Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Multinomial logit and sensitivity analysis


From   Frank Talony <frank.talony@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Multinomial logit and sensitivity analysis
Date   Sun, 8 Jul 2012 22:05:09 -0400

I am performing a multinomial logit test. The data set does not include the
dependent variable and it had to be generated from the data itself with
some conditions.

First I generate indicator variables id1, id2, id3 based on independent
variable (v1, v2, v3, ...).
For example:
id1=1 if (v1 > 100), 0 otherwise
id2=1 if (v1 < 20 AND v2>10), 0 otherwise
id3=1 if (v3 > 1.4*v2), 0 otherwise

Then the dependent variable is computed from the indicators as follow:
dep1 = 1 if (id1 OR id2)
dep1 = 2 if (id3)
dep1 = 0 otherwise

My logit is specified as:
mlogit dep1 v1 v2 v3 v4 .... vn, level(90)

Is it a valid test to regress v1,v2,v3 on dep1?
Can I perform some test (perhaps sensitivity analysis) to support using
v1,v2,v3 in my results?

The fact is the results is less interesting without v1,v2,v3 but I am not
sure if they are valid inclusions.
Thanks,

Frank
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index