or for parents being subtly different the second time round.

But look at the assumption underneath the puzzle first: that "raised in the same household" means "raised in the same environment". Does it? And a second hides beside it — that siblings have, in some loose sense, the same genes.

Take the genetic one first, since it is easier to dispose of. Full siblings share on average about half of the genetic variants that differ between people — and on average is doing real work there. It is the expected value of a lottery run afresh at each conception, so two siblings can come out considerably more or less alike than half. So "same genes, same house, different outcome" was never true at the first step.

Now the harder half. How would anyone measure what a shared household does? Twin and adoption designs give a way in: compare pairs who differ in how much genetic material they share and in whether they were reared together, and the differences among people in a trait can be apportioned into three buckets — genetic influence, shared environment (everything making people in the same family alike), and everything else, called nonshared environment, which also absorbs measurement error.

Here is the finding that surprises nearly everyone. For personality, the shared-environment bucket comes out small — often close to zero by adulthood. Unrelated children adopted into one home, raised on the same rules and the same table, end up not much more similar in personality than two strangers. Whatever a household does, making its children's personalities alike is not conspicuously among it.

So what is in the large remaining bucket? Some is plainly what one sibling had and the other did not: a friend group, one teacher, an illness, a first job, the order in which things happened. Some is the same household experienced differently — a parent at twenty-five is not that parent at thirty-two, and a divorce at four is not a divorce at eleven. Some is siblings actively pushing apart: if one becomes the sporty one, the role is taken. And a subtle one: a child's own disposition draws different treatment from the very same parent, so a placid infant and a colicky one are not living with the same mother. And some of it, honestly, is chance — developmental noise too small and too various to record.

The intuition being corrected is that a shared upbringing ought to make siblings alike, so any difference demands a special cause — a birth-order effect, a favoured child. Turn it round: the parts of an environment that shape personality are largely the parts each child meets alone, so similarity is the thing that would need explaining.

And the numbers are easy to abuse. They are variance components within a studied population, not statements about an individual: "personality is around forty per cent heritable" does not mean forty per cent of your character came from your genes. Estimates move with the trait, the age, the measure, and the range of environments sampled — where those run from severe deprivation to comfort, the shared-environment share is larger, because there is more variation to capture. And the nonshared bucket is mostly unexplained: decades of searching have not produced a tidy list of nonshared causes, and part of it is simply measurement error. It is a residual, not a discovery.

• Why shared environment looks influential in childhood for some traits and fades by adulthood.
• How a disposition builds the environment that then reinforces it — gene-environment correlation.
• Why identical twins raised together still differ, and what that residual can and cannot tell us.

Every term the collection defines is gathered in the glossary.