Where our numbers come from

The backbone is the United States federal record, because it is the deepest public naming dataset in the world: 146 years, 117,820 distinct names, national and state level.

• US Social Security Administration — National baby names
  Every name given to five or more babies of one sex in a year, 1880–2025
• US Social Security Administration — Baby names by state
  The same counts split across all 50 states and DC, 1880–2025
• US Social Security Administration — Actuarial life table (Table 4.C.6)
  Survival probability by age and sex — the basis for every “alive today” estimate

For the “around the world” figures on each name page we use the current official rankings published by 4 more agencies:

• Canada — Statistics Canada, 2024 (published top 101 per sex)
• England & Wales — Office for National Statistics, 2024 (published top 100 per sex)
• France — INSEE, 2025 (published top 100 per sex)
• Scotland — National Records of Scotland, 2025 (published top 103 per sex)

Notable people named X come from Wikidata, a public structured database maintained by the Wikimedia community. That is the one source on this list that is community-edited, and we label it as such wherever it appears.

Syllable counts come from the CMU Pronouncing Dictionary (cmudict 0.7b, Copyright © 1993–2015 Carnegie Mellon University, redistributed under the BSD-2-Clause licence). It carries about two spellings in five, and those cover roughly four in five of the births recorded that year. For a name it does not carry we count vowel groups in the spelling instead, scored at about 92% against the names it does carry, and the page prints those as “about”.

This distinction matters more than any single number, so we keep it explicit everywhere on the site.

Measured — taken directly from the agency files, with no modeling by us: every rank, every yearly birth count, peak year, first year on record, state-level counts and every foreign ranking.

Estimated — calculated by us from those measured counts. Estimates are always phrased as estimates (“about”, “roughly”, “expect”) and never presented as a record. The four models we run are below.

People alive today with a name

We take every recorded birth for the name, one year at a time, and multiply it by the probability that someone born in that year is still alive, using the Social Security Administration's published actuarial life table for that sex. Adding those products gives the estimate:

living ≈ Σ (births in year Y × survival probability at age [current year − Y])

Two honest limitations: the table is a national average, so it knows nothing about a specific family, and the SSA data counts Social Security card applications rather than every birth, so names given to people who never applied are under-counted.

How many will share the name in a classroom

We compare the name's most recent birth count with the number of babies of that sex in the same year's records — about 1.6 million girls and 1.7 million boys in 2025 — and scale it to a class of 25 children:

per class = (births with the name ÷ cohort size) × 25

It answers the real question parents ask — “will there be three of them in her class?” — but it assumes a classroom that mirrors the country. Regional taste is strong in naming, so check the state pages too.

The same name, spelled other ways

The SSA counts spellings, not names, so Jackson and Jaxson are two entries in the file and one name in a classroom. We group the spellings of a name with a fixed rule that folds interchangeable English spellings of one sound — ck for k, ph for f, x for ks, a non-initial y for i, doubled letters — while keeping every vowel apart, then add up the births.

The rule errs towards splitting. Sean and Shawn stay separate here even though they sound alike, because a family that is too small only understates a number, while a family that is too large tells you your child will share a name with children who will not. Phonetic hashes such as Double Metaphone do the opposite — they merge Chloe with Kelly and Rhett with Rudy — which is why we do not use one. Every name page prints the whole family it used, so you can judge the grouping instead of trusting it.

The typical age of someone with the name

We find the weighted median birth year across everyone ever given the name: the year by which half of all people carrying it had been born. Subtracting it from the current year gives the typical age. A name with two separate popularity waves has a median sitting between them, which is why we call it typical rather than average.

Fashion outlook

A transparent scoring heuristic, not a prophecy. It reads three measured facts — how many separate decades the name held a top-100 place, where it sits relative to its own peak, and how fast it is currently moving — and returns how likely the name is to read as “of its era” about 15 years from now. A name that charted across six or more decades scores lowest, because it has already survived several fashion cycles. We show the reasoning in a sentence next to the score; no part of it is a black box.

One rule, applied to every name identically. We compare the current rank with the rank five years earlier. A move of more than 15% counts as rising or falling; smaller moves are steady. A rising name whose peak was more than twenty years ago is labeled a comeback rather than a riser. Names outside the top 2,000 are called rare, because ranks that deep move for statistical reasons rather than real changes in taste.

The SSA publishes new US figures each spring; the other agencies publish on their own schedules through the year. We reload every source when it appears, recompute all derived statistics, and the current dataset was last refreshed on 2026-05-01. Historical years never change retroactively, so a chart you read today will look the same next year, with one more point on the end.

Everything we derive from the source counts — the rankings as we present them, spelling families, living-bearer estimates, typical ages and fashion outlooks — is published under the Creative Commons Attribution 4.0 licence (CC BY 4.0). Reuse it freely, with a link to this page (https://babynamebattles.com/data) as attribution.

The source data is not ours to license. It stays under the terms of the agency that publishes it: the SSA files are a work of the US federal government, and each of the other agencies above sets its own conditions for reuse.

If a number here disagrees with the agency that published it, that is a bug and we want to know. Write to support@babynamebattles.com and we will fix it or explain the difference.

See also: about this site · privacy · terms