st: multilevel analysis, ICC and MOR

[yori novrianto <yorinovrianto@gmail.com>]

Hello everybody,

my name is Yori, I'm using Stata 11.0 version. I have a question
regards to the ICC and MOR in multilevel analysis. my thesis is about
stunted-child-overweight-mother in urban and rural region. For my
modeling I differentiate between urban and rural region. I’ve run the
empty model with the following result, the MOR for urban was 7.85,
rural was 11.60.

In the rural model, when I started include the child characteristics
the MOR suddenly become 1.00003. Variables in child characteristics
consist of: child’s sex, child age, child birth order, breastfeeding
status, age of weaning, pre-lacteal feeding practices, age of
introduction complementary feeding.

I try to run each variable independently, with the result: child age
with the MOR 13.9998, child’s sex 11.611 and child birth order 11.84.
Whenever I include/ combine child age category with other variables
suddenly the MOR become 1.0003. I’ve checked the child age variable,
there is no missing value and there is no outlier.

Do you have any idea about this problem, because i think it is just
not make sense if the ICC suddenly became one while i just include
three variables in my model.

Regards,

Yori

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