Date: Wed, 5 Feb 1997 11:42:08 -0800
Reply-To: Yen Wei <Wei.Yen@OFM.WA.GOV>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Yen Wei <Wei.Yen@OFM.WA.GOV>
Subject: Re: PROC TTEST with WEIGHT ?
Content-Type: text/plain; charset="us-ascii"
Please consider this as a reply to the previous postings by both Peter
Luo and Bernard Tremblay.
Peter, you offered an option to replicate as many observations as the
weight indicates. If the sum of the weight is small, that may work.
What if the sum of the weight is very large, say several million? In
that case, this option becomes inefficient.
Bernard, it is not the weight variable that is a dummy. It is the very
variable in the VAR statement. The (0,1) feature of the dummy will be
a problem when you do multiplication and division with the weight
variable to create a new one.
Any other thought is appreciated.
Wei Yen
State of Washington
Office of Financial Management
----------
From: bernard@capitale.qc.ca[SMTP:bernard@capitale.qc.ca]
Sent: Wednesday, February 05, 1997 1:11 AM
To: Yen Wei; SAS-L@UGA.CC.UGA.EDU
Subject: Re: PROC TTEST with WEIGHT ?
Hi
I'm afraid I can't be of much help for that. If your dummy
variable is not continuous then using it as a weigth is not appropriate
for
sure. Could it be more appropriate to use it as a CLASS variable ?
May be there another procedure more adequate ?
Somebody has an idea ?
Bernard Tremblay
>>>From owner-sas-l@UGA.CC.UGA.EDU Wed Feb 5 12:36 EST 1997
>>>Mime-Version: 1.0
>>>Content-Transfer-Encoding: 7bit
>>>Date: Wed, 5 Feb 1997 09:19:37 -0800
>>>From: Yen Wei <Wei.Yen@OFM.WA.GOV>
>>>Subject: Re: PROC TTEST with WEIGHT ?
>>>To: Multiple recipients of list SAS-L <SAS-L@UGA.CC.UGA.EDU>
>>>
>>>Bernard,
>>>
>>>Thanks for your quick response. I am interested in this question,
too.
>>> You solution seems to work when the variable in the VAR statement
is a
>>>continuous one. Is there a solution if pre-weighted variable is
dummy
>>>variable? TIA.
>>>
>>>Wei Yen
>>>State of Washington
>>>Office of Financial Management
>>>
>>>----------
>>>From: Bernard Tremblay[SMTP:bernard@CAPITALE.QC.CA]
>>>Sent: Tuesday, February 04, 1997 10:58 PM
>>>To: Multiple recipients of list SAS-L
>>>Subject: Re: PROC TTEST with WEIGHT ?
>>>
>>>Sorry to all,
>>> It's my mistake I wanted to reply to Nicolas in french.
>>>Here the english version:
>>>
>>> You should compute yourself, in a DATA STEP, a new variable
>>>that would include the weights in it's value. Then, you may be able
>>>to use PROC TTEST with the weighted variable ...
>>>
>>> Regards,
>>> Bernard Tremblay
>>>
>>>
>>>>>>From Wei.Yen@ofm.wa.gov Wed Feb 5 11:27 EST 1997
>>>>>>From: Yen Wei <Wei.Yen@ofm.wa.gov>
>>>>>>To: "'Bernard Tremblay'" <bernard@capitale.qc.ca>
>>>>>>Subject: RE: PROC TTEST with WEIGHT ?
>>>>>>Date: Wed, 5 Feb 1997 08:22:25 -0800
>>>>>>Mime-Version: 1.0
>>>>>>Content-Transfer-Encoding: 7bit
>>>>>>
>>>>>>Would you be kind enough to provide an English version of your
reply
>>>so
>>>>>>those of us who don't speak the French language can understand?
>>>>>> Thanks.
>>>>>>
>>>>>>----------
>>>>>>From: Bernard Tremblay[SMTP:bernard@CAPITALE.QC.CA]
>>>>>>Sent: Tuesday, February 04, 1997 8:34 PM
>>>>>>To: Multiple recipients of list SAS-L
>>>>>>Subject: Re: PROC TTEST with WEIGHT ?
>>>>>>
>>>>>>Bonjour Nicolas,
>>>>>>
>>>>>> Je crois que tu n'auras pas le choix de calculer toi-meme
>>>>>>les valeurs de ta variable avec les "poids" en creeant une
nouvelle
>>>>>>variable dans un "DATA STEP" au prealable. Ensuite tu n'auras
qu'a
>>>>>>utiliser les valeurs que tu as calculees dans ta procedure "PROC
>>>>>>TTEST".
>>>>>>
>>>>>> Bonne chance,
>>>>>>
>>>>>> Bernard Tremblay
>>>>>>
>>>>>>PS: Excuse-moi pour les accents, mais notre courrier electronique
ne
>>>>>> peux pas encore les digerer!
>>>>>>
>>>>>> \\\|///
>>>>>> \\ - - //
>>>>>> ( @ @ )
>>>>>>+------oOOo-(_)-oOOo----------+---------------------------------+
>>>>>>| Bernard Tremblay | |
>>>>>>| La Capitale | Tel: (418) 646-2401 |
>>>>>>| | Fax: (418) 646-5960 |
>>>>>>| | Int: bernard@capitale.qc.ca |
>>>>>>+-----------------------------+---------------------------------+
>>>>>>| Imaginasys enr | Res: (418) 878-4447 |
>>>>>>| | Int: bertrem@quebectel.com |
>>>>>>+---------------Oooo----------+---------------------------------+
>>>>>> oooO ( )
>>>>>> ( ) ) /
>>>>>> \ ( (_/
>>>>>> \_)
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>>>>From owner-sas-l@UGA.CC.UGA.EDU Wed Feb 5 09:28 EST 1997
>>>>>>>>>X-Sender: piepho@fserv.wiz.uni-kassel.de
>>>>>>>>>Mime-Version: 1.0
>>>>>>>>>Date: Wed, 5 Feb 1997 15:20:12 +0100
>>>>>>>>>From: Hans-Peter Piepho <piepho@WIZ.UNI-KASSEL.DE>
>>>>>>>>>Subject: Re: PROC TTEST with WEIGHT ?
>>>>>>>>>Comments: To: Nicolas POUSSING <poussing@DROIT-ECO.U-NANCY.FR>
>>>>>>>>>To: Multiple recipients of list SAS-L <SAS-L@UGA.CC.UGA.EDU>
>>>>>>>>>
>>>>>>>>>>As far as we know, PROC TTEST does not support any weighting
>>>>>>(through
>>>>>>>>>>WEIGHT or FREQ). Hence, how can we test for differences
between
>>>>>>subgroups
>>>>>>>>>>means (and variances) when then data need use of weights ?
>>>>>>>>>>Many thanks in advance.
>>>>>>>>>>reinstad@droit-eco.u-nancy.fr
>>>>>>>>>>
>>>>>>>>>>--
>>>>>>>>>>Nicolas POUSSING
>>>>>>>>>>ADEPS
>>>>>>>>>>4, rue de la Ravinelle
>>>>>>>>>>CO 26
>>>>>>>>>>54035 NANCY CEDEX
>>>>>>>>>>tel : 03 83 30 58 41
>>>>>>>>>>fax : 03 83 35 83 92
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>use PROC GLM
>>>>>>>>>______________________________________________________________
___
>>>___
>>>>>>___
>>>>>>>>>Hans-Peter Piepho
>>>>>>>>>Institut f. Nutzpflanzenkunde WWW:
>>>>>> http://www.wiz.uni-kassel.de/fts/
>>>>>>>>>Universitaet Kassel Mail: piepho@wiz.uni-kassel.de
>>>>>>>>>Steinstrasse 19 Fax: +49 5542 98 1230
>>>>>>>>>37213 Witzenhausen, Germany Phone: +49 5542 98 1248
>>>>>>
| 
