5 ms·

Correlation might not imply causation, but nothing implies causation. The best we've got is a pile of correlations, a few anecdotes, and a lot of imagination.

jey 17y ago
You can infer a lot about the underlying causal processes from the "pile of correlations". Humans do this all the time. If flipping a light switch correlates highly with the lamp's state switching from "on" to "off" and vice-versa, and in every observed instance the light-switch was flipped before the light changed state, wouldn't you assign a higher probability to the causal model "flipping the switch causes lamp to change state" than "lamp changing state causes switch to flip"? In other words, we can do a lot better than to just say that "lamp changing state and switch-flipping are highly correlated". http://plato.stanford.edu/entries/causation-probabilistic/ http://plato.stanford.edu/entries/causation-probabilistic/ (Sorry for being pedantic.)
- hugh_ 17y ago
  In this case we can be reasonably confident in the causal link between the switch and the light because we're controlling the switch ourselves -- this helps us eliminate the other causal possibilities (a) that the light going on and off causes the switch to flip and (b) both the light and the flipping of the switch are separately caused by some third factor of which we're ignorant. In situations where we can't freely vary any of the parameters we're always going to have a lot more difficulty. Given a pile of correlations between, say, happiness and the countless other variables in the mere two hundred or so countries which exist (eg "average bovine thigh circumference"), it'd be impossible, in the absence of any good theories about what should make people happy, to determine what does make people happy. Luckily we have pretty good ideas from our own observations and from those of others about what actually does make people happy: health, wealth, nice weather, absence of civil war, et cetera. But we'll never be able to get anything other than the vaguest confirmation of what we already believed out of statistical methods alone.
  - jey 17y ago
    I can make the same inferences even if I'm just observing another person or some robot flipping the switch, so no, it doesn't depend on us "controlling" the switch.
    - hugh_ 17y ago
      But the robot _might_ be being controlled by some guy in another room who is also controlling the lightbulb (alternatively the person could be following instructions given by an earpiece). Or maybe they're both hooked up to a geiger counter. I admit it seems pretty unlikely that anybody would wire a robot or person up to flip a switch at exactly the same time a lightbulb comes on, but it only seems unlikely because of our pre-existing knowledge about robots and lightbulbs. If we take away that knowledge and simply make it a statement about how variable A changes when variable B does, it's much harder to infer which way, if either, the causal link goes. (I suppose if you really wanted to be paranoid you could suppose that your _own_ actions in flipping the lightswitch might be subconsciously being controlled by someone else, but this is a whole different level of skepticism)
  - mistermann 17y ago
    Or, we could just go back to the original topic of the article...people in Costa Rica seem to be happy. Disbanding their military (which I was shocked to read, honestly) and spending the money (presumably) on education, amazingly, seemed to result in a better country? Hell, most countriesd if they disbanded their military and burned the money instead would be better off. It seems to me, assuming this survey is remotely correct, is people enjoy a life where they are not sticking their noes into other countires business, and having nice weather and a strong social construct is pleasurable.
3pt14159 17y ago
Somethings imply (operative word here) causation. I type the little letter "a" on my keyboard and an "a" comes up in emacs, thats causation.
Eliezer 17y ago
A common misconception, but you'd have to read quite a bit of Judea Pearl to understand how certain patterns of conditional statistical independence do allow us to declare causation. See this page for a start: http://bayes.cs.ucla.edu/home.htm http://bayes.cs.ucla.edu/home.htm Or the opening paragraphs of this: http://lesswrong.com/lw/qr/timeless_causality/ http://lesswrong.com/lw/qr/timeless_causality/
- DaniFong 17y ago
  Thank you for your links, and writing the article that you've written. I've tried to understand it; though in honesty skimming it more than reading it, and I haven't given it the fresh thought that it likely requires. But. How can one declare causation if one isn't even sure of one's senses? One's observations? One's data? How can we avoid the possibility of an 'evil daemon' of Cartesian philosophy: who is "as clever and deceitful as he is powerful, who has directed his entire effort to misleading me." http://en.wikipedia.org/wiki/Evil_daemon http://en.wikipedia.org/wiki/Evil_daemon Sure, certain patterns patterns of our observations will nudge us one way or the other towards a higher probability of assigning some, possibly specific, causal link. But can we ever be sure of it? This is the sense in which I mean that nothing 'implies' causation. Perhaps it is better to say that some things merely suggest it.
  - Eliezer 17y ago
    In precisely the same sense, nothing implies anything. So why single out causality?
    - DaniFong 17y ago
      Mainly because people so commonly dismiss a study by stating that 'correlation doesn't imply causation,' while going about their lives with their own ideas and biases based on anecdotal evidence just the same. It's the tool of a common course of visible hypocrisy. Maybe it's just a personal thing.

jey 17y ago

You can infer a lot about the underlying causal processes from the "pile of correlations". Humans do this all the time. If flipping a light switch correlates highly with the lamp's state switching from "on" to "off" and vice-versa, and in every observed instance the light-switch was flipped before the light changed state, wouldn't you assign a higher probability to the causal model "flipping the switch causes lamp to change state" than "lamp changing state causes switch to flip"? In other words, we can do a lot better than to just say that "lamp changing state and switch-flipping are highly correlated". http://plato.stanford.edu/entries/causation-probabilistic/ http://plato.stanford.edu/entries/causation-probabilistic/ (Sorry for being pedantic.)

hugh_ 17y ago
In this case we can be reasonably confident in the causal link between the switch and the light because we're controlling the switch ourselves -- this helps us eliminate the other causal possibilities (a) that the light going on and off causes the switch to flip and (b) both the light and the flipping of the switch are separately caused by some third factor of which we're ignorant. In situations where we can't freely vary any of the parameters we're always going to have a lot more difficulty. Given a pile of correlations between, say, happiness and the countless other variables in the mere two hundred or so countries which exist (eg "average bovine thigh circumference"), it'd be impossible, in the absence of any good theories about what should make people happy, to determine what does make people happy. Luckily we have pretty good ideas from our own observations and from those of others about what actually does make people happy: health, wealth, nice weather, absence of civil war, et cetera. But we'll never be able to get anything other than the vaguest confirmation of what we already believed out of statistical methods alone.
- jey 17y ago
  I can make the same inferences even if I'm just observing another person or some robot flipping the switch, so no, it doesn't depend on us "controlling" the switch.
  - hugh_ 17y ago
    But the robot _might_ be being controlled by some guy in another room who is also controlling the lightbulb (alternatively the person could be following instructions given by an earpiece). Or maybe they're both hooked up to a geiger counter. I admit it seems pretty unlikely that anybody would wire a robot or person up to flip a switch at exactly the same time a lightbulb comes on, but it only seems unlikely because of our pre-existing knowledge about robots and lightbulbs. If we take away that knowledge and simply make it a statement about how variable A changes when variable B does, it's much harder to infer which way, if either, the causal link goes. (I suppose if you really wanted to be paranoid you could suppose that your _own_ actions in flipping the lightswitch might be subconsciously being controlled by someone else, but this is a whole different level of skepticism)
- mistermann 17y ago
  Or, we could just go back to the original topic of the article...people in Costa Rica seem to be happy. Disbanding their military (which I was shocked to read, honestly) and spending the money (presumably) on education, amazingly, seemed to result in a better country? Hell, most countriesd if they disbanded their military and burned the money instead would be better off. It seems to me, assuming this survey is remotely correct, is people enjoy a life where they are not sticking their noes into other countires business, and having nice weather and a strong social construct is pleasurable.

3pt14159 17y ago

Somethings imply (operative word here) causation. I type the little letter "a" on my keyboard and an "a" comes up in emacs, thats causation.

Eliezer 17y ago

A common misconception, but you'd have to read quite a bit of Judea Pearl to understand how certain patterns of conditional statistical independence do allow us to declare causation. See this page for a start: http://bayes.cs.ucla.edu/home.htm http://bayes.cs.ucla.edu/home.htm Or the opening paragraphs of this: http://lesswrong.com/lw/qr/timeless_causality/ http://lesswrong.com/lw/qr/timeless_causality/

DaniFong 17y ago
Thank you for your links, and writing the article that you've written. I've tried to understand it; though in honesty skimming it more than reading it, and I haven't given it the fresh thought that it likely requires. But. How can one declare causation if one isn't even sure of one's senses? One's observations? One's data? How can we avoid the possibility of an 'evil daemon' of Cartesian philosophy: who is "as clever and deceitful as he is powerful, who has directed his entire effort to misleading me." http://en.wikipedia.org/wiki/Evil_daemon http://en.wikipedia.org/wiki/Evil_daemon Sure, certain patterns patterns of our observations will nudge us one way or the other towards a higher probability of assigning some, possibly specific, causal link. But can we ever be sure of it? This is the sense in which I mean that nothing 'implies' causation. Perhaps it is better to say that some things merely suggest it.
- Eliezer 17y ago
  In precisely the same sense, nothing implies anything. So why single out causality?
  - DaniFong 17y ago
    Mainly because people so commonly dismiss a study by stating that 'correlation doesn't imply causation,' while going about their lives with their own ideas and biases based on anecdotal evidence just the same. It's the tool of a common course of visible hypocrisy. Maybe it's just a personal thing.