FactHub

FactHub curated dataset

Girls' and boys' high school sports participation by state, 2015-16 to 2025-26, with state sports-eligibility laws

NFHS participation counts by state and sex for nine school years, Census population aged 14-17 by sex, and the dates of the 27 state laws limiting girls' school teams to female students, with difference-in-differences comparisons of states with and without such laws.

Coverage: School years 2015-16 to 2018-19 and 2021-22 to 2025-26 (no NFHS survey for 2019-20 or 2020-21)

Node: dataset:girls-hs-sports-participation-by-state

Send feedback on this dataset

What this dataset measures

Built to test whether girls' high school sports participation fell in states after they passed laws like Washington Initiative IL26-638. NFHS is the only national source of participation by state and sex; Census estimates give per-capita denominators.

Geography
United States, 50 states and the District of Columbia
Unit of observation
State (or DC) by school year
Update cadence
Annual, after each NFHS survey (usually August)
Curator
FactHub Research Team

Curated data tables

These related tables form this versioned dataset release. Each table has its own downloadable file and field definitions.

Participation and population by state and school year

NFHS boys and girls participants, Census population aged 14-17 by sex, law group, and derived rates.

Unit of observation: State by school year

View this table

state-year-participation.csv text/csv · 37.5 KB (38,398 bytes)
15 field definitions
FieldTypeMeaningUnitsValue type
state string State name or District of Columbia — observed
schoolYear string School year, such as 2018-19 — observed
group string law (one of the 27 law states), noLaw (no statewide law or policy), or policy (Alaska, Virginia: state policy, not law) — derived
lawFirstYear string First school year counted as covered by the state law — derived
girls integer Girls participants, Totals by State table participation entries observed
boys integer Boys participants, Totals by State table participation entries observed
totalReported integer Total as printed by NFHS (equals boys + girls except North Dakota 2021-22, +90, and Kansas 2024-25, +1) participation entries observed
girlsPop14to17 integer Civilian female population aged 14-17 on July 1 of the school year's first calendar year persons estimated
boysPop14to17 integer Civilian male population aged 14-17, same basis persons estimated
girlsPer1000 number 1,000 x girls / girlsPop14to17 participants per 1,000 girls aged 14-17 derived
boysPer1000 number 1,000 x boys / boysPop14to17 participants per 1,000 boys aged 14-17 derived
girlsSharePct number Girls as a share of boys + girls participants percent derived
girlsFlagFootball integer Girls' flag football participants (2018-19, 2024-25, 2025-26 only; 0 where the state reported none; empty in other years) participation entries observed
girlsWrestling integer Girls' wrestling participants (2018-19, 2024-25, 2025-26 only; empty in other years) participation entries observed
girlsExFlagWrestlingPer1000 number Girls per 1,000 excluding flag football and wrestling (empty in other years) participants per 1,000 girls aged 14-17 derived

State laws limiting girls' school teams to female students

The 27 laws counted, enactment dates and first covered school year.

Unit of observation: State law

View this table

state-laws.csv text/csv · 1.7 KB (1,701 bytes)
6 field definitions
FieldTypeMeaningUnitsValue type
state string State — observed
bill string Bill number — observed
enacted date Date signed, veto overridden, or became law without signature — observed
firstYear string First school year counted as covered — derived
enjoined string yes if the law was blocked in court for all or most students for part of the period — derived
note string Effective dates, court history, scope — observed

Supporting files

Source snapshots, manifests, quality reports, documentation, and other materials used to produce or review this release.

  • parse.js Transformation manifest · text/javascript · 1.3 KB (1,350 bytes)

    Reads the Totals by State table from each NFHS survey (pdftotext -raw output).

  • sport.js Transformation manifest · text/javascript · 1.8 KB (1,829 bytes)

    Reads one girls' sport column by state from the NFHS survey (pdftotext -table output); used for wrestling and 2025-26 flag football.

  • analyze.js Transformation manifest · text/javascript · 4.4 KB (4,464 bytes)

    Joins NFHS totals, law groups and Census population into the state-year panel and computes group series.

  • analyze2.js Transformation manifest · text/javascript · 6.6 KB (6,735 bytes)

    Computes every difference-in-differences estimate, robustness check and table in this dataset.

  • build-csv.js Transformation manifest · text/javascript · 1.1 KB (1,139 bytes)

    Writes the two CSV tables from the computed panel.

  • laws.json Transformation manifest · application/json · 2.9 KB (2,984 bytes)

    Law list used by the scripts (same content as state-laws.csv).

  • f2425.json Transformation manifest · application/json · 328 bytes

    Girls' flag football by state, 2024-25, transcribed from the NFHS other-sports list; sums to the national 68,847.

  • f1819.json Transformation manifest · application/json · 112 bytes

    Girls' flag football by state, 2018-19, transcribed from the NFHS other-sports list; sums to the national 11,209.

  • README.md Documentation · text/markdown · 1.8 KB (1,823 bytes)

    How to reproduce the tables and analyses from the NFHS and Census files.

  • dataset-parts.json Documentation · application/json · 62.1 KB (63,603 bytes)

    Machine-readable copy of version 1 sources, tables, events and saved analyses (version 2 adds four analyses and two sources, listed in the changelog).

  • v2-checks.js Transformation manifest · text/javascript · 4.9 KB (5,027 bytes)

    Computes the catch-up regression, the timing table and the universal-verification comparison added in version 2.

Saved analyses

Participation per 1,000 high-school-age residents, law states vs. no-law states

Question: How did girls' and boys' high school sports participation per 1,000 residents aged 14-17 change in states that passed laws limiting girls' teams to female students, compared with states that did not?

Result: Girls per 1,000: law states 370.1 (2018-19) to 401.8 (2025-26); no-law states 453.4 to 452.7. Boys per 1,000: law states 512.2 to 527.7; no-law states 543.3 to 542.8.

Operation: Sum NFHS participants by sex across each group of states, divide by the group's Census civilian population aged 14-17 of that sex, times 1,000.

Scope: School years 2015-16 to 2025-26; 2019-20 and 2020-21 have no survey and are null.

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Question: Did girls' participation per 1,000 girls aged 14-17 change differently in the 24 law states than in the 22 no-law states between the last pre-pandemic survey (2018-19) and the latest (2025-26)?

Result: Law states +45.2 per 1,000, no-law states -0.8; difference +46.0 (95% CI 8.4 to 83.5).

Operation: For each state, rate in 2025-26 minus rate in 2018-19; mean in law states minus mean in no-law states.

Scope: 24 law states vs 22 no-law jurisdictions; states unweighted.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Law states minus no-law states, change in girls per 1,000, 2018-19 to 2025-26 46 ± 18.6 participants per 1,000 girls aged 14-17 (95% CI 8.4 to 83.5) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness checks

Sensitivity: girls per 1,000, 2018-19 to 2024-25

Question: Does the result hold if the comparison ends in 2024-25?

Result: +46.7 (95% CI 18.6 to 74.7); law states +19.6, no-law -27.1.

Operation: As did-girls-main, ending in 2024-25.

Scope: 24 law vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000, 2018-19 to 2024-25, law minus no-law 46.7 ± 13.9 participants per 1,000 girls aged 14-17 (95% CI 18.6 to 74.7) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Sensitivity: girls per 1,000 excluding flag football and wrestling

Question: Is the gap explained by girls' flag football and wrestling, which grew fastest and were sanctioned unevenly across states?

Result: +40.9 (95% CI 3.4 to 78.4); law states +24.2, no-law -16.7. Ending 2024-25: +42.3 (15.0 to 69.5).

Operation: As did-girls-main after subtracting each state's girls' flag football and wrestling participants.

Scope: 24 law vs 22 no-law; 2018-19 to 2025-26.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000 excluding flag football and wrestling, law minus no-law 40.9 ± 18.6 participants per 1,000 girls aged 14-17 (95% CI 3.4 to 78.4) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Sensitivity: excluding states with questionable NFHS figures

Question: Do states whose NFHS totals repeat a prior year or jump more than 25% in one year drive the result?

Result: +39.0 (95% CI 16.8 to 61.1).

Operation: As did-girls-main without the flagged jurisdictions.

Scope: 19 law vs 16 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000, flagged states excluded, law minus no-law 39 ± 10.8 participants per 1,000 girls aged 14-17 (95% CI 16.8 to 61.1) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 35

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Sensitivity: excluding laws blocked in court

Question: Does the result change without Idaho, Utah and Arizona, whose laws were enjoined for some or all students?

Result: +47.4 (95% CI 7.5 to 87.2).

Operation: As did-girls-main without those three states.

Scope: 21 law vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000, enjoined-law states excluded, law minus no-law 47.4 ± 19.7 participants per 1,000 girls aged 14-17 (95% CI 7.5 to 87.2) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 43

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Sensitivity: population-weighted rates

Question: Does weighting by population change the result?

Result: +32.5 (95% CI 2.0 to 74.7); law group 370.1 to 401.8 (+31.8), no-law 453.4 to 452.7 (-0.8).

Operation: Change in pooled girls per 1,000 (law group) minus change in pooled rate (no-law group); 95% interval from 5,000 state bootstrap draws.

Scope: 24 law vs 22 no-law; 2018-19 to 2025-26.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in pooled girls per 1,000, law minus no-law 32.5 participants per 1,000 girls aged 14-17 (95% CI 2 to 74.7) 95% confidence interval Pooled rates; percentile state bootstrap, 5,000 draws 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Placebo: the same comparison before any law (2015-16 to 2018-19)

Question: Were the future law states already gaining faster before the laws?

Result: -16.0 (95% CI -38.2 to 6.2): the future law states were not gaining faster beforehand.

Operation: As did-girls-main for 2015-16 to 2018-19, when no state had such a law.

Scope: 24 future-law vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000, 2015-16 to 2018-19, future-law minus no-law -16 ± 11 participants per 1,000 girls aged 14-17 (95% CI -38.2 to 6.2) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Comparison: boys per 1,000, 2018-19 to 2025-26

Question: Did boys' participation also grow faster in law states?

Result: +21.9 (95% CI -15.3 to 59.1); law states +30.7, no-law +8.8.

Operation: As did-girls-main with boys per 1,000 boys aged 14-17.

Scope: 24 law vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in boys per 1,000, law minus no-law 21.9 ± 18.5 participants per 1,000 boys aged 14-17 (95% CI -15.3 to 59.1) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Comparison: girls' share of all participants

Question: Did girls gain relative to boys in law states?

Result: +1.4 points (95% CI 0.3 to 2.6); law states +1.1, no-law -0.4. Excluding flagged states: +1.8 (0.8 to 2.8).

Operation: As did-girls-main with girls / (girls + boys), in percentage points.

Scope: 24 law vs 22 no-law; 2018-19 to 2025-26.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls' share of participants, law minus no-law 1.4 ± 0.5 percentage points (95% CI 0.3 to 2.6) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Comparison: percent change in girls per 1,000

Question: How large is the gap in percent terms?

Result: Law states +10.9% on average, no-law +0.9%; difference +10.1 points (95% CI 1.6 to 18.5).

Operation: As did-girls-main using each state's percent change.

Scope: 24 law vs 22 no-law; 2018-19 to 2025-26.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Percent change in girls per 1,000, law minus no-law 10.1 ± 4.2 percentage points (95% CI 1.6 to 18.5) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

Law-state minus no-law-state change, across specifications

Question: How robust is the gap in girls' participation between law and no-law states across specifications, and how does it compare with boys and with the period before the laws?

Result: Main: 2018-19 to 2025-26: +46 (8.4 to 83.5); Ending 2024-25: +46.7 (18.6 to 74.7); Without flag football and wrestling: +40.9 (3.4 to 78.4); Without questionable NFHS figures: +39 (16.8 to 61.1); Without court-blocked laws: +47.4 (7.5 to 87.2); Controlling for starting rate (catch-up): +39.5 (-0.5 to 79.5); 2022-23 laws, measured from 2021-22: -12.6 (-72.7 to 47.6); Placebo: 2015-16 to 2018-19, before any law: -16 (-38.2 to 6.2); Boys (per 1,000 boys): +21.9 (-15.3 to 59.1)

Operation: Collects the estimates of the individual saved analyses for charting.

Scope: See each row.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Main: 2018-19 to 2025-26 46 ± 18.6 participants per 1,000 girls aged 14-17 (95% CI 8.4 to 83.5) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46
Ending 2024-25 46.7 ± 13.9 participants per 1,000 girls aged 14-17 (95% CI 18.6 to 74.7) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46
Without flag football and wrestling 40.9 ± 18.6 participants per 1,000 girls aged 14-17 (95% CI 3.4 to 78.4) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46
Without questionable NFHS figures 39 ± 10.8 participants per 1,000 girls aged 14-17 (95% CI 16.8 to 61.1) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 35
Without court-blocked laws 47.4 ± 19.7 participants per 1,000 girls aged 14-17 (95% CI 7.5 to 87.2) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 43
Controlling for starting rate (catch-up) 39.5 ± 19.83 participants per 1,000 girls aged 14-17 (95% CI -0.5 to 79.5) ± one standard error; 95% confidence interval OLS of each state's change on a law-state indicator and its starting rate; HC1 robust standard errors; 95% t interval (see catch-up-check) 46
2022-23 laws, measured from 2021-22 -12.6 ± 29.4 participants per 1,000 girls aged 14-17 (95% CI -72.7 to 47.6) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit (see did-girls-cohort-b-post-covid-base) 32
Placebo: 2015-16 to 2018-19, before any law -16 ± 11 participants per 1,000 girls aged 14-17 (95% CI -38.2 to 6.2) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46
Boys (per 1,000 boys) 21.9 ± 18.5 participants per 1,000 boys aged 14-17 (95% CI -15.3 to 59.1) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 46

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • The rows use different base years and controls; the 2021-22-base row covers only the 10 states whose laws took effect in 2022-23.

Change by state, 2018-19 to 2025-26

Question: In which states did girls' participation per 1,000 rise or fall?

Result: Girls per 1,000 fell in 4 of 24 law states (Florida, North Carolina, Louisiana, North Dakota) and in 12 of 22 no-law jurisdictions. Median change: law +31.3, no-law -8.9.

Operation: Per-state rates in 2018-19 and 2025-26 and their difference.

Scope: All 51 jurisdictions.

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.

By year the law took effect

Question: Is the gap similar for each wave of laws?

Result: Laws first applied 2021-22 (9 states), 2018-19 to 2025-26: +62.2 (2.8 to 121.7); Laws first applied 2022-23 (10 states), 2018-19 to 2025-26: +42.6 (-7.2 to 92.5); Laws first applied 2023-24 (5 states), 2022-23 to 2025-26: +11.5 (-64.1 to 87.1)

Operation: As did-girls-main for each cohort; the 2023-24 cohort uses 2022-23 as its last pre-law year.

Scope: Each cohort vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Laws first applied 2021-22 (9 states), 2018-19 to 2025-26 62.2 ± 27.7 participants per 1,000 girls aged 14-17 (95% CI 2.8 to 121.7) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 31
Laws first applied 2022-23 (10 states), 2018-19 to 2025-26 42.6 ± 23.8 participants per 1,000 girls aged 14-17 (95% CI -7.2 to 92.5) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 32
Laws first applied 2023-24 (5 states), 2022-23 to 2025-26 11.5 ± 35.2 participants per 1,000 girls aged 14-17 (95% CI -64.1 to 87.1) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 27

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • Cohorts are small; the 2023-24 cohort has 5 states and only three post-law years.

States whose laws took effect in 2024-25 or 2025-26

Question: What happened to girls' participation in Ohio, New Hampshire and Nebraska after their laws?

Result: Ohio +114.3 (but its 2022-23 and 2023-24 NFHS figures are identical, so its base year is suspect); New Hampshire -8.9 vs no-law +41.9; Nebraska +21.2 vs no-law +26.3.

Operation: Rate in the last pre-law year and in 2025-26, beside the no-law mean change over the same years.

Scope: Ohio, New Hampshire, Nebraska.

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • One or two post-law years; single states; Ohio's pre-law figures appear carried forward.

Catch-up check: law-state gap controlling for each state's starting rate

Question: Law states started lower (370 vs 453 girls per 1,000 in 2018-19). Is their faster gain just low-rate states catching up?

Result: Controlling for starting rate, the law-state gap is +39.5 per 1,000 for 2018-19 to 2025-26 (95% CI -0.5 to 79.5) and +40.8 for 2018-19 to 2024-25 (13.5 to 68.1). Since 2018-19, no-law states that started higher gained less (slope -0.22, CI -0.40 to -0.04), a catch-up pattern absent before 2018-19 (slope -0.04).

Operation: OLS of each state's change on a law-state indicator and its starting rate (centered at 450 per 1,000); HC1 robust standard errors; 95% t interval

Scope: 24 law vs 22 no-law states.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Law-state gap, 2018-19 to 2025-26, controlling for starting rate 39.5 ± 19.83 participants per 1,000 girls aged 14-17 (95% CI -0.5 to 79.5) ± one standard error; 95% confidence interval OLS of each state's change on a law-state indicator and its starting rate (centered at 450 per 1,000); HC1 robust standard errors; 95% t interval 46
Law-state gap, 2018-19 to 2024-25, controlling for starting rate 40.8 ± 13.54 participants per 1,000 girls aged 14-17 (95% CI 13.5 to 68.1) ± one standard error; 95% confidence interval OLS of each state's change on a law-state indicator and its starting rate (centered at 450 per 1,000); HC1 robust standard errors; 95% t interval 46
Law-state gap, 2015-16 to 2018-19 (before any law), controlling for starting rate -17.5 ± 11.59 participants per 1,000 girls aged 14-17 (95% CI -40.9 to 5.9) ± one standard error; 95% confidence interval OLS of each state's change on a law-state indicator and its starting rate (centered at 450 per 1,000); HC1 robust standard errors; 95% t interval 46
Slope on starting rate, no-law states only, 2018-19 to 2025-26 -0.22 ± 0.08 change per 1,000 per unit of starting rate (95% CI -0.4 to -0.04) ± one standard error; 95% confidence interval OLS within the 22 no-law jurisdictions; HC1 robust standard errors 22
Slope on starting rate, no-law states only, 2015-16 to 2018-19 -0.04 ± 0.03 change per 1,000 per unit of starting rate (95% CI -0.09 to 0.02) ± one standard error; 95% confidence interval OLS within the 22 no-law jurisdictions; HC1 robust standard errors 22

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • 46 states; the controlled estimate for 2025-26 has an interval that just reaches zero.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

Sensitivity: 2022-23 laws measured from 2021-22, the first post-pandemic survey

Question: For the 10 states whose laws took effect in 2022-23, how did girls' participation change from the last pre-law year (2021-22) to 2025-26 compared with no-law states?

Result: -12.6 per 1,000 (95% CI -72.7 to 47.6): from the first post-pandemic survey, these law states gained about as much as no-law states (+63.3 vs +75.9), which were recovering from a deeper pandemic dip.

Operation: As did-girls-main for the 2022-23 cohort, using 2021-22 as the base year.

Scope: 10 law vs 22 no-law.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Change in girls per 1,000, 2021-22 to 2025-26, 2022-23 law states minus no-law states -12.6 ± 29.4 participants per 1,000 girls aged 14-17 (95% CI -72.7 to 47.6) ± one standard error; 95% confidence interval Difference-in-differences of state-level changes; 95% Welch t interval, states as the unit 32

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • 2021-22 was depressed by the pandemic, more in no-law states, so this base favors no-law states; the 2018-19 base is before both the pandemic and the laws.

Robustness check for: Difference-in-differences: girls per 1,000, 2018-19 to 2025-26

When the rise happened

Question: In which year did girls' participation in law states pass its 2018-19 level?

Result: Law states above their 2018-19 rate: 6 of 24 in 2021-22, 8 in 2022-23, 9 in 2023-24, 16 in 2024-25, 20 in 2025-26. Pooled law-state rate: 353.4, 353.4, 366.6, 379.1, 401.8 (2018-19: 370.1). No-law states: 400.2, 418.2, 427.9, 439.1, 452.7 (2018-19: 453.4).

Operation: Count of states whose girls per 1,000 exceeded their own 2018-19 rate, and pooled rates, by school year.

Scope: 24 law and 22 no-law states, 2021-22 to 2025-26.

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • Much of the rise came in 2025-26, a year with several erratic state reports.

States that require proof of sex from every athlete

Question: Did girls' participation fall in the states that require every athlete, or every girl in girls' sports, to document sex?

Result: Nebraska (doctor-signed document, every athlete; 2024-25 to 2025-26): girls +21.2 per 1,000 vs no-law median +10.1; boys -11.5. New Hampshire (birth certificate or other evidence; 2023-24 to 2025-26): girls -8.9 vs +30.1; boys -17.5; girls' share flat. Oklahoma (annual parent affidavit; 2018-19 to 2025-26): girls +31.8 vs -8.9; from 2021-22: girls +0.2 vs +57.2, boys +33.6, girls' share -1.6 points.

Operation: Change in girls per 1,000, boys per 1,000 and girls' share from the last pre-law year (and, for Oklahoma, also from 2018-19) to 2025-26, beside the median change in the 22 no-law jurisdictions over the same years.

Scope: Nebraska, New Hampshire, Oklahoma.

View this analysis

  • Observational comparison: states chose these laws, and the law states differ from the others (region, politics, school sports culture). The estimate is a difference in trends, not a measured effect of the laws.
  • NFHS counts are self-reported by state associations; some states repeat a prior year's figures or jump more than 25% in a year.
  • Participation entries, not unique athletes: a girl in three sports counts three times.
  • Three single states with one to four post-law years; none uses I-638's provider check based only on anatomy, genetics or testosterone. Nebraska's law took effect September 3, 2025, after fall practice began.

Sources

  1. NFHS High School Athletics Participation Survey, 2015-16 Participation counts by state and sex (Totals by State table) for 2015-16

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  2. NFHS High School Athletics Participation Survey, 2016-17 Participation counts by state and sex (Totals by State table) for 2016-17

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  3. NFHS High School Athletics Participation Survey, 2017-18 Participation counts by state and sex (Totals by State table) for 2017-18

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  4. NFHS High School Athletics Participation Survey, 2018-19 Participation counts by state and sex (Totals by State table) for 2018-19, plus girls' wrestling and flag football by state

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  5. NFHS High School Athletics Participation Survey, 2021-22 Participation counts by state and sex (Totals by State table) for 2021-22

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  6. NFHS High School Athletics Participation Survey, 2022-23 Participation counts by state and sex (Totals by State table) for 2022-23

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  7. NFHS High School Athletics Participation Survey, 2023-24 Participation counts by state and sex (Totals by State table) for 2023-24

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  8. NFHS High School Athletics Participation Survey, 2024-25 Participation counts by state and sex (Totals by State table) for 2024-25, plus girls' wrestling and flag football by state

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  9. NFHS High School Athletics Participation Survey, 2025-26 Participation counts by state and sex (Totals by State table) for 2025-26, plus girls' wrestling and flag football by state

    Retrieved October 23, 2026. NFHS copyright; factual figures used with attribution. Derived state totals republished with attribution; the PDF itself is linked, not hosted.

    Reliability: Other. The National Federation of State High School Associations compiles counts reported by its 51 member state associations. A student in two sports counts twice. The survey takes no position on state eligibility laws.

  10. NFHS news release (August 2026): Led by Girls Flag Football and Girls Wrestling, High School Sports Participation Increases for Fifth Consecutive Year National 2025-26 totals and the growth of girls' flag football and wrestling

    Retrieved October 23, 2026. NFHS copyright; quoted with attribution. Linked, not hosted.

    Reliability: Other. Has a stake in this finding. NFHS promotes high school sports and reports on its own survey; its framing of growth is promotional, but the figures match the survey tables.

  11. U.S. Census Bureau, Annual Estimates of the Civilian Population by Single Year of Age and Sex for States, 2010-2019 (SC-EST2019-AGESEX-CIV) July 1, 2015-2018 civilian population ages 14-17 by sex and state (denominators for 2015-16 to 2018-19) Free copy

    Retrieved October 23, 2026. Public domain (U.S. government work). Aggregates republished.

    Reliability: Government. Census Bureau population estimates (Vintage 2019), built from the 2010 Census plus births, deaths and migration.

  12. U.S. Census Bureau, Annual Estimates of the Civilian Population by Single Year of Age and Sex for the United States and States: April 1, 2020 to July 1, 2025 (SC-EST2025-AGESEX-CIV) July 1, 2021-2025 civilian population ages 14-17 by sex and state (denominators for 2021-22 to 2025-26) Free copy

    Retrieved October 23, 2026. Public domain (U.S. government work). Aggregates republished.

    Reliability: Government. Census Bureau population estimates (Vintage 2025), built on the 2020 Census base; not strictly continuous with Vintage 2019.

  13. Idaho Legislature, House Bill 500 (2020), Fairness in Women's Sports Act: bill status Idaho enactment date (signed March 30, 2020; effective July 1, 2020)

    Retrieved October 23, 2026. Public government record. Linked.

    Reliability: Government. Official legislative status page; the page notes status updates are unofficial until the journal.

  14. NPR, Tennessee Becomes 3rd State This Month To Enact Restrictions For Transgender Athletes (March 27, 2021) Enactment dates for the Mississippi, Arkansas and Tennessee laws

    Retrieved October 23, 2026. News report; cited. Linked.

    Reliability: News report. National news report written at the time of signing.

  15. ESPN, Transgender athlete laws by state: Legislation, science, more (August 2023) State-by-state bill numbers and enactment dates for laws enacted 2020-2023, and court blocks

    Retrieved October 23, 2026. News report; cited. Linked.

    Reliability: News report. State-by-state tracker compiled by ESPN reporters; dated August 2023, so it predates the 2024-2025 laws.

  16. Trans Athlete (transathlete.com), state policy tracker, Take Action page Wyoming enactment date; Alaska and Virginia listed as policy, not law; count of 29 states

    Retrieved October 23, 2026. Advocacy page; cited. Linked.

    Reliability: Advocacy group. Maintained by an advocate for transgender athletes' inclusion, which opposes these laws; used only for dates and classifications.

  17. Axios Columbus, Ohio bill restricting trans health care to take effect in 90 days (January 24, 2024) Ohio HB 68 veto override date

    Retrieved October 23, 2026. News report; cited. Linked.

    Reliability: News report. Local news report on the Senate override vote.

  18. New Hampshire Bulletin, Sununu signs bans on trans girls in girls' sports, gender affirming surgeries for minors (July 19, 2024) New Hampshire HB 1205 signing date

    Retrieved October 23, 2026. News report; cited. Linked.

    Reliability: News report. Nonprofit state news outlet reporting the signing.

  19. Office of Governor Jim Pillen, Governor Pillen Signs the Stand With Women Act (June 4, 2025) Nebraska LB 89 signing date

    Retrieved October 23, 2026. Public government record. Linked.

    Reliability: Government. Has a stake in this finding. Press release from the governor who signed the law.

  20. Nebraska Legislature, LB 89 (2025), Stand With Women Act, slip law Nebraska requirement that every athlete give a doctor-signed document confirming sex

    Retrieved October 23, 2026. Public government record. Linked.

    Reliability: Government. Official slip law published by the Nebraska Legislature.

  21. Oklahoma Legislature, SB 2 (2022), enrolled Oklahoma requirement of an annual parent affidavit of sex for every athlete

    Retrieved October 23, 2026. Public government record. Linked.

    Reliability: Government. Enrolled bill text published by the Oklahoma Legislature.

Limitations and reproducibility

Reproducible: All inputs are public. The scripts (artifacts) parse the NFHS PDFs, join Census estimates, and compute every analysis; the README lists the steps. Extracted columns were checked against NFHS national totals.

  • Observational: the law states differ from the others (mostly Southern, Plains and Mountain West states with Republican legislatures). A difference in trends is not proof that the laws caused it.
  • NFHS figures are reported by each state association and are not audited. Some states repeat an earlier year exactly (Ohio 2022-23 and 2023-24; Connecticut, Delaware, Louisiana, Maine, Vermont and DC in some years), and some jump by more than 25% in a year (for example Vermont +103%, Kentucky +38% and Washington +28% in 2025-26).
  • Counts are participation entries, not unique students: a student in two sports counts twice.
  • No survey for 2019-20 or 2020-21 (COVID-19). Pandemic limits on school sports lasted longer in some no-law states, which may have slowed their recovery.
  • Girls' flag football and wrestling grew fast after 2021, and flag football was sanctioned first in several Southern law states; a version excluding both sports is included.
  • The laws differ (grades covered, how sex is established, enforcement), and some were blocked in court for some or all students (Idaho, Utah, Arizona; West Virginia and New Hampshire for named plaintiffs).
  • None of these laws requires every girl to have her sex verified by a health care provider as I-638 Sec. 3(2) would; this data cannot test that requirement.
  • Census Vintage 2019 (for 2015-2018) and Vintage 2025 (for 2021-2025) use different census bases.
  • A federal executive order of February 2025 applied nationwide, so in 2024-25 and 2025-26 the comparison is between states with and without their own law.
  • Law states started with lower girls' participation (370 vs 453 per 1,000 in 2018-19), and after the pandemic gap lower-rate states gained more; part of the law states' faster gain may be catch-up (see catch-up-check).

Publications using this dataset

Version details

The exact citing versions, and the version of this dataset each cites.

  • Article girls-sports-participation-after-state-laws — version 6 (published) cites version 3
  • Article wa-i638-girls-event-placements — version 5 (published) cites version 3
  • Chart girls-sports-participation-did-estimates — version 2 (published) cites version 3
  • Chart girls-hs-sports-rate-law-vs-no-law-states — version 2 (published) cites version 3
  • Briefing wa-i-638 — version 5 (published) cites version 3
  • Briefing wa-i-638 — version 4 (published) cites version 3