Sibling deidentification
A Socratic walk-through of sibling deidentification — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does one sibling's success at something so often make the other quietly drop it?
An older brother wins a regional swimming title. Within a year the younger one, who had been swimming happily and not badly, has stopped — and if you ask, he will say he simply lost interest. Families narrate this everywhere: one is the musical one, one the sporty one, one the clever one, and the labels are treated as discoveries about the children.
The tempting reading is that the second child was not as good. Worth testing, because a rival reading sits beside it: that the abandonment answered the comparison rather than the ability, and the difference the family later reports as a discovery was partly manufactured by the family's own attention.
Reasoning it through
REASONING #Ask first what a child is optimising, if anything is. Not swimming times. In a household a child's standing is judged before an audience of two or three adults, and the comparison group is not the population — it is the sibling at the same dinner table. A hundredth-percentile swimmer with a ninety-ninth-percentile brother is, in the only forum that matters daily, the slower one.
Follow that. A domain visibly held by a sibling has three properties for the second child. Every performance invites an unfavourable comparison. Any success reads as derivative — following the brother, not doing your own thing. And the parental attention in it is already committed: the lift to the pool, the interest, the story the family tells about who they are, all point elsewhere. An unoccupied domain has none of those and buys distinct identity far more cheaply.
So a rule falls out that is not about talent: pick the domain your sibling has not claimed. Do that from both sides for a decade and you produce a family of specialists, all of whom then explain their specialism as innate.
Two mechanisms could drive that rule, and they predict differently. One is self-evaluation: the drop protects a comparison the child cannot win, so it should be strongest for close-in-age, same-sex pairs in domains central to the abandoner's sense of themself. The other is allocation: the drop responds to scarce parental time, so it should track how much resource that domain consumes and weaken where a household can fund two swimmers without contest. The evidence does not pick decisively between them.
Now the reading that must be ruled out before either can be believed, and it is embarrassingly simple. Children differ in ability. The one who is worse at swimming quits swimming. That produces the identical family pattern, requires no comparison mechanism, and is more parsimonious. Add regression to the mean — an early performance that looked like talent was partly noise, and the next season is duller — and the abandonment needs no explanation at all.
Worse, most of the evidence is retrospective. "She was always the musical one" is a sentence produced by a family that has already sorted itself; it is not an independent measurement of anything.
What would separate the readings? Measure interest and ability in a domain before the sibling's visible success, then track who abandons it. If deidentification is real, abandonment should be predicted by the sibling's success holding the abandoner's own trajectory constant, and should be stronger for close-in-age, same-sex pairs. The sharpest test compares siblings with cousins of similar ability raised apart: cousins face the same ability distribution and almost none of the shared audience. If cousins showed the same withdrawal, comparison is not doing the work.
The analogy
THE ANALOGY #Think of two shops opening on one small high street. The second owner does not choose what to sell by consulting their own tastes; they walk the street first and notice the bakery. Selling bread would mean competing on someone else's ground for customers who already have a habit — so they open a hardware shop instead, and years later everyone says the street sorted itself naturally by what each owner was suited to.
A shopkeeper knows they are avoiding competition and could enter anyway, whereas a child experiences the withdrawal as a genuine change in what interests them; and a high street has far more slots than a family has domains, so the binding constraint on siblings is parental attention rather than the number of niches.
Clarifying the model
THE MODEL #Two clarifications, and one connection.
First, deidentification is not a claim that the siblings had equal ability. It is a claim about an increment — that the sibling's visible success subtracts from the other's persistence over and above whatever the ability difference already does. That increment is the entire content of the hypothesis and what the designs above try to measure.
Second, nothing here licenses advice. No finding says a family should conceal one child's success or manufacture separate domains, and the pattern described is a population tendency assembled from wildly varying cases. Plenty of sibling pairs share a domain for decades without either withdrawing.
The connection is to a question this collection has already walked through — why siblings raised in one household turn out so unalike. The answer there was that the environments are not in fact shared: the influences shaping personality appear individual to each child, and that component remains mostly unexplained. Deidentification is a candidate filler for part of it, and an unusual one, because it would generate difference out of a shared exposure. The fixed point of difference is that the earlier account measures how big the unexplained component is, while this one proposes a process that might live inside it — and proposing is not demonstrating.
Hence the honest summary: the pattern is described repeatedly and the mechanism is not settled. I quote no effect sizes, because published estimates rest on small samples and differing definitions of what counts as a domain.
A picture of it
THE PICTURE #How to readStart at the rounded terminal, with a child who already likes something; the parallelogram is the input the ability account has no use for, a look at where the sibling stands. Both diamonds are genuine branches with both exits labelled — an unclaimed domain, or a claimed one the child thinks they can still win, leads to investment, while a claimed one they cannot leads to withdrawal. The cylinder is the family's running account of who is what, which both outcomes write into, and the dashed back-edge is the loop that matters: that account is what the next child reads when they try something. The red node is the rival explanation drawn in deliberately, since a plain ability gap reaches the same withdrawal by a shorter route.
What became clearer
WHAT CLEARED #The family's own explanation — that each child turned out suited to something different — may have the causation backwards. If a child's standing is judged against the sibling in front of them rather than against the world, an occupied domain is a bad bet regardless of talent, and withdrawing is the cheapest route to a distinct identity. Repeat that for a decade and the household produces specialists, then reads the specialisation as a discovery.
But the same pattern falls out of children simply differing in ability, and most of the evidence is retrospective narrative from families that have already sorted themselves. So what is gained is not a conclusion but a sharpened question: does a sibling's visible success subtract from the other's persistence over and above the ability gap, and does the subtraction shrink between cousins? Until a study measures the domain before the success rather than after, both answers stay on the table.
Where to go next
ONWARD #- Whether the same logic operates between children in a small classroom, where the comparison audience is also fixed and local.
Key terms
TERMS #| Term | What it means |
|---|---|
| Deidentification | withdrawal from a domain a sibling has claimed, as active differentiation from them. |
| Niche picking | selecting an activity partly for its vacancy rather than for one's aptitude at it. |
| Regression to the mean | the tendency for an unusually high measurement to be followed by an ordinary one, which mimics a real decline. |
Every term the collection defines is gathered in the glossary.