Methodology: the 1980s decade hub

What the ownership score measures — and what it does not

The ownership score is a descriptive statistic about recorded births. It ranks names by how concentrated their recorded history is inside 1980–1989, adjusted so rare names cannot dominate, and weighted by how visible the name actually was. It does not measure why parents chose a name, what a name meant culturally, or anything about the people who carry it. A high score means one thing only: an unusually large share of this name's recorded use happened in this decade, given its size.

Data source and coverage

All figures derive from U.S. Social Security Administration national birth records, 1980 through 1989 — a complete decade in the current dataset, which runs through 2025. Rankings, ownership scores, the classroom, and spelling families are computed offline from the same source and served as a precomputed profile; nothing is recalculated in your browser.

Source limitations

• The SSA suppresses name-and-sex counts below 5 per year, so very rare names are under-counted everywhere on this hub.
• Records reflect sex as recorded at birth, in two categories; they are not a record of gender identity.
• Coverage is national only — no state or regional breakdown is used here.
• “Lifetime” births mean recorded births from 1880 through 2025, the SSA file's own span, not all of history.

Eligibility

Female and male records are separate comparison sets end to end; a name recorded for both sexes is scored independently in each. A name-and-sex pair is eligible for ownership ranking if it meets either condition:

• at least 5,000 recorded births during 1980–1989, or
• a top-1,000 national rank for its sex in at least five distinct years of 1980–1989.

The ownership score, formula by formula

Raw concentration

The share of a name's recorded lifetime births that occurred in 1980–1989. Kept in the data for transparency, but never presented as the score on its own.

raw_concentration = births_in_decade / lifetime_births

Prior decade share

The baseline concentration of the whole eligible set, computed separately for the female and male comparison sets (two priors). The shrinkage step applies each sex's own prior: 10.91% for girls, 10.49% for boys. The pooled reference value is 10.68%.

prior_decade_share = Σ births_in_decade (eligible set) / Σ lifetime_births (eligible set), computed per sex

Adjusted concentration

Empirical-Bayes-style shrinkage toward the prior, so a low-volume name cannot top the table on a lucky ratio. The smoothing strength α = 2,500 was chosen by sensitivity testing, not arbitrarily (see below).

adjusted_concentration = (births_in_decade + α × prior_decade_share) / (lifetime_births + α), α = 2,500

Raw prominence

How visible the name was during the decade, on a log scale so that sheer size matters but does not dominate.

raw_prominence = ln(1 + births_in_decade)

Normalization

Both prominence and adjusted concentration are min-max normalized within each sex's eligible set, so each component lands on a 0–1 scale.

normalized_x = (x − min(x)) / (max(x) − min(x)), over the eligible set of that sex

Ownership score

The final score weights concentration (identity) at 70% and prominence (visibility) at 30%. Data retains four decimal places; this site rounds for display only.

ownership_score = 100 × (0.70 × normalized_concentration + 0.30 × normalized_prominence)

Choosing α

α controls how strongly low-volume names are pulled toward the prior. Candidate values 500, 1,000, 2,500, 5,000, and 10,000 were compared with a sensitivity script that tabulates the top 25 at each value, rank churn between values, and intrusions by names with fewer than 5,000 decade births. The selected value, α = 2,500, is the smallest candidate with zero low-volume intrusions in the top 25 that preserved intuitive ordering among substantial names.

Ties and ranking views

All sorts are deterministic: ties break toward higher 1980s births, then alphabetically. The ranking views are fixed rules, not editorial picks: Most Owned sorts by ownership score; Most Popular by decade births; Popular but Timeless keeps names at or above the median births of the pooled eligible set (both sexes) and at or below its 25th percentile of adjusted concentration; Unexpected Results keeps names whose popularity rank exceeds their ownership rank by 20 or more within their sex set.

Diversity and concentration

Name share

Every name-and-sex row's share of all recorded births in the decade, pooled across sexes.

p_i = births_i / total_births

Shannon entropy

How evenly births are spread across names. Higher means more even.

H = −Σ (p_i × ln(p_i))

Effective number of names

The headline diversity figure: the number of equally common names that would produce the same entropy. For the 1980s, 986.

N_eff = exp(H)

Diversity score

Entropy normalized to 0–100 against the maximum possible entropy for 38,615 distinct name-and-sex rows. The 1980s scores 65.3.

diversity_score = 100 × H / ln(N_distinct)

Concentration score

A Herfindahl index normalized so 0 means perfectly diffuse and 100 means one name takes every birth. The 1980s scores 0.4.

HHI = Σ p_i²; concentration_score = 100 × (HHI − 1/N) / (1 − 1/N)

Pooling note: shares are computed over name-and-sex rows pooled across both sexes, so “Michael, boys” and “Michael, girls” are separate entries. Top-10 and top-100 shares use the same pooled table.

The 1984 classroom

The classroom is a statistical reconstruction of an average classroom, not an actual class record. It uses the year 1984, 30 seats, and national records. Seat allocation:

1. femaleSeats = round(30 × F_total / (F_total + M_total)) from actual 1984 totals; boys take the remainder (for 1984: 14 / 16).
2. expected_seats_i = count_i / sex_total × sexSeats for every name recorded in 1984.
3. Whole seats are floored; the leftovers are distributed by largest remainder, ties broken by higher count, then alphabetically. If rounding still leaves a gap, the alphabetically first maximal-remainder name is adjusted — deterministic either way.

Repeated names are expected and are not suppressed. The same inputs always produce the same roster.

Spelling families

Conventional rankings separate spelling variants. This view groups manually reviewed variants to show their combined demographic footprint.

Families come from one manually reviewed curation file — there is no automatic clustering. A family ships only with review status “approved”, at least two variants each recording 1,000 or more 1980s births, and a combined 20,000 or more. A variant may appear in only one approved family. The combined decade rank is where the family's summed births would sit in the decade's per-sex rankings; each variant's own rank is shown alongside.

Missing and suppressed data

Years in which a name falls below the SSA's five-birth suppression threshold contribute zero to that name's counts here — the suppression means the true count is unknown, between zero and four. Names with no recorded 1980s births are absent rather than zeroed. Status labels reuse this site's existing name-status taxonomy and read “unknown” where no status is on record.

Updates and versioning

This profile is regenerated when the SSA releases a new annual vintage. Any change to a formula bumps the methodology version; the version, build timestamp, source vintage, and build commit (when available) are rendered from the payload itself at the top of this page.

Back to the 1980s hub