Re: st: cox regression - half siblings

[Grethe Søndergaard <dkstatstata@gmail.com>]

Thank you for your answer

I agree that the composition of families can be quite complex in that
both half and full siblings can appear within the same family.

Following up on your example, my plan is to include child 1 and 2 in
the full siblings dataset and all three children in the half siblings
dataset. However, the problem is that I only want to perform the half
siblings analysis on the children who indeed are half siblings, i.e.
´compare´ child 3 with child 1 and 2.. But is that possible? Or can I
only include half siblings, and no full siblings, in this dataset in
order to do this?

Thank you



2010/8/25 Maarten buis <maartenbuis@yahoo.co.uk>:
> --- On Wed, 25/8/10, Grethe Søndergaard wrote:
>> I have a dataset that contains information on a population
>> born in 1980-1990. Some of the persons in this dataset are
>> siblings (both half siblings and full siblings). I want to
>> perform cox regression on the entire population as well as
>> on populations of siblings,
>> i.e:
>> - a population of individuals who are full siblings
>> - a population of individuals who are half siblings (same
>> mother,
>> different fathers)
>> - a population of individuals who are half siblings (same
>> father,
>> different mothers)
>
> What about the population who have no siblings, the families
> with one child...
>
>> The dataset includes the following variables:
>> Person_ID
>> Mother_ID
>> Father_ID
>> Health indicators
>> Education
>>
>> I am going to use stcox to analyse my data. When I do the
>> analyses based on siblings, I use "strata" in order to
>> stratify the analysis on siblings. When doing the analyses
>> for full siblings I write:
>>
>> stcox i.education, strata (mother_ID father_ID)
>>
>> But what do I do, when performing the analyses on e.g. half
>> siblings
>> with the same mother but different fathers? If I only
>> stratify on mother_ID I will include all children with the
>> same mother, meaning both full siblings and half siblings
>> (with the same mother). So how is it possible to perform
>> this analysis on half siblings only?
>
> Say a family has three kids, two from the same mother and
> father and a third from another father, how do you define
> which child should appear in which dataset? You could add
> child 1 and 2 to the full sibling dataset and to the half
> sibling dataset (as they are half-siblings of child 3).
> That does not seem like a good solution, as now you are
> using the same observation multiple times.
>
> The way to solve this depends on the question and the
> theory and the data. So it can only be solved by the
> researcher.
>
> Hope this helps,
> Maarten
>
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://www.maartenbuis.nl
> --------------------------
>
>
>
>
>
> *
> *   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/
>

*
*   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/