There is something about Natália’s and Guzey’s interaction through your posts and in the comment section that doesn’t feel right to me. I haven’t been able to pin down exactly why I get this feeling but my best guess is that it seems like you are missing your actual disagreement.
To me it seems like you are both agreeing on the existing evidence is week. You bring up evidence that point in different directions sure but neither of you seem to bring up strong evidence.
However, you seem to have quite different beliefs on the matter. Natália says:
Our priors about sleep research should be high
I feel grumpy, dumb and distractable every time I sleep less than 7 or so hours. I can’t do things that require focused attention like solving physics homework problems very effectively, and I don’t get nearly as much pleasure when I attempt doing so. My memory becomes very poor: after a recent night of <6 hours of sleep, I somehow forgot the reasoning behind several Manifold Markets trades I had made the prior evening and just stared at them in utter confusion for several minutes before remembering.
and high priors would explain quite high odds (19:1 in two cases) given for the posterior despite what seems to be fairly week evidence[1]. However, note that the reason for this prior seem to be based on personal observations and on abrupt changes in sleep schedule. I agree about the observation that reducing the amount of sleep does impair my cognitive capacity in the short term, but then the same applies to the week after adding or removing daylight saving times from which I do cope quite well after that week. It seems unclear to me if a gradual change in the amount of sleep would have a significant impact without some good data.
On the other hand, Guzey also seem to put most weight into anecdotal evidence. It seems to me that one real possibility that there are large individual differences which would explain that your different priors and why you can find anecdotal evidence both for that little sleep is bad for you and that you can get by with less sleep than otherwise claimed.
The competing hypothesis is that most people are bad at judging how much sleep they need and how their sleep amounts affect their health. This would mean that priors based on own experiences and similar experiences from some other people should not carry much weight when applied to people in general. Furthermore, that the few existing studies with reasonable methodologies measuring effects which are not just reported by subject’s experiences are relatively more important.
Most of the stated posteriors in the conclusion are something I find reasonable when applied to myself but I wouldn’t use so strong posteriors when applied to people in general.
There is something about Natália’s and Guzey’s interaction through your posts and in the comment section that doesn’t feel right to me. I haven’t been able to pin down exactly why I get this feeling but my best guess is that it seems like you are missing your actual disagreement.
To me it seems like you are both agreeing on the existing evidence is week. You bring up evidence that point in different directions sure but neither of you seem to bring up strong evidence.
However, you seem to have quite different beliefs on the matter. Natália says:
and high priors would explain quite high odds (19:1 in two cases) given for the posterior despite what seems to be fairly week evidence[1]. However, note that the reason for this prior seem to be based on personal observations and on abrupt changes in sleep schedule. I agree about the observation that reducing the amount of sleep does impair my cognitive capacity in the short term, but then the same applies to the week after adding or removing daylight saving times from which I do cope quite well after that week. It seems unclear to me if a gradual change in the amount of sleep would have a significant impact without some good data.
On the other hand, Guzey also seem to put most weight into anecdotal evidence. It seems to me that one real possibility that there are large individual differences which would explain that your different priors and why you can find anecdotal evidence both for that little sleep is bad for you and that you can get by with less sleep than otherwise claimed.
The competing hypothesis is that most people are bad at judging how much sleep they need and how their sleep amounts affect their health. This would mean that priors based on own experiences and similar experiences from some other people should not carry much weight when applied to people in general. Furthermore, that the few existing studies with reasonable methodologies measuring effects which are not just reported by subject’s experiences are relatively more important.
Most of the stated posteriors in the conclusion are something I find reasonable when applied to myself but I wouldn’t use so strong posteriors when applied to people in general.