Dear viewers,
I am working on a conditional logit regression.The subject is the locational
choice of MNEs' new plants in 16 European countries.The dependent variable is
the choice, and independent variables are country level chanracteristics,like
GDP and corporate tax rate...........I add a dummy variable to distinguish the
plant of US MNEs(=1) or non US MNEs.Since Stata Clogit's requirment of data
structure, this dummy variable is invariant for each plant in the data.It
appears in the dataset like this:
location Choice Dummy ........
1 0 1
2 0 1
3 1 1
...............................................
16 0 1
I hope you could understand what I am trying to describe.
Then I run the regression with dummy entering as an independent variable(no
interaction term at this stage).But Stata claims that "note: dummy omitted due
to no within-group variance." So I am wondering if there is a way to use this
dummy in the conditional logit model?
First I try to interact choice(depedent variable) with dummy to give this dummy
a kind of variation.The results look ok but I am worrying if it is a correct
way.
Then I try to interact the dummy with all other independent variables.But
according to the principle of marginal, I still need to put dummy(main effect)
in the regression and this variable will be dropped by Stata.
So any one could help me on this point?
Thanks in advance!
Kevin in Dublin
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
