> it's not easy to find people with this many hours of practice, hence why they reuse the participants [...]. Since these subjects are very interesting, it makes sense to study them again with different techniques or reproducing previous results.
That's fine as far as it goes. It helps us confirm that a phenomenon can actually occur and perhaps tells us something about its general shape. But when studying a population of known outliers, it's not credible to assume how the results will or won't apply to a broader population. This isn't an abstract epistemological concern; scientific journals are absolutely packed with exciting preliminary findings that vanished into insignificance when somebody did a larger or more rigorous study (but nevertheless might be cherry-picked for a pop-science/self-help book based on the "revolutionary" findings).
That's fine as far as it goes. It helps us confirm that a phenomenon can actually occur and perhaps tells us something about its general shape. But when studying a population of known outliers, it's not credible to assume how the results will or won't apply to a broader population. This isn't an abstract epistemological concern; scientific journals are absolutely packed with exciting preliminary findings that vanished into insignificance when somebody did a larger or more rigorous study (but nevertheless might be cherry-picked for a pop-science/self-help book based on the "revolutionary" findings).