Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: rank of differenced matrix while testing IIA

From   "R.E. De Hoyos" <[email protected]>
To   <[email protected]>
Subject   st: Re: rank of differenced matrix while testing IIA
Date   Thu, 26 May 2005 10:40:13 +0300


Try using S. Long's post estimation command -mlogtest-. It computes Hausman and Small-Hsiao tests for IIA. If not installed type:

findit spostado

It will lead you to the ado's.

You are actually trying to compare parameters between a -mlogit- against a logit (two options). If you want to carry on the hausman test your self, the coding would have to be a bit more complicated. You will have to tell Stata which equation (out of the two that are being estimated by -mlogit-) you want to compare with the logit.

I hope this helps,

R.E. De Hoyos
Faculty of Economics
University of Cambridge

----- Original Message ----- From: "Alexander Nervedi" <[email protected]>
To: <[email protected]>
Sent: Thursday, May 26, 2005 9:52 AM
Subject: st: rank of differenced matrix while testing IIA

Hi Statlist users:

I am trying to check the IIA assumption for a case in which a patient has three options - full medication (2), some medication (1) and no medication (0). Numbers in brackets indicate how they have been coded in the data.

i specify the test as:

mlogit medlevel indepvariables, basecategory (1)
estmiate store fullset
mlogit medlevel indepvariables if medlevel !=0, basecategory(1)
estmiate store subset
hausman . fullset

the models are estimated and the test is run but with a warning message:

Note: the rank of the differenced variance matrix (6) does not equal the
number of coefficients being tested (7).

I was wondeirng what this really means - with the sub-model i am just down to just two alternatives, whereas with the main i have three - should this result in these kinds of warning messages - or is there something more serious that may be wrong?

many thanks

Alex Nerd.

Express yourself instantly with MSN Messenger! Download today - it's FREE!

* For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index