impute_describe

[Julia Hood]

Hello!

When I run the code pasted below, the output matches a complete case analysis.  I strongly suspect that the output does not utilize the imputed data.  Would you please review my code and let me know if you see any errors? 

 

Thanks!

Julia 

 

data _null_;

infile datalines;

filename setup "impute_describe.set";

file setup;

input;

put _infile_;

datalines4;

title Multiply imputed design-based statistical description;

datain feb.mi_ckd;

dataout feb.ckd_imp;

* impute keywords;

default transfer;

continuous age gfr_24sp bmi ;

categorical sex race_eth agecat3 ckd;

iterations 10;

multiples 5;

seed 100;

weight cnics_wt;

* describe keywords;

estout desexam2;

model mult;

table ckd;

/* contrast sex*race_eth*agecat3 ;*/

run;

;;;;

%describe(name=impute_describe, dir=.);

[Trivellore Raghunathan]

I think you are missing the mdata statement. The default is skip (complete case analysis)

Raghu

--
You received this message because you are subscribed to the Google Groups "IVEware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iveware+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Julia Hood]

 

You're right!   I was missing the mdata statement.  Thank you for catching this oversight!! 

[Julia Hood]

Hi,

I've run into another issue with the code.  I get the error message "bad strata" alongside the "abnormal termination of describe"  message in the log.  Several strata are flagged in the log as having only one cluster.   Do you know how I can get around this issue?

 

Thanks,

Julia

--
You received this message because you are subscribed to a topic in the Google Groups "IVEware" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/iveware/KTgvKmxAWZE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to iveware+u...@googlegroups.com.

[Trivellore Raghunathan]

You may have to combine some strata to have at least 2 clusters in the combined strata.

Another conservative approach is not to use stratification.

Raghu

[Julia Hood]

Thanks so much for your assistance!  I'm sorry, but I'm still having issues with impute_describe. 

 

After a close look at my data, I discovered that all of the strata have >2 clusters.  Note, the "impute" code/macro runs fine with my design variables.  The error message "Bad Strata" only appears when I try to run the impute_describe code. 

 

Do you have any idea of what the issue might be?