Correlated

As the objection “correlation is not causation” has ankylosed into a thoughtless reflex, it has become a confusion generator. So it’s worth taking a step back:

… whilst it is true that correlation does not necessarily equate to causation, all causally related variables will be correlated. Thus correlation is always necessary (but not in and of itself sufficient) for establishing causation.

The claim that ‘correlation does not equal causation’ is therefore meaningless when used to counter the results of correlative studies in which specific causal inferences are being made, as the inferred pattern of causation necessarily supervenes upon correlation amongst variables. Whether the variables being considered are in actuality causally associated as per the inference is another matter entirely. …

Correlation is evidence. Causation is theory (and even, inevitably, ‘speculative’ theory).

August 26, 2014admin 17 Comments »
FILED UNDER :Discriminations

TAGGED WITH : ,

17 Responses to this entry

  • Kgaard Says:

    Jesus … THANK YOU for this. I am so sick of progs self-righteously responding to any non-PC observation by saying “correlation does not equal causation,” and then accusing me of sloppy logic. The higher the IQ of the prog, the more vicious they are about it. Infuriating …

    [Reply]

    Posted on August 26th, 2014 at 3:10 pm Reply | Quote
  • Antisthenes Says:

    Correlation is not causation (when the inference to causation offends my ideology).

    A philosophical meme goes the way of all memes, becoming purely rhetorical, the syntactic virus living on after the extinction of thought.

    [Reply]

    Posted on August 26th, 2014 at 3:10 pm Reply | Quote
  • ReactionaryFerret Says:

    “Whether the variables being considered are in actuality causally associated as per the inference is another matter entirely.”

    In fairness, this is precisely what “correlation is not causation” refers to. Just because an inference of causation is made from a correlation does not necessarily mean that it is true. “Correlation is not causation” is not to say that correlation is never related to causation, but to say that just because the correlation has been found does not necessarily mean that all potential causes (some maybe with stronger correlations) have been examined. Such is the case with most of the arguments by AGW scaremongers, when they ignore stronger correlations between sun and warming and the inverse nature of carbon to warming (the warming seems to CAUSE the carbon to increase).

    But, I still see your point, when the phrase “correlation is not causation” is used to say that even when the strongest correlation of all possible known options are considered (and Occum’s Razor taken into account) that there’s still no possibility that it could be true just because, y’know, “correlation is not causation”, then the phrase no longer holds its original meaning and has become “I don’t like it, so it’s not true.”

    [Reply]

    Posted on August 26th, 2014 at 4:03 pm Reply | Quote
  • ReactionaryFerret Says:

    “Whether the variables being considered are in actuality causally associated as per the inference is another matter entirely.”

    In fairness, this is precisely what “correlation is not causation” refers to. Just because an inference of causation is made from a correlation does not necessarily mean that it is true. “Correlation is not causation” is not to say that correlation is never related to causation, but to say that just because the correlation has been found does not necessarily mean that all potential causes (some maybe with stronger correlations) have been examined. Such is the case with most of the arguments by AGW scaremongers, when they ignore stronger correlations between sun and warming and the inverse nature of carbon to warming (the warming seems to CAUSE the carbon to increase).

    But, I still see your point. When the phrase “correlation is not causation” is used to say that even when the strongest correlation of all possible known options are considered (and Occum’s Razor taken into account) that there’s still no possibility that it could be true just because, y’know, “correlation is not causation”, then the phrase no longer holds its original meaning and has become “I don’t like it, so it’s not true.”

    [Reply]

    Posted on August 26th, 2014 at 4:03 pm Reply | Quote
  • ReactionaryFerret Says:

    To Admin,

    I double posted. Please remove the first of my double post, as well as this post, since I can’t figure out how to remove my own posts.

    [Reply]

    Posted on August 26th, 2014 at 4:09 pm Reply | Quote
  • neovictorian23 Says:

    “Necessary, but not sufficient” are four golden words that would save the world untold confusion and trouble.

    [Reply]

    Deogolwulf Reply:

    “‘Necessary, but not sufficient’ are four golden words that would save the world untold confusion and trouble.”

    Too true. If only the moderns had borne them in mind. For instance: efficient cause is necessary but not sufficient for the explanation of effect.

    [Reply]

    E. Antony Gray (@RiverC) Reply:

    “He went for a walk because kinesin turn ATP into movement!”

    – no one in their right mind, ever

    (i hope)

    [Reply]

    Posted on August 26th, 2014 at 4:31 pm Reply | Quote
  • Piano Says:

    Why correlation is *usually* not causation: http://www.gwern.net/Causality

    [Reply]

    Porphy's Attorney Reply:

    I ran a VAR on this post though and Granger said “admin has a point, and not just on his head.”

    [Reply]

    Posted on August 26th, 2014 at 6:10 pm Reply | Quote
  • SanguineEmpiricist Says:

    Every discussion about this topic is meaningless without ascertaining who themselves has done the very basic amount of prepatory work necessary and gone through Judea Pearl.

    http://www.amazon.com/Causality-Reasoning-Inference-Judea-Pearl/dp/0521773628

    [Reply]

    Posted on August 26th, 2014 at 8:09 pm Reply | Quote
  • vxxc2014 Says:

    I shall now attempt to ankylose my seemingly only correlated musings into one fused deadly arm. Of solid bone, there’s no use being subtle anymore.

    [Reply]

    Posted on August 26th, 2014 at 8:14 pm Reply | Quote
  • vxxc2014 Says:

    As we speak Gnon’s Crab pincers ankylose into deadly swinging stumps, for Gnon’s Crab feels no pain or empathy.

    The Ankylosis is nigh. Who shall be it’s Xenophon?

    [Reply]

    Posted on August 26th, 2014 at 8:19 pm Reply | Quote
  • vxxc2014 Says:

    O/T but there’s a lesson here.

    Mighty Wurlitzer

    “Together, the groups formed what one U.S. agent called a “Mighty Wurlitzer,” an organ for playing variations on an anticommunist fugue. ”

    http://www.bloomberg.com/apps/news?pid=newsarchive&sid=a1M9EAly2hog

    Apparently included Gloria Steinem and Henry Kissinger.

    [Reply]

    Posted on August 26th, 2014 at 8:28 pm Reply | Quote
  • Shlomo Maistre Says:

    When small minds boast that “correlation doesn’t mean causation” they’re more often than not betraying a misguided grasp of logic.

    Thinking man capable of proving causation in a vacuum that’s devoid of mere happenstance, sophists (small minds) assume that man is capable of “observing” and, more absurd still, “isolating” a particular cause. Were this fantastic belief justified there could be no problem of induction – only its solution.

    Throw a ball in the air and it will drop to the ground. Man observes the law of gravity, but he proves no law by observation or any action whatsoever.

    All man observes is effect. All the natural world is an endless interplay of effects across space and time. Every material force is a cause only insofar as it is already an effect.

    Obviously correlation does not mean causation. And were inductive reasoning a real thing then perhaps man would be capable of “proving” causation… ever. Alas.

    [Reply]

    Posted on August 26th, 2014 at 10:59 pm Reply | Quote
  • Mike in Boston Says:

    I haven’t gone through it in detail (nor even gotten, to my and perhaps SanguineEmpiricist’s dismay, through all of Pearl) but David Jensen has some interesting work on plausibly inferring causality if your data is relational enough and you have plenty of it. See e.g. http://auai.org/uai2013/prints/papers/197.pdf .

    [Reply]

    Posted on August 27th, 2014 at 5:28 am Reply | Quote
  • sviga lae Says:

    To bayessplain, correlation is evidence to the exact extent that the lack of correlation is anti-evidence.

    [Reply]

    Posted on August 27th, 2014 at 9:54 am Reply | Quote

Leave a comment