From drwhite@orion.oac.uci.edu Fri Feb 26 07:58:23 1999
Date: Fri, 26 Feb 1999 07:00:57 -0800 (PST)
From: "Douglas R. White " 
To: Multiple recipients of list <60370-w99@ea.oac.uci.edu>
Subject: Re: Anthro174


> Hello Professor White
> 	I'm having alot of trouble figuring out what a linear correlation
> is and what a conditional relationship is.  Help. 
> 				bo yoon 

Lets say we have two variables, X and Y.  Intuitively "shape" of
how one variable correlations with the other, across ordered
categories X: a b c d e and Y : A B C D E is like this:

      Y: A B C D E
  X: a   #
     b     #
     c     # #
     d         # #
     e           #

wher # are the disproportionately large numbers of societies.  I
introduced a bit of irregularities for realism.

A conditional relationship might look like this

      Y: A B C D E
  X: a   # # # # #
     b     # # # #
     c     # # # #
     d         # #
     e           #

or like this:

      Y: A B C D E
  X: a   #
     b   # #
     c   # # #
     d   # # # # #
     e   # # # # # #

In the simplest case of only TWO categories for each variable,
the linear relationship looks like this:

     Y:  A B 
  X: a   #
     b     #

or like this: (negative linear correlation)

     Y:  A B
  X: a     #
     b   #
     
But conditional relationships could look like this:


     Y:  A B  (Y:A entails X:a)
  X: a   # #
     b     #
     

     Y:  A B  (conversely: X:a entails Y:A)
  X: a   #
     b   # #
     

     Y:  A B  (opposite or negative: Y:A entails X:b NOT X:a)
  X: a     #
     b   # #
     

     Y:  A B  (exercise: what entails what?
  X: a   # #
     b   #