I am trying to estimate a Multinomial Logistic model and it seem to me one of the explanatory variables is endogenous. Is there any way to test it?
Any help will be appreciated
Thanks you
Hi: I don't know much about the area but the link below looks possibly useful. Not sure if there such a test in R but you didn't specify the language that you use.
http://www.indiana.edu/~jslsoc/files_research/rprints/STBmlogtest.pdf
There were a lot of other things that came up also so my google was "testing for endogeneity in multinomial logistic". Good luck and hopefully there will be better answers than this.
Hi. Dont know what software are you using, however, use the instruments must be your first choice. Cheers,
Hi Sergio: Heckman 's test involves the use of instruments so I think we are on the same page.
The Regressor Endogeneity Test, also known as the Durbin-Wu-Hausman Test, tests for the endogeneity of some, or all, of the equation regressors. This test is available for non-panel equations estimated by TSLS or GMM.
A regressor is endogenous if it is explained by the instruments in the model, whereas exogenous variables are those which are not explained by instruments.
You can find this test also in EViews and in Stata.
What kind of data do you use?
Thanks all for your valuable response and intervention. I am using Stata. My data is cross-section - household survey
Hi Ebaidalla,
as it is said above, there is need to instrument if you want to test for regressor endogeniety, hope you managed to do so ..otherwise, the hausman test statistic compares OLS an IV estimators....if you are using stata you can get all IV commands using help command. Hausman test is implemented with the postestimation estat endogenous command following the ivregress 2sls command
good luck
Hi Ebaidalla,
as already pointed out, I believe you will need to find a good instrument to test for endogeneity. Since you are working with a non-linear model, I would rather suggest applying control function approach instead of 2SLS (Imbens/Wooldridge offer great lecture notes on the method) - it allows you to test for endogeneity while remaining in the logistic framework.
However, this convenience comes at the cost. The required form of exogeneity of the instrument is stronger and more restrictions are imposed on the reduced forms compared to 2SLS so one needs to be careful there.
You can easily implement the approach with any software without any special packages, you will only need to bootstrap to get the proper standard errors.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey