5 ms·

> A prospective study by contrast... That might be science, this is not. Woah there. Science n. the systematic study of the structure and behavior of the phys

> A prospective study by contrast... That might be science, this is not. Woah there. Science n. the systematic study of the structure and behavior of the physical and natural world through _observation_ and experiment. This paper does not cover a double-blind randomized controlled-trial. That does not mean that a retrospective matched sample studies are not science. The conclusions are actually pretty reasonable for this type of study they essentially say "we need to make sure the nutrition of vegetarians is adequate, in Austria," "we also need to study this more, in Austria." The different in abundance for mental illness is pretty striking, and seems worth investigating. Previous studies (e.g. [1]) have also linked low-fat diets to change in mental state, the outstanding control-intervention literature combined with retrospective studies make a compelling argument for further research. All that said, the link-bait HN title is pretty bad. [1] http://www.ncbi.nlm.nih.gov/pubmed/9505799 http://www.ncbi.nlm.nih.gov/pubmed/9505799

lutusp 13y ago
> Science n. the systematic study of the structure and behavior of the physical and natural world through _observation_ and experiment. Wait -- you're going to rely on a dictionary definition of science? Dictionaries don't define words, they can only report how people use words. This is why, as just one example, literally and figuratively mean the same thing -- because that's how people choose to use those words. > This paper does not cover a double-blind randomized controlled-trial. That does not mean that a retrospective matched sample studies are not science. Yes, in point of fact, that is exactly what it means -- it cannot distinguish causes from effects, and it cannot produce reliable conclusions about nature, or lead to testable, falsifiable theories about nature. > The conclusions are actually pretty reasonable for this type of study ... Yes, that is absolutely correct -- they're "pretty reasonable for this type of study". No dispute. In science -- real science, not the dictionary version -- observations must successfully isolate the thing being studied, and this is something that retrospective studies cannot do. The study links vegetarianism to certain health issues, just as puddles are successfully linked to rain. Science is not description, science is explanation -- explanations that are tentative, and that are open to falsification. Without a theory about an observation (an explanation) there can be no science. > Previous studies (e.g. [1]) have also linked ... There's my favorite pop-science buzzword: "linked". When unscientific people hear "linked", they think a cause-effect relationship has been established. But that's not what "linked" means, it's what the public thinks it means. It's the same with literally and figuratively -- they are also regularly confused by an ignorant public. > The different in abundance for mental illness is pretty striking, and seems worth investigating. You're missing the point that mental health issues might be a cause of vegetarianism, not an effect, or both vegetarianism and poor mental health are the result of some tertiary cause this study doesn't address. Until an explanation is offered and then tested, until causes and effects can be unraveled, it's the most sordid kind of data mining, and it's ... not ... science.

lutusp 13y ago

> Science n. the systematic study of the structure and behavior of the physical and natural world through _observation_ and experiment. Wait -- you're going to rely on a dictionary definition of science? Dictionaries don't define words, they can only report how people use words. This is why, as just one example, literally and figuratively mean the same thing -- because that's how people choose to use those words. > This paper does not cover a double-blind randomized controlled-trial. That does not mean that a retrospective matched sample studies are not science. Yes, in point of fact, that is exactly what it means -- it cannot distinguish causes from effects, and it cannot produce reliable conclusions about nature, or lead to testable, falsifiable theories about nature. > The conclusions are actually pretty reasonable for this type of study ... Yes, that is absolutely correct -- they're "pretty reasonable for this type of study". No dispute. In science -- real science, not the dictionary version -- observations must successfully isolate the thing being studied, and this is something that retrospective studies cannot do. The study links vegetarianism to certain health issues, just as puddles are successfully linked to rain. Science is not description, science is explanation -- explanations that are tentative, and that are open to falsification. Without a theory about an observation (an explanation) there can be no science. > Previous studies (e.g. [1]) have also linked ... There's my favorite pop-science buzzword: "linked". When unscientific people hear "linked", they think a cause-effect relationship has been established. But that's not what "linked" means, it's what the public thinks it means. It's the same with literally and figuratively -- they are also regularly confused by an ignorant public. > The different in abundance for mental illness is pretty striking, and seems worth investigating. You're missing the point that mental health issues might be a cause of vegetarianism, not an effect, or both vegetarianism and poor mental health are the result of some tertiary cause this study doesn't address. Until an explanation is offered and then tested, until causes and effects can be unraveled, it's the most sordid kind of data mining, and it's ... not ... science.