Arcade Driving Game

Training and Generalizability

I love sharing information like this with you. It irks me when I see all of the lofty promises from companies (that shall remain unnamed, but they advertise so much online that I am sure you know who I am referring to) that their brain training will make you smarter, and solve all of your memory limitations, and turn your kids into the next Einstein. I have shared before the results of studies that show how limited transfer of training is when it comes to cognitive processes. If you practice your working memory span, you can indeed increase your working memory span. But it doesn’t improve anything else, and if you stop practicing you lose the benefits.

So here is the new finding that I am eager to share. A team of researchers from the Catholic University of the Sacred Heart in Milan wanted to know if playing car racing video games can make players better drivers. After all, this is a direct link. Practice driving should lead to better driving, right? Especially when they used a virtual reality driving game that had good ecological validity.

One recent study found that older adults could significantly improve their ability to multi-task after playing a specially designed driving video game called NeuroRacer. Another study from researchers at the University of Rochester found that playing action-packed video games improved people’s ability to make quick decisions and ignore distractions.

What this research was interested in most was visual search. To be a good driver you need to maintain situation awareness of your surroundings. You need to check your mirrors, notice signs, and be aware when there are vehicles in your blind spot. This is not all that you need, but it is important. But in a driving video game, players don’t do this. So when they transferred them to real driving, they showed no increase in situation awareness. Experienced driving gamers did not visually explore their environment any more than non-players. Experience driving for real increased this, but not experience driving in games.

My Take

What this shows is that the benefits of training is really specific. If the video game doesn’t get you to look at your mirrors or be aware of your blind spot, then the fact that you might develop better steering control does not transfer to visual exploration.

Even if you were willing to consider the possibility that better steering control would free up cognitive resources that could then be used to maintain visual situation awareness, this study shows that it is not the case. Just having some spare capacity doesn’t do it. You need the training to force the trainee to engage in the visual exploration. Otherwise, there is no benefit.

Your Turn

Do you play any driving games? Does it seem like you improve your real-world driving skills? Would you have guessed that it would have some kind of transfer benefit? Let us know in the comments.

Image Credit: Nemo

5 thoughts on “Training and Generalizability”

  1. As a gamer, this makes sense. The skill set in a driving game is much different than real driving for the reasons you state.

  2. Good post
    So…the fidelity of the simulation is key.
    Having logged quite a few hours in FAA licensed flight simulators I see a good topic for thesis or dissertation research comparing high fidelity simulation, video games (like MS Flight Simulator) and real flying.
    One thing high fidelity simulation allows is a “sweaty palms, close to real” experience in what would be life threatening situations in the real world. It’s hard to practice engine failure on take-off any other way (instructors are understandably reluctant to cut power completely in a training flight).
    While driving is a somewhat complex task, I also wonder how simulator transfer of training fairs (even simpler simulations) for more complex tasks – like flying the space shuttle.
    Finally….I couldn’t agree more with your opening.
    There’s a whole lengthy discussion to be had on the pseudo science of those who shall not be named and their lofty claims for certain products’s ability to improve cognitive function.

  3. I seem to remember that the Royal Navy had just this problem with training SONAR operators (or possibly RADAR operators)about 1942. Eventually they discovered that the simulated images they were using were far too good. Real images were fuzzyand unstable. The real skill was in separating out signal from noise, rather than speed of response.

  4. I wonder how much it has to do the consequences. In racing games, if a person crashes… they just hit reset and drive some more – the goal is to get the best lap time. Even in real lift racing, the costs and risk to life and limb change the equation.

    I’m thinking no matter how good the simulator is… many of the reactions won’t be the same because it is just a simulator.

  5. On a Linked In group where I posted a link to this article, there was a good discussion about fidelity. I think a comment I made there would also be of interest here. There is a strong body of research that the taskflow fidelity is more important for transfer of skill than the sensory experience fidelity. On the other hand, sensory experience fidelity increases perception of immersion and therefore engagement with the simulator.

    So gaming systems, which care more about engagement, should focus on sensory experience fidelity. But training systems, which care more about transfer, should focus on taskflow fidelity.

    The problems arise when designers don’t understand that this is a fundamental difference. They need to start out with a different approach and different architecture long before they get to screen design. Which just brings up the usual HF/E whine that we need to be involved earlier in the design process.

Leave a Reply

Your email address will not be published. Required fields are marked *