metime Research · After Lexa · November 2026
Fewer queer women die on TV now. More of their shows just end.
The death rate fell from 9.5 per 100 characters on screen in 2016 to about 3 since 2021. In 2023, four in ten were on a show that ended that year.
Death rate, 2016 and 2023: deaths per 100 characters on screen
Ending rate, 2023: on a show that ended that year (11% in 2016)
These are LezWatch.TV's queer female, trans and non-binary characters on US scripted TV. Counting women alone, 36 died in 2016 and 14 in 2023.
metime Research is the data desk of metime, an app that uses AI to write interactive romance fiction for adults.
Death rate and ending rate, US scripted TV, 2010 to 2023
Death rate: deaths that year per 100 characters on screen that year.
Ending rate: the share of characters on screen whose show ended that year, planned finales included.
Source: metime Research analysis of LezWatch.TV data.
Use this chart
LezWatch.TV, whose records we use, wrote in August 2026 that "the death rate hasn't meaningfully improved". For recent years we agree: the rate has sat near 3 since 2021. The fall came between 2016 and 2021.
Key findings
- The death rate fell from 9.5 per 100 characters on screen in 2016 to about 3 since 2021. →
- In 2023, 42% of these characters were on a show that ended that year, up from 11% in 2016. →
- The count did not halve: 127 died in 2010-16 and 153 in 2017-23, with far more on screen. →
- Of ended shows started since 2020, 14% killed one off, against 32% of those started in 2010-15. →
- 82% of those killed off in 2020-23 died in their first year on screen, against 62% in 2010-16. →
What we didWe counted deaths and show endings in LezWatch.TV's volunteer record of queer female, trans and non-binary characters on US scripted TV, 2010 to 2023.
Why nowTen years ago this month, GLAAD reported that more than 25 queer female characters had died on TV in 2016. This is what came after.
The death rate fell by more than half after 2016
From 2010 to 2016, US scripted TV killed off 7.6 of every 100 of these characters on screen each year. From 2017 to 2023 it was 3.9.
Use this chart
The count did not halve. 127 died in 2010-16 and 153 in 2017-23. TV simply had far more of these characters on screen: 3,887 character-years in 2017-23 against 1,673 in 2010-16 (one character-year is one character on screen in one year). At the old rate, about 295 would have died in 2017-23. TV also moved them into deadlier genres such as horror and fantasy; allowing for that, the old rate predicts about 333.
Women alone. Leaving out the trans men and non-binary characters in the record, 127 women died in 2010-16 and 150 in 2017-23. Their death rate fell from 9.9 per 100 on screen in 2016 to 3.5 in 2023.
Deaths we flag. LezWatch.TV tags four of the dead as real people, all on American Horror Story: Elizabeth Short, Aileen Wuornos, Natacha Rambova and Valerie Solanas. We flag five more whose death is the show's starting point or a ghost's backstory: Eleanor Shellstrop (The Good Place), Hannah Baker (13 Reasons Why), Vidalia (Vida), and Flower and Cholera Victim Nancy (Ghosts). That second flag is our own reading, not a LezWatch.TV tag, and may miss some. All nine are included in the figures on this page. Without them, 123 died in 2010-16 and 148 in 2017-23, the 2016 rate is 9.2 and the rate since 2021 is 2.6 to 3.1.
The catchThe fall stopped after 2021: 3.1, 2.8 and 3.1 deaths per 100 on screen in 2021, 2022 and 2023. LezWatch.TV's August 2026 line that "the death rate hasn't meaningfully improved" is right about those years. Its own earlier tables showed the fall after 2016; what we add is the count from 2010 set against the rate TV kept before 2017. We use its figures only through 2023, because newer years are still being entered.
In 2016, TV killed off nearly one in ten of these characters
36 died on US scripted TV in 2016, out of 381 on screen: 9.5 per 100. That is the highest count and the highest rate of any year from 2010 to 2023. All 36 were women.
GLAAD's report that year counted "more than 25" queer female characters killed on TV. The dead included Lexa on The 100, Denise Cloyd on The Walking Dead, Root on Person of Interest and Poussey Washington on Orange Is the New Black. By 2021 the rate had fallen to 3.1 per 100.
One of the 36, Eleanor Shellstrop on The Good Place, dies before the story begins. Without her the count is 35.
The catchYearly numbers are small, so the ranges are wide. 2016's rate (likely between 6.9 and 12.8) overlaps 2015's (7.8).
Fewer shows kill one off. More end on a cliffhanger.
Of shows that started in 2020-25 and have ended, 13.9% killed off at least one of these characters, down from 31.5% of shows that started in 2010-15. Cliffhanger endings rose from 3.0% to 10.1%.
Use this chart
LezWatch.TV's volunteers tag each show with the tropes it uses. "Bury Your Queers" means at least one of these characters was killed off. For recent shows that is 33 of 237. Cliffhanger endings came to 24 shows, and happy endings slipped from 7.4% to 5.1% (12 shows). Batwoman, The Wilds, First Kill, A League of Their Own and Willow all ended on a cliffhanger.
Shows with three or more queer regulars are about four times as likely as shows with one or two to give one a happy ending (26% against 7%), and about twice as likely to kill one (44% against 21%).
The catchRecent shows that have already ended are mostly short ones: 44% of those started in 2020-25 ran one season, against 11% of those started in 2010-15, and a short show has fewer chances to kill anyone. Among shows that ran two seasons or more, the share that killed one off still fell, from 34% (46 of 136) to 16% (20 of 129). Most of the cliffhanger rise is in one-season shows (18 of the 24). Tags are assigned by volunteers and newer shows may not be fully tagged. A tag is per show, so a show with more of these characters has more chances of both a death and a happy ending.
In 2023, four in ten were on a show that ended that year
Of 456 of these characters on US scripted series in 2023, 193 were on a show that ended that year (42%). In 2016 it was 43 of 381, or 11%. We call this the ending rate. Counting women alone gives the same 42%.
The story moved from losing the character to losing the show. The L Word: Generation Q, The Owl House, Riverdale and The Rookie: Feds all ended in 2023.
The catch"Ended" includes planned finales, not only cancellations; the records do not say which. 2023 was also the year of the writers' and actors' strikes. The still-incomplete records put 2024 at 38% and 2025 at 34%.
One in three given a regular role in 2018-21 lasted a single year
Of these characters who became series regulars in 2018-21, 149 of 420 (35.5%) appear in only one year. For those introduced in 2010-13 it was 17 of 113 (15%).
Fewer are killed off. More simply leave the screen early.
The catchLezWatch.TV may not yet have every recent appearance entered, and characters introduced in 2021 have had only two to three years to come back.
Four in five killed off since 2020 died in their first year on screen
Of the 73 of these characters who died on US scripted TV in 2020-23, 60 died in the same year they first appeared (82%). In 2010-16 it was 79 of 127 (62%).
The deaths that remain are more often of characters viewers had little time to know.
The catchPart of this is new and short shows: half of the 2020-23 deaths (37 of 73) came on a show in its first year, against a quarter (33 of 127) in 2010-16. Leaving out anthologies, limited series, The Fall of the House of Usher and The Sandman, the share is 76% (42 of 55), against 59% (67 of 114) in 2010-16. This counts calendar years, not seasons, and the numbers are small. Counting the still-incomplete 2024-26 records too, the share for 2020-26 is 72% (73 of 101).
The CW gives the largest share of these shows a happy ending. NBC's most often last one season.
Among networks with at least 25 ended shows, 2010 to 2023, 7 of The CW's 37 (19%) carry a happy-ending tag. Netflix has the most happy endings in number, 12, only because it has the most shows: 12 of 118 is 10%. 19 of NBC's 45 ended shows (42%) ran one season, against 29% of Netflix's (34 of 118).
US scripted series, 2010-2023. "Characters" counts LezWatch.TV's queer female, trans and non-binary characters. A show on two networks counts for both. Trope columns count ended shows with that volunteer tag; percentages are of ended shows. Tap a heading to sort. Source: metime Research analysis of LezWatch.TV data.
HBO Max gave none of its 23 ended shows a happy ending, and killed few: 3 deaths in 166 character-years. SyFy killed one off in 9 of its 17 ended shows. Peacock killed none in 64 character-years, and 9 of its 20 ended shows (45%) ran one season, the highest share, on a small sample.
FX has the highest raw death rate, mostly from American Horror Story, which killed 15 of its 27 such characters. NBC's comes largely from Law & Order: SVU: 38 characters in 27 years, 10 of them killed.
The catchHBO Max and Peacock are small samples. Happy-ending and one-season counts depend on volunteer tags and on what LezWatch.TV records as one season. One season is not always a cancellation: some are limited series.
Look it up
The 280 deaths on US scripted TV, 2010 to 2023
280 deaths of 277 characters from LezWatch.TV's records, by year; a few died in more than one year. Search a name, a show or a network. Nine are flagged: four real people (LezWatch.TV's own tag) and five whose death is the show's starting point or a ghost's backstory (our reading). They are included in the counts on this page; the last button leaves them out. Three of the dead are not women, and are marked.
From the record
Three odd facts
Sara Lance has died four times: in 2014, twice in 2016 and again in 2021, across Arrow and Legends of Tomorrow. She is one of four characters in the database to die more than once.
LezWatch.TV death dates
The deadliest single day: 12 October 2023, when Netflix released The Fall of the House of Usher and four of its queer women died at once.
Binge release, so all deaths share one date
October is the deadliest month on TV for these characters: 86 recorded deaths, against 36 in February.
Whole database, all years. Death dates are often air or release dates.
Checked and dropped
What did not hold up
Some networks are far deadlier than others.
Raw rates differ a lot, from FX's 14.6 per 100 character-years to Cartoon Network's 0.5. But genre explains it. After adjusting for the mix of shows, no network stands out once you allow for testing many at once.
Casting queer actors keeps characters alive.
The database marks actors who are publicly queer now, not when they took the role. In a spot check of 30, only 12 were out before the role. The link may run the other way.
Open questions
What we could not test
Deaths are climbing again since 2024.
LezWatch.TV's records for 2024 to 2026 are still being filled in. On one measure 2025 looks like 2018, but that figure is provisional, so we leave those years out. We cannot say yet.
She dies in the episode where her romance finally comes together.
A common complaint, and it may be true. The data cannot test it: there is no record of when a relationship starts, or of the exact episode of a death.
The shows that ended were cancelled.
The records mark a show as ended without saying whether it was cancelled or finished as planned, so we cannot split the ending rate that way.
Method
How we counted
"The yearly death rate for queer female, trans and non-binary characters on US scripted TV peaked at 9.5 per 100 on screen in 2016 and has sat near 3 since 2021. The count did not halve, because TV had far more of them on screen: 127 died in 2010-16 and 153 in 2017-23. Meanwhile the ending rate, the share on a show that ended that year, rose from 11% in 2016 to 42% in 2023."
Main source
The analysis was done in-house by the team that owns and runs metime. LezWatch.TV, a volunteer database of 7,467 queer female, trans and non-binary characters, read in November 2026. We used US scripted series (TV shows and mini-series), leaving out anime, web series and TV movies, from 2010 to 2023: 5,560 character-years, where one character-year is one character on screen in one calendar year. The headline says "queer women": counting women alone (cis and trans women) gives the same pattern, and we give those figures beside the main ones. Elsewhere, "these characters" means the whole group.
What we measured
- The death rate: deaths in a year per 100 characters on screen that year. A death counts in the year it happened, for characters on screen that year. A character who died twice in one year counts once.
- The ending rate: the share of characters on screen in a year whose show ended that year. "Ended" includes planned finales.
- The old rate is the 2010-16 death rate, 127 deaths in 1,673 character-years (7.59 per 100), applied to each later year's characters. The adjusted figure (about 333) comes from a model of the 2010-16 deaths by genre and role.
- Endings use LezWatch.TV's volunteer trope tags on ended shows, grouped by the year the show started.
- Network figures are raw. Our check against genre compares each network's deaths with the number its mix of genres would predict.
- Flags: four deaths carry LezWatch.TV's "Based on a Real Person" tag. Five deaths that are a show's starting point or a ghost's backstory are flagged by us by hand. Both are counted in the main figures; recounts without them are in finding 1.
- Corrections to the source: LezWatch.TV spells one name "Aileen Wurnos"; we use Wuornos. Four death dates are written without dashes in the source; we read them as dates.
Credit
LezWatch.TV has tracked this for a decade. Its posts The Relationship of Growth and Death (2020) and The Rise and Fall of Representation (August 2026) published per-character death tables before us. Our figures extend theirs; they are not a correction. GLAAD's Where We Are on TV report for 2016-17 is the source of the "more than 25" count.
Limits
- We use LezWatch.TV figures as levels only through 2023. Later years are still being entered.
- Yearly death counts are small, so single-year rates have wide ranges.
- Trope tags are assigned by volunteers. Newer shows may be under-tagged.
- The records do not separate cancellations from planned finales. Death dates often match air or release dates.
- There is no published death rate for TV characters in general to compare with.
For journalists
Cite, embed, download
Cite this study
Source: metime Research, "After Lexa", November 2026, https://metime.ink/research/after-lexa/
According to an analysis of LezWatch.TV data by metime Research,
Please link to this study page, https://metime.ink/research/after-lexa/, which is safe for work, not to the metime home page. Charts and data are free to use under CC BY 4.0 with credit and that link.
Download the data
- after-lexa-by-year.csv2010-2023, one row per year: characters on screen, deaths, the death rate with 95% range, the ending rate, the old-rate expectation, and the same for women alone and without the nine flagged deaths.
- after-lexa-deaths.csvThe 280 deaths: character, show, network, year, role, date, gender group and our flags, with a link to each LezWatch.TV entry.
Every chart marked "Use this chart" has an embed code and a PNG.
A line for your network
Copy-ready lines
Press: hello@metime.ink
Analysis: metime Research, November 2026.
metime writes private, interactive romance fiction for adults in the US. Visit metime.ink