Date:         Tue, 23 Apr 1996 21:16:02 +0100
Reply-To:     Jose Luis Alabart <joseluis@MIZAR.CSIC.ES>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Jose Luis Alabart <joseluis@MIZAR.CSIC.ES>
Subject:      Re: Geometric Row Mean

        What about this code?
 
DATA D1;
INPUT X1 X2 X3 X4;
gmean=exp(mean(log(x1),log(x2),log(x3),log(x4)));
CARDS;
1  1  1    1
2  2  2    2
1  2  1    2
3  4  5    6
1  10 100 1000
;
PROC PRINT;
RUN;
 
                     The SAS System                                1
             20:46 Tuesday, April 23, 1996
 
       OBS    X1    X2     X3      X4     GMEAN
 
        1      1     1      1       1     1.0000
        2      2     2      2       2     2.0000
        3      1     2      1       2     1.4142
        4      3     4      5       6     4.3559
        5      1    10    100    1000    31.6228
 
 
 
        Perhaps it is a little tedious, but it works... nad there is nothing
that can't be solved by "copy-paste" and "replace"...  :-)
 
 
 
>I posted a question last week about computing the row geometric mean of a
>series of 30 or 40 adjacent variables.  Several kind souls came up with
>what seemed to be fairly easy solutions to the problem, except when I
>tried them they didn't work correctly.
>
>One solution was:
>
>gmean = exp(mean(log(x1 -- x40)));
>
>This solution takes advantage of the fact that the arithmetic mean of
>logarithms is equal to the logarithm of the geometric mean, which can in
>turn be reconstituted to its original form via the exp function.
>This solution is imminently logical and causes no problem, except that it
>gives the wrong answer.  When tested with a small number of variables, it
>appears to take the logarithm of the sum of x1+x2+x3+x4 (in a four
>variable case), instead of summing the logs of x1, x2, x3, x4.  I'm using
>SAS 6.11.
>
>Any other ideas, or comments on what is actually wrong with the code above.
>
>Thank you very much for your time.
>
>  ***********************************************************************
>  * Dr. Marc R. Feldesman     | email: feldesmanm@pdx.edu| Beyond every *
>  * Professor and Chairman    | fax:  (503)-725-3905     | credibility  *
>  * Anthropology Department   | voice: (503)-725-3081    | gap lies a   *
>  * Portland State University | pager: (503)-870-2515    | gullibility  *
>  * Portland, Oregon 97207    | home:  NOT               | fill         *
>  ***********************************************************************
        J.L. Alabart
        Servicio de Investigacion Agraria
        Apdo. 727
        50080 Zaragoza (Spain)
        FAX:   (Spain) 76 57 55 01
        Voice:  (Spain) 76 57 63 36
                    (Spain) 76 57 63 11
        Tx.    58194 IDAE E
        E-mails: JOSELUIS@mizar.csic.es (or)
                    ALABART@cc.unizar.es