How SSA name data works
Almost every U.S. baby-name statistic you will read, here or elsewhere, comes from one source: the Social Security Administration’s baby-name files. In 2025 they recorded 3,315,343 births under 31,219 names. Knowing what those files hold, and what they leave out, is the difference between reading a name’s numbers and misreading them.
Where the numbers come from
When parents apply for a baby’s Social Security card, the application records the first name, the sex and the state of birth. Each spring SSA counts those applications and publishes two files: a national file with every first name given to at least 5 babies of one sex in each year since 1880, and a state file with the same for each state and the District of Columbia since 1910. SSA published the 2025 files on May 8, 2026. Nearly every baby now gets a card at birth, so recent counts cover almost every birth; early years do not, as the next sections explain.
SSA national and state files, published May 8, 2026: SSA’s notes on the national file.
What the files leave out
A name given to fewer than 5 babies of a sex in a year is not in the files at all, to protect privacy. That hides most rare names and many unusual spellings, and it makes a rare name’s history patchy: a name with 6 births one year and 4 the next shows a gap, not a fall to 4. On NameAlmanac those years are drawn as gaps, never as zeros.
Names are counted exactly as written, so Sofia and Sophia, or Aiden and Aidan, are separate names with separate ranks. Names run from 2 to 15 letters, so longer names are cut short. And the files include records that are not names at all, such as Baby, Unknown or Infant, entered when a card was issued before a name was chosen. NameAlmanac counts those in totals but never lists or ranks them.
Thresholds and truncation: SSA’s readme for the national file (NationalReadMe.pdf, in the files SSA published on May 8, 2026).
Why the early years are thin
Social Security numbers were first issued in late 1936. People born before then appear in the files only if they applied for a number later in life, and many never did, especially those who died young. So the counts for the 1880s and 1890s are far below the real number of births: SSA recorded 201,484 births for 1880. Compare shares within a year rather than raw counts across the early decades, and treat the living estimates for those years with care; the methodology explains how NameAlmanac adjusts for it.
Births recorded per year: SSA national file, 1880-2025.
How to read a name’s numbers
Three figures do most of the work. Births are the count of babies given the name that year. Rank is the name’s place among names of the same sex that year, with ties broken alphabetically, as SSA does. Share is the name’s births per 1,000 babies of that sex: it corrects for the number of births, which rose from about 2.3 million recorded in 1920 to 4.2 million in 1957 and is 3.3 million now. A name with the same births in 1957 and 2025 was relatively rarer in 1957.
Rank can mislead on its own. American naming has spread out: in 1950 the ten most given girls’ names went to 24.1% of girls born, and in 2025 to 7.0%. So the #1 name of 2025, Olivia with 13,544 girls, was given to fewer babies than the #10 name of many years in the 1950s. When a page shows a rank, it also shows the births and the share.
Top-10 shares: How American naming spread out, from the SSA national file.
What the files cannot tell you
The files record the sex marked on the card application, not anyone’s identity today. They do not record ethnicity, the parents’ reasons, a family’s pronunciation or middle names. They count only people born in the United States who got a Social Security number, so anyone born abroad is missing. And they say nothing about meaning: NameAlmanac quotes meanings from Wiktionary, with credit, and says so when there is none.
Meanings: Wiktionary (CC BY‑SA 4.0), kaikki.org extract of Sep 25, 2026.