Hi Sarang,

hope you are doing well, too.

according to the standards of research, all participants should come from different "classes" of human beings. If you expect results in your research, which could systematically (and theory derived, not significant) not be influenced by {age, gender, experience...} then it is OK to find participants in the best and closest radius (i.e. University students). But if you could not statistically ignore the influence of {...} then you must choose your participants accoring to the society. Then your research has a high ecological validity. But maybe this research is too expensive?

To plan your research please consult cohens d or partial squared eta. Maybe this provides you with an idea of the number of participants and the expected results.

Best regards,

Christopher.

Hi Sarang,

I have ran a few experiments with a driving simulator, and although I don't know of any specific papers which suggest guidelines for participant inclusion criteria - I can give you a few examples of things the research team I worked with the be important co-variates.

Handedness (left or right).

Years driving experience.

Estimated average millage over the last year.

Age.

Gender.

Depending on your study, it might also be important to consider other variables; such as familiarity with automated vehicles. Often you may also wish to issue your participants with questionnaires such as a scale for trust in automation.

I would say that aiming to get a range of ages and gender parity is important, but often studies conducted in universities may be limited in this respect.

One important thing to note when recruiting for these studies, is also "simulator sickness", there is an informal consensus that women (particularly older women) may be more prone to this, so have a procedure in place for withdrawal.

Thank you very much Christopher Frank and Kyle James Harrington for your time.

Thank you very much for your time Professor Joost de Winter