Date:         Wed, 14 May 1997 10:58:10 EDT
Reply-To:     Christopher Zogby <chzogby@FDIC.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Christopher Zogby <chzogby@FDIC.GOV>
Subject:      Re: T-tes to test if mean=a certain value
Comments: To: salvruiz@FCU.UM.ES

Salvador:
 
You may be able to perform a variation of a matched pair t-test.  You could
test the hypothesis (null) that the difference between your variable and your
constant is 0.  However I don't know if this is the same as determining if a
population mean is different from a constant value.  If the former suggestion
suits your needs see the code below.  Note, you should make sure the data
meet the assumptions of your test.
 
Data diff;
  set raw_data;
 
 diff = somedata - constant;
 
 
label diff = 'Ho: DIFF = 0';
 
run;
 
proc means  data=diff  n mean std min max t prt;
var diff;
title 'If prt < 0.05 then somedata not = constant.';
run;
 
 
 
Regards,
 
Chris
 
--
________________________________________________________________
Christopher Zogby                       Voice:  (202) 898 - 3520
SAS Consultant                          Fax:    (202) 898 - 8636
Federal Deposit Insurance Corporation   E-mail: chzogby@fdic.gov
 
 
 
Salvador Ruiz wrote:
 
>Hi folks,
 
>Does anybody know how to test in SAS if the mean of a variable is equal to a
>certain value?
 
>I could only find how to test the comparison between to means.
 
>Thanks a lot
>_________________________________________________
>
>Salvador Ruiz
>Prof. de Comercializacion e Inv. de Mercados
>Dpto. de Economia de la Empresa
>Universidad de Murcia                      Tel: (9)68-363802
>Avda. Ronda de Levante, 10            Fax: (9)68-363792
>30.008 Murcia. SPAIN                       e-mail: salvruiz@fcu.um.es
>_________________________________________________