Home › Players

Gaming Accessibility Statistics 2026: 90+ Sourced Data Points on Which Features Actually Ship

By the 16Best Gaming Data Desk · Published 3 August 2026 · Last reviewed 3 August 2026 · About 32 minutes to read. Every figure below is attributed to a named source and dated as precisely as the original research allows. Where we calculate something ourselves, it is labelled 16Best analysis and the inputs are shown, so you can check the arithmetic. Where a source’s own published percentage does not reconcile with its own counts, we say so rather than quietly picking one.

Four of the 30 most-played games on Steam tell you, inside the game, which accessibility options they have. That is last of 24 audited features on full implementation, and it is the cheapest item on the entire list to build. Accessibility ships in proportion to how much of the work a standard game build already does for you — not in proportion to cost, and not in proportion to need. Features that arrive with the engine average 91% adoption. Features that need somebody to stand up in a planning meeting and ask for them average 30%. The item ranked second-from-last, a toggle for haptics, is also close to free. Whatever is short here, it is not budget.

Which accessibility features actually ship, ranked?

Ranked by independent audit, the best-supported feature is universal and the worst is nearly absent: 100% of 30 top-played Steam games let you navigate the whole UI with your gameplay input device, while 13% told you in-game what accessibility options they had. Everything interesting on this page lives in the shape of that fall.

The dataset is the only one of its kind. Nico Trick and colleagues installed and played 30 of the top 50 most-concurrently-played games on Steam, spent roughly two hours in each, and scored 24 “basic” items from the Game Accessibility Guidelines on a four-point scale, with a paid independent rater re-coding 10 of the 30 games (Cohen’s weighted kappa 0.78, linear weighting). Sample date: 29 June 2022. Published online May 2025 in the International Journal of Human–Computer Interaction, vol. 42 no. 1.

Read the limits before the numbers, because they are load-bearing. The sample is PC versions only, tested on keyboard and mouse alone; VR titles and VR-specific guidelines were excluded; only games released from March 2017 onward were eligible; and popularity was defined by concurrent players on a single day, so the sample skews toward live-service and multiplayer titles. The 24 items are the guidelines’ basic tier — the authors note there are more than 92 further features at the intermediate and advanced levels, which commercial games implement far less often. It is a four-year-old snapshot of thirty games. It is also the most rigorous per-feature evidence that exists, because everything more recent is a publisher ticking its own box.

Denominators vary because the study assigned “not applicable” where a rating made no sense — a game with no spoken dialogue was not penalised for lacking subtitles. Which metric matters: this table ranks by share of applicable games rated fully implemented. The study itself flags the lowest-scoring feature on its four-point mean score, and on that measure the bottom two swap: the haptics toggle averages 1.7 and in-game accessibility documentation 1.8, because more games did something partial about documentation than about haptics. Either way both sit at the bottom, and both are cheap. Source: Trick, Phillips, Gutwin & Bradford, Content Analysis of Basic Accessibility Features in 30 Top Played Video Games, International Journal of Human–Computer Interaction 42(1), 425–440; underlying thesis, University of Saskatchewan, December 2024. The 24 percentages are 16Best calculations from the study’s published per-feature counts — the study publishes counts and category means, not per-feature percentages.

Does implementation cost explain the ranking?

No — and we tested that first, because it is the explanation everyone reaches for. The intuitive story is that cheap text-level options (subtitles, font size, colourblind palettes) are everywhere while expensive engineering (remapping, assist modes, screen readers) is rare. Line the audit up against that hypothesis and it breaks in three places.

Break one: full control remapping ranks 6th at 87%. It is one of the more involved things on the list to retrofit, and it is nearly universal — because competitive PC players demanded it long before accessibility advocates did. Break two: the cheapest item in the entire guideline set ranks last on full implementation. Writing down your accessibility options on a menu screen costs a designer an afternoon. Four games did it. Meanwhile “wide choice of difficulty levels” — which costs balancing work, extra QA passes and design argument — ranks 20th at 40%, comfortably above three cheaper items beneath it.

Break three is the study’s own, and it is the one that should end the argument. The authors split the 30 games by developer headcount — seven small studios (1–20 staff), seven mid-size (21–50) and sixteen large (50+) — and compared mean implementation scores. Small studios scored 3.0 out of 4. Mid-size scored 3.2. Large studios, with hundreds of times the budget, also scored 3.2. A one-way ANOVA found no statistically significant difference by studio size (p = .056). If accessibility were bought, that gap would be enormous. It is 0.2 of a point, and it does not clear significance.

What does order the ranking is who does the work. Group the same 24 features by how much of each one a standard, competently-built PC game produces without anyone thinking about disability, and the ranking snaps into place.

16Best analysis. The three-tier grouping is ours and is a judgement call; the adoption scores inside it are unmodified from Trick et al. Tier means are unweighted averages of the per-feature percentages in the ranking table above, so each feature counts once regardless of how many games it applied to. A reader who moves two or three features between tiers will narrow the gap; they will not close it, because the top and bottom of the ranking are not in dispute.

The cleanest demonstration sits inside a single category of the guidelines. The study’s “General” category holds exactly three items: remembering your settings between sessions, offering a wide choice of difficulty levels, and documenting your accessibility options in-game. They cost roughly nothing, they serve every impairment type, and they are graded on the same scale by the same rater in the same two-hour session. They score 93%, 40% and 13% — a 7.0× spread from top to bottom of one category (16Best analysis, from 28/30 and 4/30 in Trick et al.). The one at 93% is the one a save system produces on its own. The one at 13% is the one somebody has to decide to write.

That is the spine of this page, and it reframes the whole conversation. Accessibility in games is not primarily a budget problem. It is an agenda problem: the features that show up are the ones nobody had to put on a roadmap. Everything that required a person to stand up in a planning meeting and ask for it lands at roughly a third.

Which accessibility categories score best and worst?

Speech accommodations score highest (97% fully implemented) and general accessibility practice scores lowest (49%) — and the top of that list is largely an accident. The speech category contains exactly one item: that voice input must not be mandatory. Twenty-nine of thirty games passed, but as the study itself points out, that is mostly because voice control is rare in games, not because developers designed around it. The one failure, Phasmophobia, is built on talking to a ghost; it later added a text alternative.

Both percentage columns are shares of applicable coded cells (680 of the 720 game×feature slots; 40 were rated not-applicable). Mean score runs from 1 (not implemented) to 4 (fully implemented). Two notes on provenance. First, the study reports its headline both ways — 63.7% fully implemented and 16.6% not implemented against applicable cells, or 60.1% and 15.7% against all 720 including not-applicable. Mixing the two bases is the easiest error to make with this dataset, so we use the applicable base throughout. Second, the published General-category row gives 39% not-implemented, but its own counts (28 of 90) give 31%; we use the counts. Category totals recomputed by 16Best from the per-category counts in Trick et al.; fully-implemented shares match the study’s published values exactly.

Game-level scores tell the same story about variance: the highest-scoring title in the sample, Age of Empires II: Definitive Edition, averaged 3.7 of 4; the lowest, Football Manager 22, averaged 1.9. Both were among the most-played games on the platform on the same day.

How many players have a disability?

Nobody has counted. Every number you have seen — 400 million, one in five, take your pick — is a general-population disability rate multiplied by a player base, and the rates themselves disagree by more than a factor of two. This is worth stating plainly, because the accessibility conversation is usually anchored to a confident-sounding headcount that no survey ever produced.

Start with how unstable the underlying rate is. Two United States federal surveys, same country, same adult population, return 28.7% (CDC Behavioral Risk Factor Surveillance System, 400,000-plus adults a year, telephone only, 44.7% response rate in 2023) and 13.6% (US Census American Community Survey, 2023 data, about 2 million households and 4.7 million people, mail, web and phone, 84.7% response rate). The Northeast ADA Center attributes the gap to question wording, question placement, collection mode and priming — BRFSS is a health survey, so respondents arrive already thinking about limitation. Crucially, the gap does not close when you try to fix it: matching the ACS to BRFSS’s demographic scope lifts the ACS to roughly 16% — which closes about a sixth of the 15.1-point gap and leaves the rest to wording, mode and priming (16Best analysis on the Northeast ADA Center’s figures).

16Best analysis: implied-player figures are each published rate applied to the 3.6 billion global player base used across our gamer demographics statistics. This is deliberately crude — a US adult rate should not be applied to a global, all-ages player base, and a 2008 casual-games rate should not be applied to anything — which is precisely why the spread, not any single row, is the finding. The Microsoft figure is not derived; it is quoted. Provenance flag: our source for the BRFSS/ACS comparison, the Northeast ADA Center, prints the 28.7% figure and a 2023 response rate but does not state a collection year alongside it, so we do not assign one; the ACS 13.6% is explicitly 2023 data. Sources listed in full below.

Only one row in that table ever asked players anything. PopCap Games commissioned Information Solutions Group to run an international online survey of 13,296 casual-game players between 2 and 17 April 2008; 2,728 of them — 20.5% — self-identified as having a physical, mental or developmental disability, with more than three quarters describing it as moderate or severe. The press release compared that to 15.1% of the American population, which is itself a scope error: an international sample benchmarked against one country’s rate. This is the origin of the “one in five gamers is disabled” line still circulating in 2026 — a self-selected online sample, casual-game players only, eighteen years old. It is not a measurement of today’s player base and should never be presented as one. We apply the same instrument discipline here that we apply in our video game addiction statistics: a prevalence number tells you which survey was run before it tells you anything about people.

Where that leaves the honest answer: nobody knows, the plausible range is wide, and the useful move is not to pick a number but to notice that no single feature on the audit ranking above serves a fixed slice of that population anyway. Subtitles are used by deaf and hard-of-hearing players, by people in shared rooms, by non-native speakers and by anyone playing with the sound off. Remapping is used by players with limited hand function and by competitive players chasing an edge. Sizing the “accessibility market” assumes a boundary the usage data does not support.

What do players with disabilities actually use?

Software options, overwhelmingly — 68% of surveyed players with disabilities used no assistive hardware at all. In a survey of 230 players with disabilities run by researchers at the University of York with the AbleGamers charity (published 2018), 156 respondents reported using no assistive technology; 77 reported using no accessibility features; 47 used neither. The two most-used accommodations were both in-game settings.

Beeston, Power, Cairns & Barlet, Characteristics and Motivations of Players with Disabilities in Digital Games, University of York and The AbleGamers Charity, 2018. Respondents could select multiple items; the ten shown here are a selection from a longer list that also included mouse-cursor enlargement (24), text-to-speech (14), speech-to-text (17), alternative controllers (10) and VR headsets (3). Shares are 16Best calculations from the published counts over the full 230 respondents. Provenance warning: this is a recruited advocacy-charity sample, not a representative one — 128 of 230 reported upper-limb and 120 lower-limb physical disabilities, so motor needs are over-represented relative to the population of people with disabilities as a whole. Average respondent age 31; 150 described themselves as gamers and 138 said gaming was their primary hobby.

Which storefront labels accessibility best?

PlayStation offers the most tags, Steam publishes by far the largest count of tagged listings, and none of the four schemes verifies a single claim. Ranked by breadth of the labelling scheme — noting that Sony and the ESA do not publish how many titles have actually been tagged, so “most tagged” is a statement about disclosure as much as about volume:

Sources: PlayStation Blog, 3 April 2023 (50+ tags across six categories: visual, audio, control, gameplay, online communication, subtitles and captions); ESA press release, 20 March 2025 (founding members Electronic Arts, Google, Microsoft, Nintendo of America and Ubisoft, later joined by Amazon Games, Riot Games, Square Enix and Warner Bros. Games); Xbox Wire, 16 November 2021; Access-Ability, 20 June 2025 (16 tags across gameplay, visual, audio and input, all exposed in Steam’s “narrow by accessibility feature” filter) and Can I Play That?, 26 January 2026 (three tags added, one renamed). Tag counts are not comparable like-for-like — the schemes slice the same underlying features at different granularity, so a bigger number is a finer grid, not more accessibility.

Volume is where Steam wins, and the growth is real — but the two published figures are not the same object, so read the arithmetic with that in mind.

Sources: Valve, reported at launch June 2025; Can I Play That?, 26 January 2026. Both figures are Valve-reported and rounded to the nearest thousand, and the two use different baskets, so treat the growth rate as an upper bound rather than a measurement. Steam’s catalogue passed roughly 125,000 released titles by 2026 (SteamDB cumulative releases), which puts declared coverage at under one in five — again a ceiling, not a floor, because the 22,000 includes non-game apps and the 125,000 does not.

What has adaptive hardware actually delivered?

It has delivered a cheaper entry point and a great deal of coverage, for a category that roughly a tenth of surveyed players with disabilities actually use. The headline devices, with published US launch prices:

Sources: Xbox Wire (4 September 2018; 16 November 2021), PlayStation Blog (13 July 2023; 3 April 2023), Naughty Dog / PlayStation Blog (9 June 2020), Can I Play That? (10 March 2025; 18 March 2025; 26 January 2026), ESA (20 March 2025), Directive (EU) 2019/882.

Does any law require games to be accessible?

No law in the US or EU currently requires the gameplay of a video game to be accessible. Both regimes regulate the perimeter — the chat window and the checkout. Getting this wrong is the most common error in accessibility write-ups, so here is the precise scope of each.

United States — the CVAA. The Twenty-First Century Communications and Video Accessibility Act became law in 2010. Video game software held an FCC class waiver from its advanced communications services obligations from 2012, renewed twice; the FCC granted a final extension on the stated understanding that it would be the last, and the waiver expired on 31 December 2018. From 1 January 2019, games offering advanced communications services — in-game voice chat, text chat, messaging — must make those communication features accessible. Two limits matter. The gameplay around the chat window is untouched by the statute. And games released before the waiver lapsed remain covered by it even while they continue to sell, so a large share of the existing catalogue is permanently grandfathered; new releases, and in some cases substantial updates to grandfathered games, are not.

European Union — the European Accessibility Act. Directive (EU) 2019/882 applies to products and services placed on the market from 28 June 2025. Video games are not named anywhere in it, and they are not exempted either — they are caught, if at all, through two side doors. The first is real-time communication: a game offering voice or text chat has to make that chat accessible, which is the same perimeter the CVAA polices. The second is e-commerce, and this is where most write-ups overreach. Selling through a website or a mobile app is clearly in scope. Whether a storefront inside a PC or console game — a battle pass screen, a currency shop — counts as an e-commerce service under the directive has not been settled by guidance or by a court, and practitioner analyses published around the applicability date say so plainly. Anyone telling you your in-game shop is definitely covered, or definitely not, is ahead of the law. Microenterprises providing services are exempt: fewer than 10 people, plus annual turnover or balance-sheet total no higher than €2 million. That exemption covers services, not products. Service contracts concluded before 28 June 2025 may run on unchanged until they expire, to a maximum of five further years.

Is there evidence that accessibility sells?

There is good evidence that accessibility features get used, decent evidence that they affect purchase intent, and almost no published evidence tying them to revenue. We went looking for a clean sales-lift number and it does not exist in public. Saying so is more useful than inventing one.

The strongest measured evidence is engagement telemetry. Ubisoft’s David Tisserand posted in 2019 that across a sample of Ubisoft titles, 90–95% of players kept subtitles on where they were on by default, against 50–60% who switched them on where they were off by default; a 2018 figure from the same source put Assassin’s Creed Origins at about 60% turning them on. Note what that actually measures. It is a default-setting effect and a statement about all players, not about disabled ones, and it is the most-abused pair of numbers in this field. If you see it quoted as evidence of a disabled-player market, the quote has changed the claim. Be equally wary of the “95% of players use subtitles” and “97%” figures that circulate without attribution — we could not trace either to a primary source and have not used them.

On purchase behaviour, Scope’s December 2020 survey of 1,326 UK gamers (812 disabled, 514 non-disabled, all playing at least monthly) is the most usable dataset:

Scope, Accessibility in gaming, fieldwork December 2020, UK only, n = 1,326. Self-reported. Scope also values the UK “purple pound” — total spending power of disabled households — at £274 billion.

Separately, the ESA and Video Games Europe’s Global Power of Play report (October 2025, 24,216 players aged 16 and over across 21 countries, nationally representative samples) found 78% of players agree video games create accessible experiences for people with different abilities. Hold that against the audit: 78% of players believe games are doing this, while the only independent per-feature check found the median audited feature shipping in about two games in three and a third of feature slots partial or absent. That is an attitude measurement, commissioned by the industry it measures, not an accessibility audit — and it is measuring what players assume, not what they have tested.

Why do gaming accessibility statistics disagree?

Because six different things get called “accessibility data” and none of them is measuring the same object. If you take one thing from this page for citation purposes, take this table.

The last row is the one that quietly ruins most comparisons. The audit used a four-point scale precisely because “has subtitles” is not a yes-or-no fact: a game can display subtitles in six-point white text over a bright background and truthfully tick the box. When you see a claim that some percentage of games are “accessible”, ask which threshold, which sample, and who checked. Almost always the answer is: no threshold, self-declared, nobody.

Who funds the numbers in this field?

Overwhelmingly, the companies being measured — and the direction of the bias is consistent. This is worth a section of its own, because the sourcing ecosystem here is unusually lopsided.

• Platform holders supply the market-size figure. The “400 million players with disabilities” number that anchors most accessibility pitches comes from Microsoft, which sells the Xbox Adaptive Controller. To Microsoft’s credit, that estimate is conservative relative to every survey rate — but no methodology has been published for it.
• Publishers supply the feature data. All 22,000+ Steam declarations, all PlayStation and Xbox tags, and every Accessible Games Initiative label are self-certified by the company selling the game. There is no audit function anywhere in the chain.
• Hardware vendors and consultancies supply most of the “accessibility is a market” research. Samsung UK’s 500-respondent study (OnePoll, published October 2023) is a genuine primary survey — and Samsung sells screens and phones.
• Trade associations supply the sentiment. The 78% “games create accessible experiences” figure comes from a report published by the ESA and Video Games Europe — the industry’s own trade bodies — and it measures agreement with a statement, not a tested game. It is a well-run, nationally representative survey. It is also the single most quotable positive number in this field, and its publisher has an interest in it.
• Charities and academics supply the barrier data. Scope, AbleGamers and university HCI groups produce the least commercially motivated numbers on this page. They are also the smallest samples, the oldest fieldwork and the most self-selected recruitment.

The pattern that leaves: supply-side numbers are large, self-reported and flattering; demand-side numbers are small, independent and unflattering. When those two collide — 22,000 tagged Steam listings against 4 of 30 games documenting their own options — trust the one where somebody actually played the game.

That cuts at us too, so here is our own exposure. This page leans hard on one academic study of thirty PC games from a single day in June 2022. It has no commercial sponsor, it published its counts, and a second rater re-coded a third of it — which is why we use it. But it is thirty games, four years old, keyboard-and-mouse only, and its authors would be the first to say a sample that size cannot speak for a 125,000-title catalogue. If somebody runs it again on 300 games across console and mobile, we will rewrite this page around their numbers, and some of what is above will not survive.

Key takeaways

• Accessibility ships by agenda, not by budget. Features a standard build produces anyway average 91% adoption; features needing a deliberate accessibility decision average 30% — a 3.0× gap across the same 24 audited items.
• The two cheapest things on the list finish last. Four of 30 top-played games documented their accessibility options in-game; seven offered a haptics toggle. Both cost almost nothing. The cost hypothesis dies here.
• The study tested budget directly and found nothing. Small studios averaged 3.0 of 4, mid-size 3.2, large 3.2 — not a statistically significant difference (p = .056). Within one guideline category, the three items score 93%, 40% and 13%: a 7.0× spread with no cost difference to explain it.
• 63.7% of applicable basic features were fully implemented across the audit, 16.6% were not implemented at all, and a further 19.7% only partially — so more than a third of the sample falls somewhere a binary store tag cannot express.
• The headcount of players with disabilities is modelled, not measured. Published rates imply 490 million to 1.03 billion of 3.6 billion players; the only player-side survey is a self-selected casual-games poll from April 2008.
• Players reach for settings, not devices. 68% of surveyed players with disabilities used no assistive hardware; key remapping (51%) and subtitles (47%) led. Nothing on the hardware timeline changes that ratio.
• Storefront tags are discovery, not measurement. Steam went from 5,000+ games to 22,000+ apps and games in about seven months — different baskets, so treat the growth as an upper bound. None of the four schemes verifies a claim.
• No law requires a game to be playable. The CVAA covers in-game communications (from 1 January 2019, with the pre-2019 catalogue grandfathered); the European Accessibility Act (from 28 June 2025) reaches games through real-time chat and e-commerce, and whether an in-game shop counts is unsettled.
• The commercial case is thinner than advocates claim and stronger than publishers assume: 63% of disabled respondents bought in-game add-ons versus 50% of non-disabled ones, but no public study ties accessibility features to a sales lift. We looked; there isn’t one.

So the spine, proven three ways. Cost does not order the ranking: the cheapest item finishes last and one of the most expensive finishes sixth. Budget does not order the studios: small teams score 3.0 and large teams 3.2, a gap that fails significance. And inside a single guideline category, three items that cost roughly the same land 7× apart, separated by nothing except whether a standard build produces them on its own. What decides whether an accessibility feature ships is not what it costs. It is whether anyone had to decide to build it.

For the wider context these numbers sit in — player counts, ages and platforms — see our gamer demographics statistics, video game industry statistics, console market share and average time spent gaming.

Frequently asked questions

What percentage of gamers have a disability?

Nobody has measured it on a current, representative player sample. The only player-side survey is from April 2008, when Information Solutions Group found that 2,728 of 13,296 casual-game players — 20.5% — self-identified as having a physical, mental or developmental disability. Everything else applies a general-population rate to players: 13.6% (US Census ACS, 2023 data), 16% (WHO, 2023) or 28.7% (CDC BRFSS). Applied to 3.6 billion players those rates imply roughly 490 million to 1.03 billion people, which is a modelled range rather than a count.

Which accessibility features are most common in games?

In an independent audit of 30 top-played Steam games, the best-supported features were: all UI reachable with the gameplay input device (100% of games), speech input not required (97%), settings remembered between sessions (93%), separate volume controls (93%), appropriate default field of view (91%) and remappable controls (87%). The rarest were in-game accessibility documentation (13%), a haptics toggle (23%) and simple clear formatting (27%).

Are games legally required to be accessible?

The gameplay itself is not, in either the US or the EU. In the US, the CVAA requires that advanced communications services inside games, such as voice and text chat, be accessible; the FCC class waiver for game software expired on 31 December 2018, and games released before then remain covered by it even while they continue to sell. In the EU, the European Accessibility Act has applied since 28 June 2025 and catches games through real-time communication and e-commerce provisions rather than gameplay. Whether a storefront inside a game counts as e-commerce under the directive is not yet settled. Microenterprise service providers are exempt: fewer than 10 people, plus turnover or balance-sheet total no higher than 2 million euro.

How many games on Steam list accessibility features?

More than 22,000 apps and games had completed Steam's accessibility questionnaire as of January 2026, up from more than 5,000 games when the feature launched in June 2025. The two figures use different baskets, so the apparent growth is an upper bound. Against a catalogue of roughly 125,000 released titles that is under one in five, and because the 22,000 includes non-game apps the true games-only share is lower still. All declarations are self-reported and none are independently verified.

Do players with disabilities need special controllers?

Most surveyed players did not use one. In a 2018 survey of 230 players with disabilities run by the University of York and the AbleGamers charity, 156 respondents (68%) used no assistive hardware at all. The most-used accommodations were in-game settings: key remapping (51% of respondents) and subtitles (47%). Customised controllers were used by 11% and eye-gaze tracking by 2%.

Does adding accessibility features increase sales?

No published study establishes a sales lift. The available evidence is adjacent: Ubisoft reported in 2019 that 90 to 95% of players kept subtitles on where they were on by default, against 50 to 60% who enabled them where they were off. Scope's 2020 UK survey found 50% of disabled gamers say accessibility information affects their purchase decision and 63% bought in-game add-ons or subscriptions, against 50% of non-disabled gamers.

Why do gaming accessibility statistics disagree so much?

Four reasons. Disability prevalence instruments differ by more than two times on the same population. Rates measured on all adults get applied to players without adjustment. Feature audits cover a few dozen titles while storefront tags cover tens of thousands of self-declared listings. And there is no agreed threshold for calling a game accessible, so a partial implementation and a thorough one produce the same tick in a tag list.

Sources

1. Trick, N., Phillips, C., Gutwin, C. & Bradford, L. — Content Analysis of Basic Accessibility Features in 30 Top Played Video Games, International Journal of Human–Computer Interaction 42(1), 425–440 (2026). Peer-reviewed independent audit; sample drawn 29 June 2022.
2. Trick, N. — Accessibility in Video Game Design, MSc thesis, University of Saskatchewan, December 2024. Full per-feature tables.
3. Beeston, J., Power, C., Cairns, P. & Barlet, M. — Characteristics and Motivations of Players with Disabilities in Digital Games, University of York and The AbleGamers Charity, 2018. Primary survey, n = 230.
4. Scope — Accessibility in gaming, UK survey, fieldwork December 2020, n = 1,326. Primary survey by a disability charity.
5. Samsung UK / OnePoll — Research on gaming’s inclusivity gap, published 4 October 2023, n = 500. Vendor-commissioned survey.
6. PopCap Games / Information Solutions Group — casual games and disability survey, fieldwork 2–17 April 2008, n = 13,296, as reported by the IGDA Game Accessibility SIG. Self-selected sample.
7. World Health Organization — Disability fact sheet, 7 March 2023. Modelled global estimate.
8. Northeast ADA Center — How common is disability among adults in the US? Understanding survey differences. Secondary analysis comparing CDC BRFSS (28.7%) with US Census ACS 2023 data (13.6%), including response rates and the scope-matched ~16% figure.
9. Xbox Wire — Xbox Adaptive Controller launch (4 September 2018) and Game Accessibility Feature Tags (16 November 2021); Disability Pride (5 July 2022, the 400 million estimate).
10. PlayStation Blog — Accessibility Tags on PlayStation Store (3 April 2023) and Access controller launch (13 July 2023).
11. Naughty Dog / PlayStation Blog — The Last of Us Part II accessibility features detailed, 9 June 2020.
12. Entertainment Software Association — Accessible Games Initiative, 20 March 2025; and Power of Play: 2025 Global Video Games Report, October 2025, n = 24,216 across 21 countries.
13. Valve — Steamworks accessibility features documentation; and Can I Play That?, Steam updates accessibility tags, 26 January 2026 (22,000+ apps and games; three tags added, one renamed). Provenance flag: the 5,000+ figure for June 2025 is Valve’s own, quoted in contemporaneous trade coverage of the launch; we could not locate a permanent Valve-hosted page stating it, which is one reason we treat the growth rate as an upper bound rather than a measurement.
14. Access-Ability (Laura Dale) — Accessibility Tags are Now Live on Steam, 20 June 2025; and Not All Subtitle Users are Deaf, 25 April 2022 (Ubisoft subtitle telemetry, David Tisserand, June 2019).
15. Can I Play That? — Microsoft releases Xbox Adaptive Joystick, 18 March 2025; and AbleToPlay open beta, 10 March 2025.
16. Perkins Coie — New Accessibility Requirements in Effect for Video Game Software (CVAA waiver expiry, 31 December 2018).
17. Taylor Wessing — Accessibility in the Gaming Industry: Implications of the European Accessibility Act, March 2025; and Player Research, European Accessibility Act and Video Games, June 2025.
18. SteamDB — Steam game release summary by year (catalogue size).