st: Failure time analysis- dependent data coding

["Oliver M. Beckers" <beckersom@unlserve.unl.edu>]

Hello,

I seek help with a failure time analysis I am trying to do with Stata.

Here is my experimental design:
From each female I got multiple juvenile offspring. These juveniles  
were split into two groups and each group was subjected to one  
treatment. Within each treatment, I measured how long each of the  
juveniles survived. I repeated this procedure with offspring of  
multiple (independent/unrelated) females.

Here is an example with numbers:
I got 7 offspring from female #1. I split the offspring in one group  
of 3 and one group of 4. The first group was subjected to treatment  
one (dry) and the second group to treatment 2 (wet). I measured how  
long each of the 7 individuals survived. I did the same thing with the  
offspring of a total of 5 females.

I want to measure if survival differed between the two treatments (dry/wet).

The problem:
I have two levels of dependence among my data and I am not sure if my  
procedure (stata command and arrangement of the data) accounted for  
the dependences of the data:
1.) I have multiple full siblings within each treatment,
and 2.) I have multiple full siblings between the two treatments.

I used the "stcox treatment, shared (female)" command, and assigned  
all offspring from the same female that were used in both treatments  
the same "female" number to account for dependence of data. So, for  
the above example all 7 offspring were assigned the same female number  
(#1). Furthermore, I assigned for each juvenile in separate columns a)  
the treatment, b) juvenile number, c) time till death, and d) censored  
(1/0; some juveniles survived longer than the maximum time of our  
observations).

My question is:
does this command/arrangement of data account for the within and  
between treatment dependence (frailty) of the data, or not? If not,  
what would be the correct command and data arrangement in Stata to  
compare the survival between the two treatment groups?

Thank you very much for any help,
Oliver Beckers



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/