Originally Posted by
Philbert
It does matter to the precision of the findings, and would make the results more valuable from a basic science perspective, but in the real world we would rarely be working with a trainee with the exact sticking point studied. A less narrow study design with more generalizable results would not be substantially more likely to produce a false positive. I agree that it might fail to show a benefit where one existed, but if it requires such precise comparison to detect a result the result is probably not worth pursuing.