On Mon, 17 Apr 2000 10:52:46 -0800, Shareef Siddeek
<[log in to unmask]> wrote:
> I do not agree with your statement that correlation can never be evidence
> of causality. It is a sign  for the scientist to investigate further to
> find the causality.

It is reasonable to conclude that, given a mechanistically plausible
model, high correlation is _preliminary_ evidence of causality, and
it is a sign for the scientist to investigate further whether
causality really does exist.  That investigation generally requires
perturbing the system and then (1) determining whether the changes
predicted by the model really do occur, and (2) tracing those changes
through any intermediate causal pathways that are thought to exist.

> My understanding is that
> with high causality you will get high correlation.

However, the far more useful reverse inference is _not_ necessarily true.
That is, high correlation does not 'prove' causation.

><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><>
       To leave the Fish-Sci list, Send blank message to:
        mailto:[log in to unmask]
 For information send INFO FISH-SCI to [log in to unmask]
><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><>