I'm no expert; I would use the ODS statements, then manipulate the output in data steps. Something like this, maybe it's helpful for what you need:
*Create a dataset with the SAS output*;
ods output ParameterEstimates=parest;
proc reg;
model weight=height age; *INSERT YOUR OWN MODEL HERE*;
run; quit;
*Given a model with n explanatory variables, the dataset PAREST will have n+1 records*
*The n explanatory records + Intercept*;
*Prepare the output for outputting to Excel*;
data prepared;
set parest;
*For each record: Output one line with Estimate/StdErr*;
var1=Estimate;
var2=StdErr;
output;
*And another line with t-stats*;
var1=tValue;
var2=Probt;
output;
run;
*dataset PREPARED will have 2(n+1) records*;
proc print data=prepared;
var Variable df var1 var2;
title "Is this better?";
run;





Reply With Quote
