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 at the end of May, and its replacement, statalist.org is already up and running.


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

st: Telling what is significant in mprobit


From   Laura Cyron <lauracyron@yahoo.de>
To   statalist@hsphsun2.harvard.edu
Subject   st: Telling what is significant in mprobit
Date   Sun, 25 Jul 2010 23:17:10 +0000 (GMT)

Hello Statalist,

I am currently trying to evaluate a multinomial probit model regarding the usage 
of mutually exclusive financial service options with the help of -mprobit-. And 
I wish to give the marginal effects at the means in the written data analysis. 


Now I am confused how to tell which independent variables are significant. 
Normally, I take a look at P > |z| of the coefficients and check whether the 
values are below the alpha value of, for instance, the significance level of 10% 
(so, a value > 0.1).

A simplified (and shortend) version of my code is as follows:

mprobit fin_services age female log_expenditure risk_attitude
mfx2

(where fin_services takes on 8 different values)

In the first very basic test runs I undertook using -mfx2-, I stumbled  across 
the problem that some of the P > |z| values of the coefficients in mprobit 
indicate that "female" and "log_expenditure" are significant for fin_services 
category 4.
Then I ran -mfx2- and in that output P > |z| was only significant for "female" 
and no longer for "log_expenditure" when Pr(fin_services==4).

What does this tell me? Or rather am I making a mistake here?
And which coefficients would than be the significant one which I could mention 
in my analysis. 


I would greatly appreciate your input.

Thank you,
Laura



*
*   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