One college sport is vanishing, and the fastest-growing one has no name

Tennis is the college sport that is actually going away, and it is close to the only one.

Hold the institution set fixed and the file is unambiguous. On the 1,910 colleges that filed EADA in both AY 2014-15 and AY 2024-25, the number fielding tennis fell from 1,117 to 1,020. Go back twenty years, to the 1,779 that filed in both AY 2004-05 and AY 2024-25, and tennis goes from 1,099 to 975. That is minus 124 institutions, more than double any other decline in the file. The two next-largest losses, cross country at minus 54 and swimming and diving combined at minus 25, turn out to be colleges switching which survey code they file under. Outside those three, no sport lost more than five institutions in twenty years.

Now the list the raw comparison produces, which is the one most people run: softball down 39 institutions, basketball down 48, golf down 28, football down 10. All four are wrong. Fix the institution set and softball is plus 5, basketball plus 10, golf plus 7 and football plus 12. The raw list was measuring 163 colleges dropping out of the file and 127 different ones arriving.

The other half of the story is the largest gain in the file, and nothing in EADA can explain it. A survey bucket called Other Sports went from 60 institutions to 432 in ten years after twelve flat years, and it now carries more athletes than any single swimming or diving code. There is no free-text sport field anywhere in the collection, so the data cannot say what those teams play.

EADA is filed every year by every co-educational institution that takes Title IV aid and fields an intercollegiate athletics program. In AY 2024-25 that is 18,045 sport rows across 2,037 institutions, against 17,238 rows across 2,073 institutions in AY 2014-15. The net loss of 36 filers looks like a rounding error. It is not: 163 institutions in the AY 2014-15 file are gone by AY 2024-25 and 127 new ones have appeared, and those two flows are what manufactured the four false declines above.

So every trend here runs on a balanced panel: the 1,910 institutions that filed in both years, 92.1 percent of AY 2014-15 filers and 93.8 percent of AY 2024-25 filers, carrying 16,215 and 17,451 sport rows.

Three measurement notes, because EADA has a column for none of them. There is no team-count column, no roster-size column and no sport-status column. Institutions fielding a sport means distinct institutions with a row for that sport carrying nonzero participation. A men's team is a row with male participants and a women's team a row with female participants, so one sport row can be two teams. Participation counts roster spots rather than people, so a two-sport athlete is counted twice.

College athletics grew over the decade, which means a flat sport is a shrinking sport. On the same fixed 1,910 institutions, teams rose 9.4 percent (24,336 to 26,620) and participation rose 20.4 percent (619,105 to 745,223). Men were up 21.2 percent and women 19.2 percent. Those team counts are at the raw survey-code grain, one team per sport code per sex, which is the right grain for a sport-by-sport post and the wrong one for counting cuts.

Stretch the window to twenty years and both growth channels show. This one runs across all filers rather than a panel, so read it as context and not as a trend: teams per institution went from 11.57 to 13.50 and the average roster from 22.2 athletes to 27.8. Colleges are running more teams and putting more people on each of them. National participation in AY 2024-25 is 443,453 men and 322,383 women, 765,836 roster spots in all.

Read every sport below against plus 9.4 percent on teams and plus 20.4 percent on athletes. A sport growing at 11 percent is losing share while it grows.

Tennis peaked at 1,186 institutions in AY 2012-13. Of the twelve year-over-year changes since, nine are falls, one is flat, and the two rises both sit inside the rebound from the pandemic year. It was down to 1,124 by AY 2019-20, before COVID had anything to do with it, and it has not recovered since either: 1,074, then 1,076, 1,055 and 1,040 across the four years after the pandemic year.

The decline is symmetric across the sexes. On the balanced panel, men's tennis teams went from 934 to 839 between AY 2014-15 and AY 2024-25, and women's from 1,097 to 1,008.

Squad size held. Average tennis roster on the balanced panel went from 9.8 to 10.1 while the number of teams fell from 2,031 to 1,847. Nothing hollowed out; colleges stopped fielding the sport. Which colleges, and at what resource level, is the companion post's question, and its answer is the opposite of the intuitive one.

Among sports at 100 or more institutions, tennis is the only one where drops badly outnumber adds. On the balanced panel it was dropped at 156 institutions and added at 59, a ratio of 2.6 to 1. Cross country records more raw drops, 241, but almost as many adds, 225, which is the signature of a coding migration rather than a retreat.

None of which makes tennis small. It is still the seventh most widely fielded sport at 1,040 institutions, with 19,010 athletes. A large sport shedding programs steadily is harder to notice than a small one disappearing all at once.

The obvious next question is where the losses concentrate. They do not. Pin each institution's AY 2024-25 classification to both years, so reclassification cannot contaminate the comparison, and run it on the balanced panel: seventeen of the nineteen classifications either lose tennis or hold flat. The largest single loss is 16 institutions, in NCAA Division III with football. Even Division I-FBS, the best-funded group in the file, is down two.

Sector gives the same answer. On the balanced panel, public two-year institutions lost 18.9 percent of their men's tennis teams (95 to 77), public four-year 14.1 percent (284 to 244) and private nonprofit four-year 6.4 percent (550 to 515). Women's teams fall in every sector that has more than a handful of tennis programs.

Whatever is happening to college tennis is happening everywhere at once, and EADA records the disappearance without recording a reason. There is nothing in the file about cost, facilities, coaching supply or student demand, and the financial columns will not fill that gap: EADA revenue is an allocation that balances against expense for most filers, as an earlier post works through.

A sport can hold its team count while its rosters hollow out, and one large sport did exactly that. Rowing gained institutions on the balanced panel, 140 to 143, and lost 7.8 percent of its athletes, 10,023 to 9,244. Average roster fell from 48.7 to 44.4. It is the only large sport where squad size dropped materially.

The interesting cases run the other way. Softball, basketball and golf are flat on institutions (plus 0.3, plus 0.5 and plus 0.6 percent) and up on athletes (plus 14.6, plus 9.1 and plus 11.0 percent). Three of the four sports the raw comparison called dying are in that sentence, and football, the fourth, is plus 1.4 percent on institutions and plus 12.7 percent on athletes. None of them is dying. None of them is keeping up either: all four sit below the plus 20.4 percent panel baseline, growing and losing share at the same time.

Roster inflation is the second growth channel and it is uneven. Baseball added 7.8 athletes per team, 34.3 to 42.1, and football 11.5, 103.0 to 114.5. Soccer went 26.3 to 30.4, lacrosse 28.8 to 32.7, volleyball 15.3 to 17.8.

Net change hides most of the movement, too. Across the balanced panel 1,375 institution-sport offerings disappeared over the decade and 2,611 appeared, a net of plus 1,236 on gross activity 3.2 times that size. Every growing sport got cut somewhere: golf was dropped at 122 institutions and added at 129, softball dropped at 57 and added at 62, football 31 and 43.

That gross-churn number is a sport-level count and should not be read as a count of cuts. It is at the raw survey-code grain, so a college moving from the cross country code to the combined track code shows up in it twice, once as a disappearance and once as an appearance. Collapse the running and aquatics code families and the same panel shows 790 offerings gone and 1,853 gained rather than 1,375 and 2,611. It also counts an offering once whether the college fielded one team in it or two. The companion piece counts the other way, in teams by sex on a stricter panel of colleges that filed all eleven years, and lands on 1,143 drops. Neither number corrects the other; they answer different questions, and the difference between them is mostly the label families.

One note on the last two columns of the table below: roster is athletes divided by teams, and a sport fielded for both sexes at one college counts as two teams there. Both are panel figures, so they run a little apart from the national rosters quoted later.

Wrestling gained 93 institutions in the raw comparison and 104 on the balanced panel, and the fast-moving half of it is women's wrestling. Nationally, women's wrestling went from 19 teams to 179 between AY 2014-15 and AY 2024-25, while the men's side went from 348 to 433. Every team count in this section is a national count rather than a panel one, which is why they run above the panel figures quoted elsewhere here and above the eleven-year-panel figures in the companion post. In athletes the two sides added almost the same number, men plus 3,212 (9,745 to 12,957) and women plus 3,087 (358 to 3,445), which is plus 33.0 percent against a 9.6-fold rise. Women's wrestling has grown in every year since AY 2013-14, including through the pandemic year, when it went from 56 institutions to 66 while the file as a whole lost 229 filers.

Volleyball looks flat and is not. Institutions fielding it went from 1,757 to 1,766. Inside that, men's volleyball went from 170 teams to 300 and doubled its participation, 2,693 to 5,415, while women's volleyball went from 1,756 teams to 1,763. Men's volleyball has climbed for eighteen years, from a low of 130 institutions in AY 2006-07, with the usual pandemic-year break.

Golf's raw loss of 28 institutions is two trends as well. Men's golf teams went from 1,234 to 1,190, down 44, while women's golf went from 889 to 976, up 87.

Two more growth curves are worth knowing, both under their own survey codes. Beach volleyball held at 4 to 9 women's programs through AY 2010-11, jumped to 24 in AY 2011-12 and has grown in every year since except the pandemic year, reaching 202. Women's participation in it went from 36 to 3,128, and nothing in the data explains that jump. On the balanced panel no named sport grew faster in percentage terms, 71 institutions to 197. Bowling comes next, plus 62.7 percent on institutions and plus 79.7 percent on athletes.

And one has stopped. Lacrosse grew from 315 institutions in AY 2004-05 to 595 in AY 2019-20, dipped to 567 in the pandemic year, then 597, 598, 595 and 594. Calling it the fastest-growing college sport describes the last decade, not the current one.

Other Sports is survey code 29 and has been in the collection since AY 2002-03. It was flat for twelve years and then broke. From AY 2004-05 through AY 2015-16 it sat between 45 and 72 institutions, shrinking over the first half of that stretch. Then: 124, 165, 195, 241, 235, 324, 367, 395, 432. No other sport has a series shaped like that.

The growth is not institutional churn. On the balanced panel it goes from 55 institutions to 405, which is 350 of the raw 372.

It is now large enough to matter and small enough to keep in proportion. In AY 2024-25 it is 432 institutions, 704 teams and 16,785 participants, the thirteenth largest sport code in the file by athletes, ahead of swimming and diving combined (16,513) and wrestling (16,402). As a share, it went from 0.39 percent of teams and 0.31 percent of participants in AY 2014-15 to 2.56 and 2.19 percent. Roughly one athlete in 46. It held up through the pandemic year too: institutions dipped from 241 to 235 while participation rose from 7,392 to 7,788, in the year the file lost 229 filers.

One explanation goes first, because it would invalidate everything else. The sport code list did not change. All 37 codes in the sport dimension appear in every one of the 23 years except diving (21 years) and table tennis (22), and nothing was added, retired or renamed between AY 2004-05 and AY 2024-25. No sport discussed here is new because of a vocabulary change.

The composition of the bucket changed in the same year its size did. Coed participation was 4.2 percent of Other Sports in AY 2014-15 and 5.3 percent in AY 2015-16, then 24.8 percent in AY 2016-17 and 45.7 percent in AY 2024-25.

That shift is national. Coed participation across all sports went from 2,527 to 9,989 over the decade, and Other Sports accounts for 76.8 percent of the AY 2024-25 total. Coed participation everywhere else barely moved, 2,443 to 2,322. All growth in coed college athletics sits inside one uncoded bucket.

The sex balance flipped at the same moment. Women were 37.5 percent of Other Sports participants in AY 2014-15, 57.6 percent in AY 2016-17 and 58.3 percent in AY 2024-25, against 42.10 percent of all athletes nationally, a figure that belongs to the earlier post on women's participation.

Split the bucket in half and it is at least two different things. Its coed half is men-majority, 4,469 men against 3,198 women in AY 2024-25. Its single-sex half is heavily women's, 6,592 women against 2,526 men. Rows reporting coed participation inside the bucket went from 5 in AY 2014-15 to 217. The teams are not oversized either: average roster went from 20.2 to 23.8, below the national 27.8, so the growth is new teams rather than bigger ones.

Where those teams sit is the last usable clue. The bucket is concentrated away from the biggest athletics programs, and NAIA institutions alone carry 5,019 of the 16,785 participants, 29.9 percent, from 106 colleges. By sector it is a private-college phenomenon: 30.5 percent of private nonprofit four-year institutions field it against 12.8 percent of public four-year, and 254 of the 432 adopters sit in that sector. The table below lists all nineteen classifications. The small ones are in there too, on bases of nine to forty-seven institutions, so read their shares with that in mind. Read the whole thing as an AY 2024-25 cross-section only, because the classification vocabulary is not stable across the window.

Two readings of that growth were obvious enough to test, and both failed.

The first was reclassification: colleges moving existing teams into a catch-all. If that were happening, named-sport counts would fall at the colleges doing it. They rise instead. The 360 balanced-panel institutions that added Other Sports teams added 1,398 teams net, of which only 583 are Other Sports and 815 are named sports, and they grew at plus 3.88 teams each against plus 0.57 for the 1,550 institutions that added none. Other Sports is additive.

The second was Title IX roster padding: colleges parking women's roster spots somewhere convenient. A bucket that is 58.3 percent women makes that attractive as a story, and it does not survive contact with the data. Institutions that field Other Sports carry a wider proportionality gap than institutions that do not, 13.98 points against 12.77 (women 55.46 percent of undergraduates and 41.48 percent of athletes, against 55.09 and 42.33). The direction is the opposite of what padding would produce. The gap itself is the earlier post's finding and is used here only as a test.

What is left is a shape without a name. The break at AY 2016-17, the coed share going from 4.2 to 45.7 percent, the men-majority coed half and the women-majority single-sex half are all consistent with esports arriving alongside a set of emerging women's sports. They are consistent with other explanations too. EADA has no free-text sport field and the sport dimension carries no description column, so the growth can be described precisely and cannot be identified.

Analysis of the Equity in Athletics Disclosure Act data collection, US Department of Education Office of Postsecondary Education, academic years 2004-05 through 2024-25, at the institution by sport by year grain. Enrollment counts used in the proportionality test are IPEDS undergraduate fall enrollment, carried on the EADA sport record. Data accessed 28 August 2026; see the EADA data source page for what is loaded and when it refreshes. Full query log: docs/data-evidence/which-college-sports-are-growing-and-which-are-vanishing/queries.md, with table and column definitions in schema.md alongside it.

Related reading on this site: the Title IX proportionality gap, measured, why EADA athletics revenue is not revenue, which colleges are cutting teams and what goes first, and the federal data release calendar for when the next EADA vintage lands.