Pooling 1950–2024 birth years adds 0.172 percentage points to minimum recorded sex disagreement in listed US SSA name counts
We measured the effect of combining birth years in a fixed mirror of US Social Security Administration name counts. The test used all listed counts for 1950 through 2024. A separate majority rule for each name and year gave 2.042% disagreement with recorded sex. One rule for each name across all years gave 2.214%. Pooling added 470,519 disagreements in 273,549,115 listed births, or 0.172 percentage points. Fifty names accounted for 53.7% of the increase. We proved sharp limits under a stated completion assumption. Each absent opposite-sex count could range from zero through four within a listed name-year. Under that assumption, the added rate ranged from 0.149 to 0.172 percentage points. Names absent for both sexes in a year were excluded. An exploratory reference fixed each name's pooled sex counts and annual totals. Its exact expected penalty was 27,366.1, compared with 470,519 observed. This mean is not a p-value. All 1,687,529 source rows matched a second pinned mirror. Independent counts and repeat runs checked the results. The package preserves all primary source data, code, hashes, figures, tables, and repeat instructions. These results describe recorded sex in listed counts. They do not measure gender identity, naming intent, or prediction accuracy in new people.
birth names · cohort pooling · recorded sex · cultural change · omitted counts · reproducible benchmark
References
5- supportsdoi:10.63744/2h8m9cv8t9y4
The cohort comparison gives a quantitative case for use of birth time in name methods. The historical principle and its limits come from this prior work.
- doi:10.1145/2467696.2467720
Prior work studies temporal name associations and a pure error baseline.
- doi:10.32614/RJ-2016-002
Prior work discusses source coverage and clear error measures for name methods.
- doi:10.1086/210431
Earlier work studies changes in shared name use.
- doi:10.1016/j.poetic.2020.101501
Prior work studies stable shared name use and different accounts of name adoption. We do not test those accounts.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- A 1950s name rule had 4.53 percentage points more disagreement than a 2000s rule on shared names in listed 2010–2017 SSA counts
This prior study measures retrospective pooling disagreement. Our study tests fixed past rules on later birth years and reports a separate coverage comparison.
- usesTen prior birth years gave lower Brier loss than all prior years since 1950 in listed US name counts for 2000 to 2017
This prior study measured retrospective annual versus pooled majority disagreement. It informed the distinct test of rules trained only on earlier birth cohorts.