1. Forum
  2. Usenet
  3. SCI.STAT.MATH

  • Q relation between corr, causation and stat sig

    From Cosine@21:1/5 to All on Mon Sep 21 18:44:51 2020
    Hi:

    Is there a relationship between correlation, statistical significance, and causation?

    Are there examples of the existence of one but not either of the other two?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rich Ulrich@21:1/5 to All on Tue Sep 22 01:15:20 2020
    On Mon, 21 Sep 2020 18:44:51 -0700 (PDT), Cosine <asecant@gmail.com>
    wrote:

    Hi:

    Is there a relationship between correlation, statistical significance, and causation?

    Are there examples of the existence of one but not either of the other two?

    I immediately think that you are inviting a tangle of
    semantic confusion.

    "Measured" correlation as opposed to "true"? ... correlation
    meaning, "any relationship"?

    "Causation" within what we know of science should yield
    "correlation" (association) if measured using relevant measures
    ... and using a sufficiently large sample produces "statistical
    significance."

    Statistical significance exists 5% of the time for a 5% test
    when there is no true relation at all (correlation or causation
    in any form).

    Semantic confusion.

    --
    Rich Ulrich

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bruce Weaver@21:1/5 to Cosine on Tue Sep 22 07:10:04 2020
    On Monday, September 21, 2020 at 9:44:54 PM UTC-4, Cosine wrote:
    Hi:

    Is there a relationship between correlation, statistical significance, and causation?

    Are there examples of the existence of one but not either of the other two?

    Here are some resources you might find helpful.

    https://www.tylervigen.com/spurious-correlations http://core.ecu.edu/psyc/wuenschk/stathelp/Correlation-Causation.htm https://www.edwardtufte.com/tufte/hill https://pdfs.semanticscholar.org/262b/854628d8e2b073816935d82b5095e1703977.pdf/

    In the essay/speech by Sir Austin Bradford Hill, note the distinction between scientifically establishing causation and deciding to ACT as if causation has been established (e.g., in the section headed "The Case for Action").

    Note too that I have not read the last item. I found it just now, but judging by the abstract, I'm guessing it's pretty sound.

    HTH.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)