Hello All,
I have a problem of collinearity in my data. One of my binary variables (Nairobi) in a set of four regional variables (Nairobi, Other urban, Central rural, Other rural) is dropped. Thus, when I run:
xi: mlogit cuse i.age1525 i.age2630 i.age3135 . . . . . i.Nrb i.Oturb i.Centrur i.Othrur [pweight=wv005], cl(v001) basecat(3)
outreg using test.doc, coefastr
I get the response:
(sum of wgt is 1.8791e+03)
note: _INrb_1 dropped due to collinearity.
But I do not want this variable to be dropped; instead I exclude the variable for other rural areas from the final regression. My reasoning is that other rural areas would in any case be negatively associated with my outcome variable (which I confirmed).
So I seek your assistance:
1. Am I right?
2. What other options do I have to do it right? For example, I notice that when I include the command "noconst", all the variables are included and none is omitted due to collinearity. Is this recommended?
Thx.
*
* 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/
