Analysis
How SSA baby-name data differs from NCHS state birth records
For celebrity families, children and baby names, SSA state lists and NCHS birth files answer different questions about what names are rising.
What to take away
- The SSA baby-name database counts names written on Social Security card applications. NCHS state birth files count birth certificates filed in each state or territory.
- SSA hides every name given to fewer than five babies in a year and sex, so its state lists understate variety rather than overstate it.
- Federal NCHS public-use files normally drop the child's name, so name-level counts come from a state or city office.
- A rank can climb while a name's share of births barely moves. The state birth count separates a real shift from noise.
- Access differs by state. One office posts county counts, another posts an annual name list, a third takes a data request.
What each dataset counts
They are not two versions of one thing. The unit each dataset counts decides what a sentence about a name can prove.
SSA baby-name database
- Unit counted
- Social Security card applications
- Smallest published unit
- Five babies per year per sex
- Coverage start
- National 1880, state lists 1910
- Geography
- Nation and state
- Name in federal release
- Yes, above the threshold
- Question it answers
- Order and frequency
NCHS state birth files
- Unit counted
- Birth certificates filed
- Smallest published unit
- Varies; public files may suppress
- Coverage start
- Federal collection, current years
- Geography
- State, county and city
- Name in federal release
- Normally no
- Question it answers
- Scale of births
Read the rows as a division of labor. SSA tells you order. NCHS tells you scale.
[figure 1]
SSA vs NCHS: What Each Counts
SSA state name list
- Unit counted
- Card applications
- Coverage
- 5+ uses in state-year
- Geography
- State, some sub-state
- Timing
- Birth year, application lag
- Best use
- Ranking names in state
NCHS state birth file
- Unit counted
- Birth certificates
- Coverage
- All filed births
- Geography
- State, county, city
- Timing
- Occurrence, certificate date
- Best use
- Name share of births
A famous family's announcement is a claim about a certificate almost nobody can see. celebrity baby names that matter explains why the published name and the data rarely line up, which is the right starting point before any figure is quoted.
How the SSA files are built and where they stop
The Social Security Administration builds its lists from card applications for people born in the United States. Its background information for popular names states the five-baby rule: names given to fewer than five babies in a year and sex are left out.
That threshold is the first line to read on any state list. Wyoming, Vermont and Alaska have far smaller birth populations than California, so their lists are shorter and jumpier. A high rank in a small state can rest on a handful of applications.
Two limits follow. The national files begin in 1880 and the state files in 1910. Applications also lag, and the files are organized by birth year rather than by the date a card was issued.
What state birth files add, and where to get one
The National Vital Statistics System gathers certificates from the 50 states, the District of Columbia, New York City, Puerto Rico and the island territories. The NVSS birth data page is the federal entry point.
Each certificate carries the child's name as filed, both parents' names, and the date and place of birth. Federal public-use files usually remove the name, so a name-level question goes to a state or city office.
Florida publishes county birth counts through FLHealthCHARTS, an online query system. It reports births, not name frequencies, so name-level counts come through a data request to the state health department.
New York City's health department publishes an annual list of top baby names. That is a city file, and it is counted apart from New York State.
The SSA popular names by state tool shows top names in a state for a chosen year, and it still applies the five-baby floor.
Those releases reach the press through a chain of filings. How CDC vital records shape a US celebrity birth announcement follows that chain, which explains why the announcement and the file often disagree.
Example: testing a claim that a name is exploding in Illinois
Take an illustrative case. The counts below are placeholders that show the method.
A name sits near rank 60 in the SSA Illinois list five years ago and near rank 40 in the latest file. Illinois recorded roughly 140,000 births in each year. The name's count moved from about 90 to about 140.
Its share of births went from about 0.06 percent to about 0.10 percent. The rank rise is real. The word "exploding" is not supported, because rank responds to the pool around it.
Rank is a statement about the pool. Share is a statement about the state.
A second case: a name common in California and Texas but absent from the national top 100. The national list is not wrong, it is national. The naming trends source guide sets out which source to quote for which claim.
Checklist before you publish a name claim
Run these six checks before any name claim goes out.
[figure 2]
Name Claim Checklist
- Name the datasetSSA state list or NCHS birth file
- Give the year and the state, not just the country
- Note the five-baby threshold if the source is SSA
- Check whether the name's spelling varies by state
- Separate rank from share of births
- Avoid causal language unless a source supports it
- Link the reader to the primary source, not a summary
- Name the file you counted and the year, whether an SSA state list or a state birth file.
- State the threshold you accepted, usually the SSA five-baby rule.
- Put the state's total births for that year next to the name count.
- Mark whether the release is state or city, since New York City files separately.
- Label a rank change as rank, and a share change as share.
- Say whether the state office can confirm the name at all.
