Programming the statistical procedures from SAS

resampling with proc GLM (or any analytical proc)

Reply
Occasional Contributor
Posts: 9

resampling with proc GLM (or any analytical proc)

Hello group:

 

I'm trying something new here and running into the usual programming snafu.

 

step 1 - use proc surveyselect to create several replicates of a dataset

 

step2 - use proc glm and ods to obtain a file of parameter estimates. Here's the code which works without the "by" clause.

 

ODS trace on;
ods output glm.anova.t2pcl_total.ParameterEstimates=foo.bar1;
proc glm data=boot1 outstat=foo.bar;
class cage (ref=first) ceduca (ref=last) cincome crace sex (ref=first);
model t2pcl_total = sex cage ceduca cincome crace /*_INTO_*/ overall_score /solution;
weight tot_weight;
by replicate; /* works without this on a single file */
run;
ods trace off;

 

When computing one file the output "glm.anova.t2pcl_total.ParameterEstimates" gets created just fine but with multiple

replicates it creates a path like "glm.bygroup3.anova.t2pcl_total.ParameterEstimates" - this is clear from the trace in the

attached log file. See error message in attached file also.

 

I'm thinking the answer might lie in macro programming but that's just a guess. Any help greatly appreciated.

 

Thanks,

OldSASguy100

 

 

 

 

Respected Advisor
Posts: 4,606

Re: resampling with proc GLM (or any analytical proc)

[ Edited ]

Just use

 

ods output ParameterEstimates=foo.bar1;

 

and everything, including the by variable, will be output to foo.bar1.

PG
Occasional Contributor
Posts: 9

Re: resampling with proc GLM (or any analytical proc)

Thanks, worked like a charm.

##- Please type your reply above this line. Simple formatting, no
attachments. -##
Ask a Question
Discussion stats
  • 2 replies
  • 109 views
  • 2 likes
  • 2 in conversation