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: RE: RE: RE: missing standard error in multinomial logit


From   "Macquet, Laure (Student)" <macqulmn@aston.ac.uk>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: RE: RE: RE: missing standard error in multinomial logit
Date   Wed, 8 Aug 2012 20:37:47 +0000

Hello Dave, thanks for answering !

Well I thought of another variable that would be NATURE coded as follows:
1 if online firm
2 if brick and mortar firm
3 if the observation is a greenfield transaction

in that case, there is no more problem with mlogit. (if I don't mention the fact that all of a sudden all my z-statistic for Greenfield equal 0!)

But if I want to get the marginal effects with mfx command, how can I deal with NATURE? I'm not sure but is what is called a categorical variable? Therefore in that sense, I guess I cannot interpret the results of the marginal effects saying "should NATURE increase by one unit, then the marginal effect on the dependent variable is..."

I'm not sure I made myself clear enough...

Thanks again in advance !
________________________________________
From: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] on behalf of Jacobs, David [jacobs.184@sociology.osu.edu]
Sent: 08 August 2012 19:45
To: 'statalist@hsphsun2.harvard.edu'
Subject: st: RE: RE: missing standard error in multinomial logit

One indicator of problems with dummies in such an analysis is extremely large coefficients on one or more dummies particularly if you ask for odds ratios.

In any case, if you're right that it is a problem with the dummies, is there any way you could recode that would make sense?

Dave Jacobs

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Macquet, Laure (Student)
Sent: Wednesday, August 08, 2012 2:08 PM
To: statalist@hsphsun2.harvard.edu
Subject: st: RE: missing standard error in multinomial logit

Hello all,

I am trying to estimate the impact of some specificities of e-commerce firms on their choice for entry mode through a multinomial logit regression.

The entry modes are the dependent variable: Acquisition, Joint Venture, Partnership, Greenfield

In my independent variables, amongst others, I have two dummies:
WEB (1 if the acquired/partnering firm is an online firm, 0 otherwise) BM (1 if the acquired/partnering firm is a brick and mortar firm, 0 otherwise)

When I run the regression with all the independent variables, the number of iterations is really big (around 45) and some lines within the table have missing values except from the coefficient. I suspect the two dummies to be the issue but I cannot really get rid of any of them...
Indeed, WEB=0 does not necessarily means that BM=1 for the same observation since in the case of Greenfield, WEB=BM=0 (there is no acquired/partnering firm).

Can anybody help me on this?

Thanks very very much !

Laure












*
*   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/

*
*   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/





*
*   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