[baddox]

Wait, what? Why would self-reported feelings not be taken as strong evidence? It seems to me like self-reported feelings are great evidence, and need particularly strong evidence (e.g. evidence of long-term damage that isn’t felt in the short term, like for smoking cigarettes) to contradict.

[judofyr]

> Wait, what? Why would self-reported feelings not be taken as strong evidence?

Because you might be stuck in a local maxima that's close to your global minima. I've been reading "Why We Sleep", and while I don't have the citations here at the moment he mentions that one of the tricky things with sleep deprivation is that people are often not very aware of it themselves. Your body gets used to the new, lower, level and thinks that it's normal.

It should also be mentioned that most people don't try to distinguish between correlation and causation, and as such it can be hard to draw any conclusions. Imagine a person who reports that "I'm so happy when I drink alcohol", but it turns out that he has no social contact (e.g. working at night, sleeping through the day; no friends) outside of the bar setting, and it's actually the social element that he most desires.

I guess it depends on what you mean by the word "evidence", but I wouldn't really say that his self-reported feelings show any strong evidence that alcohol makes him happy.

[0] https://www.goodreads.com/book/show/34466963-why-we-sleep

[pbhjpbhj]

Because we're really good at fooling ourselves, and we lie to present an idealised version of ourselves.

Example: [Apparently] having coffee in the morning doesn't wake you up, it's that caffeine dependency (and maybe a little dehydration) makes you feel terrible. Coffee temporarily reverse this negative [local maxima].

Coffee appears to wake you up, but it's doing the opposite.

There are many common examples - 'drinking doesn't affect my driving', 'I don't drink heavily', 'advertising doesn't affect me', 'I get enough exercise', 'my diet is healthy', 'I go to church regularly'.

Reports here https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5639921/#!po=7...., for example, have measured overreporting of exercise/church attendance at double.