Given that you want to assess a real-life scenario, it might make sense to actually simulate one. You could look at virtual reality scenarios to place your participants on either a crowded or a half-empty bus, train or subway. In terms of visual attention, you could use two scenarios in which the other passengers are either quiet and non-engaging vs. a scenario where they are a bit more active and visually engaging for the participant to draw their visual attention.

That would leave a few questions:

- is such scenario already available? I'm not aware of any, so it might have to be created.

- how to implement traditional neuropsych tests in such scenario? Maybe look at publications by Thomas Parsons, he has done so with various real-world scenarios.

- do you have to use traditional neuropsych tests? Not really, as you could also implement more naturalistic (ecologically relevant) assessments of attention (or whatever you intend to measure). However, you might want to correlate whatever you measure in the scenario with traditional (validated) neuropsych measures since that's what this field still fixates on.

Let me know if you have any questions. Best of luck.

Sebastian Koenig Thanks a lot for your answer. And pointing out Prof. Thomas Parsons's research. But for our application, we have to use real life scenarios, since we are going to test some augmented reality app during commuting (so we can not use virtual reality), by controlling the tiredness, visual attention and crowdedness. Actually we need to control levels of each of these independent variables to validate our application.

Do you have any suggestions?

Alright, that makes things more complicated.

Let's go through the independent variables one by one:

Tiredness: This one is easy - instruct your participants to get a set amount of sleep prior to the experiment.

Crowdedness: Select an empty or mostly empty train/bus/subway line and fill the vehicle with confidants (e.g. research assistants, volunteer students, etc.). This will also solve your visual attention problem. You simply give them instructions to act out various interactions that correspond to a higher visual attention load vs. lower load. The volunteers will not make this as realistic and ecologically valid as normal commuters, but at least you have more control this way.

Thanks Sebastian Koenig . Since our design is with-in subject fully-crossed, we are thinking to use Stroop Effect (for visual attention) , and do n-back test (to induce tiredness) for fixed interval. Do you think it's reasonable?

To be honest, I am not sure.

Do you mean that your participants have to solve a stroop task to put demands on their visual attention?

The tiredness/fatigue induced by completing an n-back is probably different compared to not getting enough sleep. It all depends on what kind of hypotheses you have and how you plan on generalizing your results to the real world.

Hi Sebastian Koenig , The experiment is see how two augmented reality applications differ each other (i.e. interaction effects) when the tiredness, visual attention and crowdedness increases. So we need to manually induce these factors in different proportions to verify the interaction effects.

Like you mentioned, we need to induce tiredness which occur in commuting scenarios due to long travel duration. Similarly we need to induce crowdedness and visual attention which occur in commuting scenarios due to passengers behavior. So they will do the tests before hand (during commuting) and once we are confident that they have passed some threshold we test our two applications on same participant order counter balanced.