When looking at representation and determining if there is advserse impact 
you need to look at the percentages as below:


                 Total              1                     2                  
3                      4                     5                 NR  
Males          26         4  (15.4%)      16 (61.7%)           6 
(23.1%)         0  (0.0%)         0 (0.0%)       0 (0.0%)  

Females       6           0   ( 0.0%)        3 (50.0%)            2 
(33.3%)         1 (16.7%)        0 (0.0%)       0 (0.0%)


You want to determine if women (or minorities, or over 40), based on their 
representation, are being adversely effected.  You look at the total women 
and how their distribution over the ratings equates, then you look at the 
total men and how their distribution equates--you are looking at them 
individually and not overall.  After this is calculated you determine by 
rating category if women are within 80% of the ratings men received-- if they 
are within 80% there is no adverse impact.  If it is less than 80% there is 
adverse impact.  For example, let's look at the "2" category.  There are 16 
of 26 men that received "2's" or 61.5% of the male representation.  There are 
3 of 6 females that received "2's"  or 50% of the female representation.  To 
determine if the females are being adversely effected you take 50% and divide 
it by 61.5% and get 81.3%--as this is more than 80% there is no adverse 
impact.  Now, let me throw a curve ball.  In calculating adverse impact it is 
assumed that if women (or minorites) are being effected more it is a bad 
thing.  In the case of looking at ratings this may not be the case so a 
standard formula cannot be applied.  For example,  if more women are getting 
"1's" than the males this is not a bad thing but if more women are getting 
"5"s this is a bad thing!  You have to use some judgement on this because if 
more female's based on their representation are getting 3's maybe that is bad 
and maybe that is good--you have to look at the overall picture.  When you 
are calculating adverse impact on employment actions like hires or 
terminations it is very straight forward but when looking at it from a rating 
perspective it can get tricky because you are making comparisions over 
mutiply categories which mean different things--the 1's being good and the 
5's being not so good. 

I would recommend that you proceed with the chart as I typed it above and not 
try to build in the adverse impact piece.  The HR person looking at the data 
in the PRC meeting will need to do some quick calculations based on how the 
distribution is for determining adverse impact.  If you have any questions or 
if I can help please let me know.




 -----Original Message-----
From:  Yowman, Andrea  
Sent: Tuesday, June 05, 2001 10:05 AM
To: Acevedo, Felecia
Subject: FW: PRC Demographics

Please give me your thoughts

 -----Original Message-----
From:  Inglis, Elspeth  
Sent: Tuesday, June 05, 2001 9:28 AM
To: Yowman, Andrea; Cash, Michelle
Cc: Corteselli, Gina
Subject: FW: PRC Demographics

FYI

this is the revised demographic screen. Please review and advise. thanks
Elspeth

 << OLE Object: Picture (Device Independent Bitmap) >>