> "Authoritative recommendations" aren't the only thing we get out of science.
I know, but when is the last time you saw a diet recommendation article of the form "Our Bayes factor on kale doing such and such is..."?
The only things 99.9% of people pay attention to (or know how to interpret) are authoritative-sounding recommendations, so that's what diet writers write.
> science's job is to attempt to give us a best guess at what the most likely answer to a question is.
No, I think you actually misunderstand what science can do for us here. Science can give us confidence intervals that let us maximize expected utility given what we know. We don't just assume that the most likely hypothesis is true.
> advice like Pollan's is a good balance between simple and as effective as we can hope for at this pt
Much of the advice in the article is too specific and based on weak evidence. For example, the suggestion not to eat much meat; there are huge amounts of research suggesting that eating mostly meat is a very good approach. I'm not going to claim this is true, because as I said, it's actually pretty weak evidence, but there's certainly no good justification to authoritatively claim the opposite.
> I know, but when is the last time you saw a diet recommendation article of the form "Our Bayes factor on kale doing such and such is..."?
The only things 99.9% of people pay attention to (or know how to interpret) are authoritative-sounding recommendations, so that's what diet writers write.
Attacking writers (to no disagreement from me) is a pretty odd response to a defense of science's role in society. Science coverage is garbage, but that has approximately nothing to do with your point, nor my response to it.
> No, I think you actually misunderstand what science can do for us here. Science can give us confidence intervals that let us maximize expected utility given what we know. We don't just assume that the most likely hypothesis is true.
Oh for the love of God, this is a masterpiece of pedantry. You're mistaking "I assume everyone here is intelligent enough to understand what I'm getting at despite my lack of precision in a brief three-sentence comment" for "I don't understand what I'm talking about". Just because you don't fit the former statement doesn't mean I fit the latter.
> Science can give us confidence intervals that let us maximize expected utility given what we know.
Is this possible in the field of diet, considering the long timespan needed for experimentation, the numerous contributing factors, and the fact that study participants are very often not telling the truth?
I know, but when is the last time you saw a diet recommendation article of the form "Our Bayes factor on kale doing such and such is..."?
The only things 99.9% of people pay attention to (or know how to interpret) are authoritative-sounding recommendations, so that's what diet writers write.
> science's job is to attempt to give us a best guess at what the most likely answer to a question is.
No, I think you actually misunderstand what science can do for us here. Science can give us confidence intervals that let us maximize expected utility given what we know. We don't just assume that the most likely hypothesis is true.
> advice like Pollan's is a good balance between simple and as effective as we can hope for at this pt
Much of the advice in the article is too specific and based on weak evidence. For example, the suggestion not to eat much meat; there are huge amounts of research suggesting that eating mostly meat is a very good approach. I'm not going to claim this is true, because as I said, it's actually pretty weak evidence, but there's certainly no good justification to authoritatively claim the opposite.