st: including the same variable as fixed and random in xtmixed

[José Maria <jmpsouza@usp.br>]

dear Statalisters:
   maybe this is more a statistical question rather a Stata one, but it is 
related to the use of the command xtmixed. I will use the example 4, on 
pages 293-294 of the Longitudinal/Panel data manual, for release 10.
   Let's admit that that the variable state covers *all* the states of the 
country and the variable region represent the nine official regions of the 
country. So, the categories of these variables represent the total of 
possibilities for each one and for the combination, that is, they are really 
the population.   In this case, would it make sense to include both  in the 
equation (as dummies), writing
    xtmixed gsp private ...  state2 state3 state4... region2 region3 
region4...||region: || state:
assuming that there is also interest in comparing the means of gsp among 
states and among regions?
   I know that the decision involves a question of number of parameters and 
parsimony. But, if to compare the means is important, what is the best 
statistical way?
Thank you,
José Maria

Jose Maria Pacheco de Souza, Professor Titular (aposentado)
Departamento de Epidemiologia/Faculdade de Saude Publica, USP
Av. Dr. Arnaldo, 715
01246-904  -  S. Paulo/SP - Brasil
fones (11)3061-7747; (11)3768-8612;(11)3714-2403
www.fsp.usp.br/~jmpsouza 


__________ Informação do ESET NOD32 Antivirus, versão da vacina 4613 (20091116) __________

A mensagem foi verificada pelo  ESET NOD32 Antivirus.

http://www.eset.com



*
*   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/