Please, for all of us, make sure your test set contains a wide diversity of people - doing this testing on recruits from a local college or among friends would make it very easy to bias your results for accuracy on particular populations. it's critical that your device be equally accurate on tall people, short people, white people, black people, men, women, thin, fat, and people with various illnesses or other disorders that may affect either THC metabolism or produce abnormal breath results (e.g., excess ketones in the breath for a diabetic). and don't just listen to some random person on the internet for this, find a pro.
Excellent point. This is something we have talked about internally a lot. Early on, we are really testing whichever people we can get but when we move into a full clinical trial we are definitely going to design the study to be as diverse as possible. That is why we reached out to Curebase, to make sure the professionals were involved.
