Video game engines as the new “virtual” Skinner box
Game engines are ready-made Skinner boxes that let you test reinforcement ideas on humans with richer stimuli and instant data.
01Research in Context
What this study did
Adriaanse et al. (2026) wrote a think-piece. They asked: what if we swap metal cages for video game engines?
The paper maps how Unity or Unreal can work like a high-tech Skinner box. Rich 3-D worlds replace lights and levers.
Loot boxes, foraging quests, or social avatars become the new reinforcers. Every click, gaze, or step is logged in real time.
What they found
The authors did not run an experiment. Instead they showed that game engines already give us free, ready-made operant chambers.
Built-in physics, random loot tables, and eye-tracking plug-ins let us test reinforcement schedules faster, cheaper, and flashier than ever.
How this fits with other research
Alcañiz et al. (2022) already proved the idea works. They turned a VR room into a diagnostic tool and spotted autism with 86 % accuracy by tracking eye-gaze contingencies.
Miller et al. (2020) and McGonigle et al. (2014) pushed the same tech into therapy. VR airport and job-interview modules shaped real-world travel and conversation skills after only a few short sessions.
Smits-Engelsman et al. (2023) sounds like a contradiction at first. Wii games improved balance inside the screen, but kids could not transfer the skill to real-life balance beams. The gap shows that richer game engines (what Young wants) may need extra real-world bridging tasks to make behavior stick.
Why it matters
If you can program it, you can measure it. Next time you need to test a new reinforcement idea, build a five-minute mini-game instead of renting rats or cages. Run dozens of human participants tonight, tweak the loot table tomorrow, and see the data plot live. Start small: drop a virtual vending machine into your session and let the learner earn tokens for accurate responses. You will get millisecond-perfect data and a client who thinks science is fun.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Open a free Unity tutorial, build a 30-second coin-collect game, and track how many coins your client earns across five trials.
02At a glance
03Original abstract
The Skinner box has provided a standardized method of conducting experiments on operant behavior throughout the history of behavior analysis. As technology has advanced, these operant chambers have become increasingly complex to allow for the study of new stimuli, behavior, and outcomes. The present article takes this variability one step further by advocating for the use of video game engines in the study of operant behavior in humans. Game engines provide high levels of flexibility, control, and realism as well as continuous behavioral tracking, dynamic stimulus presentation, and complex reinforcement schedules that greatly expand the range of research questions that can be addressed. Importantly, the potential for increasing stimulus, response, and outcome variability provides the basis for assessing and maximizing the generalizability of operant and related principles. This article illustrates the use of video game engines to study causal inference, delay discounting, loot boxes, foraging, and multiplayer dynamics. Adopting game engines in behavioral research not only expands the scope of behavior analysis but also increases its relevance to real‐world behavior, offering a promising path forward for innovation.
Journal of the Experimental Analysis of Behavior, 2026 · doi:10.1002/jeab.70081