Prioritising Accessibility Considerations

A common question around accessibility is where to start. What to concentrate on if you're late in development, have limited resources, aren't familiar with the topic. I've often seen developers paralysed by this, wanting to do good work but overwhelmed by the possibilities and worry of getting it wrong. 

There are many ways to prioritise accessibility. How many people a given consideration benefits. How much difference it makes to the experience. How innovative it is. How difficult it is to implement. How well it fits with the unique set of kind of barriers that the game in question presents. 

This post summarises a short piece of work to look at a generalised set of priorities, commissioned for two purposes - use in teaching, and use in development of a set of indie games. The work was in two parts. First, a brief to examine available data on which features disabled gamers say they use, and identify any patterns from that which may indicate which are the most used / most useful. Second, some broader discussion of more general prioritisation issues and methods that sit outside that data. 

Surveys

The first piece of this work was to look at available data. Three sources were used - 

There isn't much else available beyond these three. Other survey-based research does exist, but it covers questions like types of impairment, types of games played and so on, rather than individual accessibility accommodations. 

Limitations

The data from all three sources must be taken with a huge grain of salt. 

In the paper the authors mention likely selection bias, as their survey is not a survey of the general population or representative of it; their sample is self-selecting, described as people who:

Similar applies to the two twitter threads. Small sample size, large selection bias. Someone replying to either twitter thread is someone engaged enough to have seen it in the first place, someone who identifies as having accessibility needs, and someone free enough from stigma to feel comfortable talking about it - people who will not publicly admit to having serious difficulty reading for example, are common.

The nature of the questioning also means that respondents are thinking in general terms about features that are applicable across all games, so there is little representation of game-specific considerations.

But these caveats aside, it is still some data. It just has to be looked at realistically, in the context of its limitations. The top considerations from these sources cannot be taken as a canonical prioritised list; but the considerations that are frequently mentioned can still be taken as considerations that matter to lots of people, and so are worth treating with some degree of priority. 

Breakdowns

https://docs.google.com/spreadsheets/d/1joV5_VbfssdTYD-HfZCABYbGiy3dY3a_kkR5OjmkqgI/edit?usp=sharing

https://docs.google.com/spreadsheets/d/1ovgepgfj39vNsk0jFue9AIxY0Hywu-Z2FagurQDGrN4/edit?usp=sharing 

https://docs.google.com/spreadsheets/d/1-S6Pp0XJl0Ocndzsx3_Iyq-t3a5bAUCWpmKqEWutgcE/edit?usp=sharing 

Survey conclusions

Again while the ordering in those breakdowns can’t be taken as prioritised lists, here are ten considerations that crop up consistently across the three. 

I'm aware that I'm repeating myself here bit it's important to be as clear as possible that caveats apply to this data. They cannot be regarded as the ten most important considerations, they are likely neither the ten with the most reach or the ten with the most impact, either in general or for any specific game. However what they can be taken as is ten considerations that will usually make a significant difference to the experience of a significant number of people.

Some of these are backed up by other sources. Anecdotal social media observations of the most commonly complained about issues being text size followed by remapping, subtitling, and colourblindness. And a publisher’s analysis of the top three issues most commonly complained about through their feedback mechanisms - text size, remapping, and colorblindness. 

Anecdotally, simulation sickness and photosensitivity often result in people regarding the problem being with themselves rather than the design, so it likely under-reports in the data. But its impact is far greater than any other. It can cause discomfort, pain, serious injury or even worse, so it is critically important to consider too.

So for a top 5 it might be worth prioritising these -