Model Selection in SAS Enterprise Guide and SAS Enterprise Miner - Ask the Expert Q&A

[Olfatshiekh]

Hi,

My research question is to assess the influence of grandparents on grandchildren's health.

My research question is to assess the influence of grandparents on grandchildren's health.

My outcome is binary: Good health vs. Poor health.

My main predictor is presence of older adults (Yes or No)

however, the data was collected through proxy adults on all the household's members, so one person is giving all the information about all the members living in the same house and there are many children in one household.

So, I think that I need to adjust for the household, since many children living in the same household, using fixed effect logistic regression, is that right? Or do I need to adjust for the grandparents who is available in the household and may have an influence on the children's health living in the same household?

I'm not sure how to do such adjustment if I need to, is this can be done using PROC GENMOD or PROC LOGISTIC? Which one is better? and can you provide me with the SAS code?

Thanks,

Olfat

[Ksharp]

I think you need Generalize Linear Mixed Model.

Check PROC GLIMMIX .