topic Re: how to get pooled median and q1 and q3 using Proc mianalyze in Statistical Procedures

how to get pooled median and q1 and q3 using Proc mianalyze

jojo — Wed, 05 Mar 2025 07:19:09 GMT

after proc means procedure by imputation based on 50 imputed dataset, I need to get the pooled results for means, median, q1, q3 from proc mianalyze. Is the following code correct? e.g. use the same standard error for mean, q1, q3.

proc mianalyze data= ccc;

by treatment;

modeleffects mean_val q1_val q3_avl;

stderr std_val std_val std_val;

run;

Re: how to get pooled median and q1 and q3 using Proc mianalyze

Ksharp — Wed, 05 Mar 2025 08:20:39 GMT

Nope.

Different quantile have different stderr.

data Orig;   /* restrict to 5000 births; exclude extreme weight gains */
set Sashelp.BWeight(obs=5000 where=(MomWtGain<=40));
run;

title;
ods noproctitle;
ods select  ParameterEstimates;
proc quantreg data=Orig ci=sparsity; 
   model Weight =   / quantile = 0.25 0.5 0.75;
run;

Re: how to get pooled median and q1 and q3 using Proc mianalyze

jojo — Sun, 09 Mar 2025 17:22:42 GMT

Thanks! So I need to use proc quantreg to get the standard error to each quantile and then use proc mianalyze to get the pooled q1 and q3.