Bachelor’s thesis · University of Granada
My final-year project: two versions of the same video game, the usual one and one accessible to people with cognitive disabilities, designed, built and tested with 28 people.
TLDR
Context
My Computer Engineering thesis studies accessibility in video games and puts it into practice: two versions of the same game, one following the industry’s usual patterns and one accessible by design.
Why it matters
The industry is huge, and one in five people who play run into barriers.
Video games are now a place to meet, talk and share. When a game isn’t accessible, it shuts many people out of something that goes beyond playing.
Even so, accessibility tends to arrive at the end of development as an add-on, or not at all.
The approach
With Fundación Purísima Concepción in Granada, part of the Hermanas Hospitalarias network, I focused the work on players with cognitive disabilities, one of the groups the industry addresses least. Their accessibility specialists helped me refine the prototypes during development.
Hypothesis: what helps some, helps everyone.
Solutions designed for cognitive barriers work as enablers for any player.
Cognitive barriers
I started from AccessibleEU’s cognitive accessibility guide. Every design decision in the game answers one of these barriers.
The game
A three-lane endless runner: a gardener moves forward collecting seeds, soil and water to make flowers bloom, while dodging leaves and sticks.
The standard prototype replicates the industry’s usual patterns, with no accessibility measures. The accessible one adds simple, high-impact measures. Both versions share the same mechanics.
| Standard | ✓ | ✓ | ✓ A flower blooms | End | — | — | — |
|---|---|---|---|---|---|---|---|
| Accessible | +1 | +1 | +1 A flower blooms | −25 | +1 | +1 | +1 A flower blooms |
Simple, high-impact measures
Ability
Barrier in endless runners
Accessible prototype’s measure
Ability
Barrier in endless runners
Ability
Barrier in endless runners
Ability
Barrier in endless runners
Accessible prototype’s measure
Other barriers found
Other measures in the prototype
How it’s built
Adding it to a finished game is expensive, limited and rarely fits well. In Hazlo Florecer it’s part of the architecture from the design stage: the standard prototype was made by removing the measures from the accessible one, not the other way round.
Both versions share the same code. An AccessibilityManager holds the settings for each mode (speed, dynamic difficulty, tutorial, typeface and theme) and the rest of the game just reads them. A MetricsManager logs every run to a JSON file for the evaluation.
match current_mode: Mode.STANDARD: settings = { "base_speed": 150.0, "progressive_speed": true, "dynamic_difficulty": false, "show_tutorial": false, } Mode.ACCESSIBLE: settings = { "base_speed": 225.0, "speed_on_collect": 25.0, "speed_on_hit": 25.0, "dynamic_difficulty": true, "show_tutorial": true, }
Simplified excerpt from the repository. The accessible version starts a little faster, but doesn’t speed up on its own over time: its speed goes up 25 when collecting and down 25 on a collision.
The evaluation
We formed two groups of 14, each with 7 participants with cognitive disabilities and 7 without. Each group played one version.
I combined objective measures, which the game logged on its own, with subjective ones: a mental workload survey based on NASA-TLX, adapted with one simple question per dimension and faces instead of numbers as answers.
28 participants in total
With a cognitive disability (strong colour)Without a cognitive disability (light colour)Results
No dimension got worse with the accessibility measures.
The biggest gap is in perceived performance (+0.84 points out of 5): with the accessible version, players felt they had done better. Together with the improvement in “did you have fun?”, that’s exactly what makes someone want to keep playing.
With 14 people per version, this is a clear trend, not a statistical test. I analyzed the results with box plots, for the time and speed of each run, and a radar chart that brings together the average for each dimension.
Every question is phrased positively: a high score is always good, including Effort and Frustration.
Every player is a world of their own.
The objective metrics show every run is different depending on who plays. In the standard version, some last a long time and others lose right away. In the accessible one, dynamic difficulty takes each person to their own challenge, and the experience is just as satisfying for every profile.
Conclusions
Theoretical framework
Disability affects us all
We think of the best-known disabilities, but anyone can face a barrier at some point. Pick a type to see the barriers it creates in a game and the measures that overcome them.
Is it applied?
Sources: Aguado-Delgado et al.’s systematic review (2018) and the industry analysis in my thesis.