[ianai]

They want to measure the benefits of medication.

[garmaine]

That’s missing the point. ADHD meds make everybody concentrate better, regardless of whether the person actually has ADHD.

[ianai]

And even that would give them a measure of the benefits. If it’s the same benefit that actual ADHD receives then it’s still a useful metric. If it’s less than the benefit when used for actual cases then, in aggregate, the benefits would be a low bar. Arguably that’s a better bias as it would make the benefits/costs decision more conservative. They’re taking the difference in measurements for each individual.

[eternauta3k]

I think his point was that the test saturates at a normal person's level of concentration. So medicating normal people does not improve their score.

[ianai]

For ceteris paribus you need data on all variables for all observations. The more variety the better, too.