I've been trying to get my parameter estimates into a more standardized format, specifically where all I have is the coefficient with some sort of indicator of whether it is significant at the 1, 5 or 10% level, all above the standard error.
I've gotten something similar with the following code, but the output is not excel/word friendly, a word table being the ultimate destination of the results:
picture stderrf (round)
low-high=' 9.99)' (prefix='(')
ods output ParameterEstimatesF(persist) = t;
proc robustreg data=work.auto method = lts fwls;
model &input = &model1a;
title 'MODEL 1_A';
retain modelnum 0;
if parameter="Intercept" then modelnum = modelnum + 1;
proc tabulate data=r2 noseps;
class modelnum parameter;
var estimate stderr;
table parameter=''*(estimate =' '*sum=' '
stderr=' '*sum=' '*F=stderrf.),
/ box=[label="Parameter"] rts=15 row=float misstext=' ';
I will be running more than one regression in this with some of the same variables in each, so any advise on how to get the results and combine them into one table would be greatly appreciated.