Beer and " statistical correlation "

Forgive the pedantry for a moment : the dictionary defines

    " correlation " as  the degree to which two or more attributes  or

    measurements  on the same group of elements  show a tendency to vary

    together . ( Webster's College Dictionary )

               Again and again in the HEALTH NEWS , I encounter this

    flaw in logical reasoning : a statistical correlation is established

    -for example  between church going  and longevity or in today's news,

    between - of all things ! - beer drinking and a health heart. Then

    the naive scientifically illiterate  assumes that medical science has

    advanced in real knowledge.

                Common sense would caution doctors  against telling the

    public  that an  ADDICTIVE substance- like beer or wine - is good for

    them in " moderate " quantities. My own experience is that this sort

    of pseudo-science will only reinforce problem drinking : " If a little

    of the stuff is good , then more must be better !  "

          .
             I'll bet a strong statistical correlation can be established

    between people  with cancer who visit Elvis Presley's  Graceland mansion

    and the percentage among them of MIRACLE CURES - a correlation that

    would put the famous Fatima to shame !

               What about a statistical correlation between work as a

    newspaper editor and stomach ulcers ?

                  Or between being left-handed ( sinister in Latin ) and

    being a serial killer ?

              Unless you know the CAUSE of a phenomenon, a statistical

    correlation only suggests further investigation.

                  Now I need a nice cold beer. Good for heart, you know.