3 ms·

> That's what a positive correlation means. Not so, at least, not without an explanation backed up by evidence. You're missing something fundamental -- a posit

by lutusp 13y ago

> That's what a positive correlation means. Not so, at least, not without an explanation backed up by evidence. You're missing something fundamental -- a positive correlation does not carry evidentiary weight until it's been explained. Science is not a collection of descriptions, it's a discipline that is sometimes able to uncover explanations -- where they exist. Alongside the old saying "correlation doesn't prove causation" should be another similar cautionary note about uncovering evidence for the meaning of a correlation before describing it as meaningful. http://www.businessweek.com/magazine/correlation-or-causation-12012011-gfx.html http://www.businessweek.com/magazine/correlation-or-causatio...

gwern 13y ago
> a positive correlation does not carry evidentiary weight until it's been explained. What on earth? A positive correlation is a fact. This is like looking in a microscope and saying, 'ah, none of this data is true until we have a theory which can explain it'. Uh... no. The data is the data. The correlations are the correlations.
- lutusp 13y ago
  > What on earth? A positive correlation is a fact. You know, you could just learn how science works. Did you follow the link I provided? It shows that correlations mean nothing without an explanation. > This is like looking in a microscope and saying, 'ah, none of this data is true until we have a theory which can explain it'. Yes, and that is exactly how science works -- observation means nothing unless and until it addresses a theory. But to show you what you're missing, let's do a little experiment. Let's say I'm a doctor and I've created a revolutionary cure for the common cold. My cure is to shake a dried gourd over the cold sufferer until he gets better. The cure might take a week, but it always works. My method is repeatable and perfectly reliable, and I've published my cure in a refereed scientific journal (there are now any number of phony refereed scientific journals). And, because (in this thought experiment) your version of science can get along without defining theories, I'm under no obligation to try to explain my cure, or consider alternative explanations for my breakthrough — I only have to describe it, as you seem to think. Because I've cured the common cold, and because I've met all the requirements that you feel are necessary for science, I deserve a Nobel Prize. Yes or no?
  - gwern 13y ago
    > You know, you could just learn how science works. Your condescension is hilarious. > Because I've cured the common cold, and because I've met all the requirements that you feel are necessary for science, I deserve a Nobel Prize. Yes or no? No. My observation of your claimed data is fine, but it is an additional matter of interpretation to make any causal claims, or claim external or internal validity. What you are not getting is that the correlation stands on its own. It's when you try to read into it some other claim (and I use the pronoun 'you' very deliberately here), like 'the rich are rich solely because they are smart' or 'only smart people are rich' or 'the rich are morally better', and are revolted by them, that you try to pretend the data doesn't exist. It does. Deal.
    - lutusp 13y ago
      >> You know, you could just learn how science works. > Your condescension is hilarious. When one seen someone assume a correlation has a meaning not proven in the evidence, it's not condescension, it's charity. > What you are not getting is that the correlation stands on its own. Yes, as does the causal relationship between puddles and rain in advance of an attempt to offer an explanation: puddles predict rainfall. > but it is an additional matter of interpretation to make any causal claims, or claim external or internal validity. Yes, and now that you're reduced to making my points for me, we're done.
      - gwern 13y ago
        > Yes, as does the causal relationship between puddles and rain in advance of an attempt to offer an explanation: puddles predict rainfall. Indeed, it does. Which is why a puddle-denier is a sad sight. 'There can't be puddles because no one has offered a fully-comprehensive theory which predicts puddles and which I find acceptable! Why must people be so puddleist?' > Yes, and now that you're reduced to making my points for me, we're done. Denying your fallacious original overreadings and adopting your opponent's criticisms and claiming to have held them all along, while flattering, is still very annoying.
      - gwern 13y ago

by lutusp 13y ago

gwern 13y ago

> a positive correlation does not carry evidentiary weight until it's been explained. What on earth? A positive correlation is a fact. This is like looking in a microscope and saying, 'ah, none of this data is true until we have a theory which can explain it'. Uh... no. The data is the data. The correlations are the correlations.

lutusp 13y ago
> What on earth? A positive correlation is a fact. You know, you could just learn how science works. Did you follow the link I provided? It shows that correlations mean nothing without an explanation. > This is like looking in a microscope and saying, 'ah, none of this data is true until we have a theory which can explain it'. Yes, and that is exactly how science works -- observation means nothing unless and until it addresses a theory. But to show you what you're missing, let's do a little experiment. Let's say I'm a doctor and I've created a revolutionary cure for the common cold. My cure is to shake a dried gourd over the cold sufferer until he gets better. The cure might take a week, but it always works. My method is repeatable and perfectly reliable, and I've published my cure in a refereed scientific journal (there are now any number of phony refereed scientific journals). And, because (in this thought experiment) your version of science can get along without defining theories, I'm under no obligation to try to explain my cure, or consider alternative explanations for my breakthrough — I only have to describe it, as you seem to think. Because I've cured the common cold, and because I've met all the requirements that you feel are necessary for science, I deserve a Nobel Prize. Yes or no?
- gwern 13y ago
  > You know, you could just learn how science works. Your condescension is hilarious. > Because I've cured the common cold, and because I've met all the requirements that you feel are necessary for science, I deserve a Nobel Prize. Yes or no? No. My observation of your claimed data is fine, but it is an additional matter of interpretation to make any causal claims, or claim external or internal validity. What you are not getting is that the correlation stands on its own. It's when you try to read into it some other claim (and I use the pronoun 'you' very deliberately here), like 'the rich are rich solely because they are smart' or 'only smart people are rich' or 'the rich are morally better', and are revolted by them, that you try to pretend the data doesn't exist. It does. Deal.
  - lutusp 13y ago
    >> You know, you could just learn how science works. > Your condescension is hilarious. When one seen someone assume a correlation has a meaning not proven in the evidence, it's not condescension, it's charity. > What you are not getting is that the correlation stands on its own. Yes, as does the causal relationship between puddles and rain in advance of an attempt to offer an explanation: puddles predict rainfall. > but it is an additional matter of interpretation to make any causal claims, or claim external or internal validity. Yes, and now that you're reduced to making my points for me, we're done.
    - gwern 13y ago
      > Yes, as does the causal relationship between puddles and rain in advance of an attempt to offer an explanation: puddles predict rainfall. Indeed, it does. Which is why a puddle-denier is a sad sight. 'There can't be puddles because no one has offered a fully-comprehensive theory which predicts puddles and which I find acceptable! Why must people be so puddleist?' > Yes, and now that you're reduced to making my points for me, we're done. Denying your fallacious original overreadings and adopting your opponent's criticisms and claiming to have held them all along, while flattering, is still very annoying.
    - gwern 13y ago