0

Game accessibility data [WIP]

A frequent ask when working in game accessibility is data. Prevalence data, usage data – this post shares some examples, and some reasons why they should be treated with caution.

Background

If you’ve had any kind of involvement with accessibility you’ve likely come up against questions about data. “But how many people are actually Deaf?”, and so on.

It may seem like a sensible question, but actually it’s worth looking at why that question is being asked. It’s often asked when looking for a way to justify deprioritising something. A more sensible question to ask would be “how many of our players are we comfortable with excluding?”.

As companies and teams and individuals progress along their accessibility journey this thinking tends to fade into the background. Particularly as it moves earlier in development and part of design decisions rather than expensive feature workarounds. Moving more towards accessibility just being part of what day to day development looks like, more towards a success criteria on other work rather than some separate thing battling for priority.

Caution required!

As the old saying goes – ‘lies, damned lies and statistics’. Data can often be extremely misleading.

For example around 2% of the population are deaf, prioritising based on that would mean that many games wouldn’t have subtitles. But 60% of Assassins Creed Origins players turned on the off-by-default subtitles, and 95% of Assassins Creed Odyssey players left the on-by-default subtitles turned on.

Around 1% of the population have one hand, yet Uncharted 4’s one handed functionality was used by 33% of their players.

There isn’t any data available on how many people are physically unable to operate a gyroscope, but it would not be large – yet Into The Dead’s non-gyro alternative input methods were used by 75% of their players.

And 8% of males have some form of colourblindness, but 100% of The Outer Worlds’ players made use of the game’s colourblind accessibility – it’s just how the game is designed, by default.

Research data can be misleading too. Was the demographic representative? Was the same size large enough? Were the questions leading?

And even usage data itself can be misleading. If a feature is used by 2% of players, why is that? Because it’s useful to 2% of players? Or because it should be useful to 10% of players, but it’s poorly executed? Or because it is hard to find? Because the naming and description of the setting itself is unclear?

In all; exercise caution, be careful about what you use, how you use it, and why it is being requested.

But now that those caveats are out of the way, on with the data.


Data

Prevalence

  • Around 30% of players in the USA and UK identify as disabled. This is currently the best source of data as it is a consistent finding across different years.
    [caveat: there are many people who come up against accessibility barriers in games who don’t identify as disabled, reasons ranging from low reading age to colourblindness]
    Newzoo gamer sentiment study

Usage data

  • 35% of Sniper Elite 5’s players increased text size, 95% left subtitles on, 8% enabled aim assist, 92% moderate aim assist sensitivity, 6% aim assist snap, 6% auto perk unlock, 2% auto climb, 12% turned off motion blur.
    Sniper Elite 5 GAAD data

  • 16% of MudRammer’s players are blind. The work took 1 day, the game retailed for $5, meaning instant profit.
    [source no longer available]

  • Around 1% of Solara’s playerbase were blind, this 1% spent far more on microtransactions than anyone else.
    [source no longer available]

Research

1

Comments are closed.