Weekend ‘catch-up sleep’ is a lie

The file-drawer problem with small lab studies is serious, and can't be solved with effect sizes; it's much easier to run a few of these and publish the hits than to do that with large-population studies.

But small lab studies provide a lot of value alongside population studies. They help reveal issues with bad self-report data, which are major concerns for sleep and weight questions. They also offer an efficient way to replicate and extend large studies. If you're spending a decade with 10k+ subjects, gathering a huge range of data is understandable (and I would argue, preferable). But it leaves you with a choice between salami slicing across all your variables, or applying multiple-comparisons corrections that destroy your sensitivity. Using large multi-factor studies as exploratory work, then validating effects with small focused studies is actually a pretty healthy outcome.

A final thought: I have a lot of problems with this study, but file-drawer issues aren't one of them. Looking at a bunch of correlated variables (insulin sensitivity, weight gain, outside-meal eating, etc) and reporting them all is a solid way to mitigate that - we know those things should correlate for basically any real effect, which makes it much harder to stumble into a false positive and publish.

This is the issue, and it's why the person who titled this article ought to be ashamed.

We know sleep-deprived people sleep extra if you let them, in amounts predicted by their sleep deprivation. We know the recovery sleep produces better results on a bunch of metrics than getting ~8 hours without extra recovery time. We also know that they don't make up lost sleep hours one-to-one; they add a few hours over a few days then return to baseline, and the fraction recovered declines as the amount of deprivation rises. So.. what does all of that mean?

I know my athletic performance responds to recovery sleep; I've seen studies saying cognitive performance does too. Does insulin sensitivity recover fully? (The author's past work finds "sorta".) Does that recovery provide a better baseline when deprivation starts again, or does it just shift to a new target level regardless? (This implies it just moves to the lower level.) How about the same two questions for memory? For muscle and aerobic exercise? And then: does time with recovered functioning correct long-term damage, mitigate it, or just stop worsening it? Is the short-term "elevated amyloid plaque levels" result actually tied to Alzheimer's, or do they just share a mechanism?

Summarizing a one-weekend insulin sensitivity result as though it settles 50 different questions sharing a broad label is ridiculous.

You need activity/exercise for good sleep.

They've done some previous work on catch-up sleep and demonstrated that insulin sensitivity recovers to baseline, though not necessarily fast. Their claim here is that catch-up sleep doesn't seem to offset prior effects (e.g. by overcorrecting) or buffer against new deprivation (that is, you fall off abruptly to where the no-weekend people are). They didn't have anybody do 8 hours, no recovery on weekends, which is a shame, but they also didn't claim anything beyond "recovery sleep doesn't counteract/prevent the damage".

The person who wrote the WP article, on the other hand, has an awful lot to answer for... That garbage title is only the start of the horrible misrepresentation.

Knocking him for collaborating is ignorant and disingenuous. Scientists collaborate. His record shows he is a real scientist. It doesn't have to show he is the best or that he deserves a Nobel Prize. At this point, he's probably a better candidate for the Nobel Peace Prize if anything, but this only goes to show what he's been working on lately, which is more activism than scientific research. More Neil deGrasse Tyson than Einstein, but that doesn't make Neil a fake scientist.

JDP's last book he did write for a broader audience. Great move. Even I read it. And he brought with him the scientists and the research that otherwise would not have reached his audience.

His book is maybe 75% work of other scientists, 20% story telling, and 5% what he draws from it. Basically his 12 rules. He demonstrates how to draw meaning and meaningful conclusions rationally, with well grounded premises. If you have low tolerance for the more abstract genre of ancient story analysis, then skip it. Or maybe there is some truth to it? Decide for yourself.

Did you read his book? Fine if you didn't but I wouldn't claim to understand him if I hadn't. Maybe you have issues with him, which is fine, but again, that is a separate matter than the substance of his lectures. Which are all highly specific by the way. So to say they're all based on weak papers or that he's selling snake oil is an immediate tell that, again, you either don't understand him or have other issues/motives.

(Thanks HN for not letting me reply, claiming I'm posting too fast, when clearly I'm not. Not sure how you're suppose to have a conversation.)