Standard errors from DISTRIB macro

Reply
Regular Learner
Posts: 1

Standard errors from DISTRIB macro

I have using two macros to analyze some complex survey data, MIXTRAN and DISTRIB macros, and from the user guide in order to generate standard errors I need to incorporate values into "vcontrol=, and start_val2=," if I am running an amount model. I have 16 replicate weight variables and my code for the mixtran and distrib macros are below. The user guide is also attached to this post. 


%mixtran (data=example,
response=DRTVK,
modeltype=amount,
subject=seqn,
repeat=day,
outlib=mylib,
covars_amt= DRTKCAL,
covars_prob=,
foodtype=VK_MALES_edu,
seq=day2,
lambda=0.4,
replicate_var=repwt_1,
weekend=weekend,
vargroup=,
numvargroups=,
vcontrol=,
start_val1=,
start_val2=,
start_val3=,
nloptions=qmax=61,
titles=4,
printlevel=2);
run;

 

title4 "Output from the DISTRIB macro calculating usual intake PK for FEMALES by edu";

%Distrib (call_type=Full,
modeltype=amount,
outlib=mylib,
subject=seqn,
mcsimda=mylib._mcsim_VK_MALES_edu_Repwt_1,
pred=mylib._pred_unc_VK_MALES_edu,
param=mylib._param_unc_VK_MALES_edu,
seed=5454768,
nsim_mc=100,
food=DRTVK_MALES,
byvar=, /* MIXTRAN was not fit separately for by groups */
cutpoints= 10 12 14 17 22 25 33 38 42 47,
ncutpnt=10,
wkend_prop=3/7,
add_da=example,
subgroup=edu,
recamt=VK_AI,
recamt_co=ge,
recamt_hi=, /* comparison is not a range */
titles=4);

**** note: The parameters byvar and recamt_hi can be omitted in this call since they will not be used. ***
**** They are in the example only for the sake of documentation. ***;

run;

 

The macros are able to run successfully, and the output for the DISTRIB macro provides the mean, and various percentiles, I also need the standard error. 

 

Any help will be greatly appreciated.

 

Thank you,

 

Stephanie

Ask a Question
Discussion stats
  • 0 replies
  • 109 views
  • 0 likes
  • 1 in conversation