Date: Wed, 24 Jun 2009 16:46:56 -0700
Reply-To: Dale McLerran <stringplayer_2@YAHOO.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Dale McLerran <stringplayer_2@YAHOO.COM>
Subject: Re: getting parameter std errors out of proc reg
Content-Type: text/plain; charset=iso-8859-1
--- On Wed, 6/24/09, batgrrl <unquenchable.fire@GMAIL.COM> wrote:
> From: batgrrl <unquenchable.fire@GMAIL.COM>
> Subject: Re: getting parameter std errors out of proc reg
> To: SAS-L@LISTSERV.UGA.EDU
> Date: Wednesday, June 24, 2009, 10:21 AM
> (I evidently sent this only to Steve
> so will post it here as well)
> Thanks so much for your help. I am at my wit's end and trying to get
> this working before my thesis defense, and I am grateful for any
> coaching I can get. I am running SAS 9.2 on Windows.
> I tried your suggestion and get an error saying "Output
> "parameterestimates" was not created. Make sure that the output
> object name, label, or path is spelled correctly. Also, verify that
> the appropriate procedure options are used to produce the requested
> output object. For example, verify that the NOPRINT option is not
> used." Changing the output name to parmest (in case parms was somehow
> reserved) had no effect.
> I found an earlier sas-l email that said you need to be sure to quit
> any earlier processes. I tried this (on my proc import) and still get
> the error. Here is my code, cut and pasted in:
> PROC IMPORT OUT= WORK.BATS
> "C:\Documents and Settings\Data Repository\Desktop
> \Jennifer's thesis data\JKModelDataCorr.csv"
> ods trace on / listing;
> ods output parameterestimates=parmest;
> proc reg data=bats outest=est tableout;
> model tactivity = effort size edge distH2o distpark nativepl /
> include=1 selection=adjrsq sse aic;
> title "Total Activity All Species Full Model";
> output out=out p=p r=r;
> ods trace off;
> With trace on I don't see that ParameterEstimates is even created.
> The only data sets that trace reports are Nobs and SubsetSelSummary.
> That would the explain the error message, but why isn't it created?
> Moving the ODS Output command to different locations doesn't make any
> difference. Removing TABLEOUT makes no difference. The outest data set
> is created, but not with the full info that TABLEOUT should include.
> I'm utterly unable to find any references online to tableout
> malfunctioning, or parameterestimates missing.
The reason you are not generating a ParameterEstimates table is
due to use of the SELECT=ADJRSQ option. This is an unintended
consequence of using a model selection process.
Is there a specific model for which you want the parameter estimates?
Then invoke PROC REG with only the variables that belong in that
model and with no SELECT= option. You will then get a
ParameterEstimates table. For instance,
ods output parameterestimates=parmest;
proc reg data=bats outest=est tableout;
model tactivity = effort size edge distH2o distpark nativepl;
title "Total Activity All Species Full Model";
output out=out p=p r=r;
ods trace off;
will produce parameter estimates for the full model (with all 6
predictor variables in the model).
Fred Hutchinson Cancer Research Center
Ph: (206) 667-2926
Fax: (206) 667-5977