Athletes with march birthdays from major sports

Ask a large population how old everyone is and the answers pile up on round numbers. Not because people cluster at those ages, but because that is how people answer when they are estimating. The pattern is old, well documented and measurable, and it explains why some recorded ages are softer than others.

What to take away

• Where a birth was never registered, an age is a report rather than a reading, and reports round.
• The rounding leaves a signature in the aggregate that statisticians can measure, which is how the quality of an age dataset gets assessed.
• A single entry cannot be corrected by knowing this, but it can be weighted properly.

Where a number comes from when there is no document

Civil registration is the system that records births as they happen, and it is not universal in time or place. Coverage has been built up over generations, interrupted by conflict and displacement, and reconstructed after losses. Where a birth fell outside it, an age is established later: from a household's memory, from a relation to a remembered event, from a school enrollment, from a document issued long afterward on the basis of one of those.

Every one of those routes produces an estimate. Estimates round, and they round to particular numbers: multiples of ten first, then multiples of five. This is measurable in the aggregate, and demographers have a standard tool for it. The standard index of digit preference compares how many reported ages end in the preferred digits against how many would if reporting were even, and turns the result into a score for how much of it a dataset carries.

Why this reaches a sports page

Athletes come from everywhere, and a squad list mixes people whose ages rest on very different foundations. A census or a registration system in one country may be old and complete; in another it may have been rebuilt within living memory. Both produce a document. The documents look the same on a team sheet.

The basis for the age	What it is evidence of
A birth registered at the time	The date, to the standard of that system
A later registration from family testimony	What a family reported, at the time they reported it
A document issued for a specific purpose	That the issuing body was satisfied then
An age given at first registration with a club	A number someone supplied
A reconstruction after records were lost	A careful estimate, or a rounded one

None of these are dishonest. They are different evidentiary situations wearing the same clothes, and a list that shows only the resulting number erases the difference entirely.

Reading a disputed age fairly

The default assumption when two documents disagree is that someone lied. That assumption is wrong more often than it is right, and it lands hardest on people from places where registration came late through no fault of theirs. The January athletes page sets out the three different age questions that get confused, and the February athletes page covers what a record claim actually rests on. Both apply here, with one addition: before treating a disagreement as a discrepancy, ask what each document was evidence of.

What a page should do

• Never state a date, an age or a place of birth. This site does not, for the reasons the January index gives.
• Where a governing body has recorded a dispute, note that the dispute exists and who recorded it, without adopting either value.
• Do not describe a whole country's records as unreliable. Registration coverage varies by period and region within countries as much as between them.
• Treat a listed age as a claim with a provenance, not a measurement. The March index makes the same point about calendar dates.

Common questions

Does digit preference mean an individual's age is wrong?

No. It is a property of a large collection of reported ages, not a test that can be applied to one person. Knowing a dataset rounds tells you how much to trust the aggregate, not which entries moved.

Why not just use the earliest document?

Because the earliest document is not automatically the best one. A record made contemporaneously beats a later one; a later record made from good evidence beats an early one made from a guess. Age of the document is a proxy, not the thing itself.

Is this a problem that has been solved?

Registration coverage has improved a great deal over time, so it is much smaller for people whose births were recorded recently and unchanged for people whose were not. It will be with sport for as long as careers include people born under older systems.

What should a reader conclude from an athlete's listed age?

That someone recorded a number and a governing body accepted it. That is genuinely worth knowing and it is less than the number appears to say.