Causation vs. Correlation

From a brief conversation with John Sterman

Me:

"I enjoyed our brief conversation - I was wondering if you had any useful articles/resources used in the proof of Causation.

The only absolute case of causation I can think of is something that is, by definition, biconditional. Using the example in class, a person is either infected or uninfected, so clearly there is a casual link between the two populations.

I found the Causation vs. Correlation section (5.2.1) in the book helpful, I was just curious if you had any preferred methods/approaches of proof for causation."

Professor Sterman's Response:

"Thanks Bryan,

this is a deep issue, and statisticians, economists and others continue to struggle with it.

IMO, there are no absolute proofs of causation (this goes back to Hume at least in philosophy).

You might jump ahead and read Ch. 21 in the book (on model testing and validation).

If you are really interested in the philosophy of knowledge/phil. of science, check out the Duhem-Quine thesis (google it). I also recommend the work of Thomas Kuhn highly (specifically, his short but classic Structure of Scientific Revolutions). and see the system dynamics models of his theory I built years ago (attached)."

Attachment 1: The Growth of Knowledge  ·  Attachment 2: Path Dependence, Competition, and Succession in the Dynamics of Scientific Revolution