Working out this much seems contrary to good health. Nature may give generously at one crack at this level of intensity in the belief you are trying to survive. But only is only fooled once. Maybe what should be studied more and shared in popular culture is the real health risk of overtraining. Marathons and full triathlons are popular but at some point, probably for psychological reasons, too many people venture way over the line of diminishing returns physically and socially.
This is the within-person version of a problem I keep running into in the data: "responder" looks like a stable trait until you measure the same person twice, and then most of it turns out to be context and noise.
The Odden design you walk through (same people, same block, twice) is the right way to expose it, and the group-stable but individual-unstable split is the whole story.
My bioinformatics read is that a single measurement of a noisy trait is mostly regression to the mean wearing a costume, which is also why one VO2max test, like one ApoB, can mislead you.
The signal lives in the trajectory, the repeated measure over time, not the snapshot. So your "human problem, not a reproducibility problem" framing is right, and I would push it one step further: it is an argument for tracking your own response longitudinally and ignoring whether you were a "responder" on any single block.
Did the paper report within-person test-retest error as a share of the individual variance?
Working out this much seems contrary to good health. Nature may give generously at one crack at this level of intensity in the belief you are trying to survive. But only is only fooled once. Maybe what should be studied more and shared in popular culture is the real health risk of overtraining. Marathons and full triathlons are popular but at some point, probably for psychological reasons, too many people venture way over the line of diminishing returns physically and socially.
Perhaps! Thanks for the comment, Shawn.
This is the within-person version of a problem I keep running into in the data: "responder" looks like a stable trait until you measure the same person twice, and then most of it turns out to be context and noise.
The Odden design you walk through (same people, same block, twice) is the right way to expose it, and the group-stable but individual-unstable split is the whole story.
My bioinformatics read is that a single measurement of a noisy trait is mostly regression to the mean wearing a costume, which is also why one VO2max test, like one ApoB, can mislead you.
The signal lives in the trajectory, the repeated measure over time, not the snapshot. So your "human problem, not a reproducibility problem" framing is right, and I would push it one step further: it is an argument for tracking your own response longitudinally and ignoring whether you were a "responder" on any single block.
Did the paper report within-person test-retest error as a share of the individual variance?