Statalist


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

Re: st: loglikelihood and loglikelihood ratio


From   jjc.li@utoronto.ca
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: loglikelihood and loglikelihood ratio
Date   Tue, 17 Mar 2009 10:18:17 -0400

Hi Maarten,

1. My loglikelihood is positive, not negative. Sorry for the mistake in the previous email.

2 A is the unrestricted model, B is the restricted model. Following is the test result. What's wrong with it?


. lrtest A B
(log-likelihoods of null models cannot be compared)

Likelihood-ratio test                                  LR chi2(3)  =     10.94
(Assumption: B nested in A)                            Prob > chi2 =    0.0121

Jingjing



Quoting Maarten buis <maartenbuis@yahoo.co.uk>:


--- On Tue, 17/3/09, jjc.li@utoronto.ca wrote:
I used -sureg- to estimate a system.

1. After using -display e(ll)- to show the loglikelihood,
it gives a negative value. But I remember I have saw a PPT
says loglikelihood is a typical negative value. Is mine
abnormal?

All the information you have given us is that your log
likelihood is negative. That is not abnormal, but a positive
log likelihood would not be a reason for concern either, so
all this tells you exactly nothing.

2. In order to compare unrestricted and restricted models,
I used -lrtest- to do the loglikelihood ratio test. The LR
chi2() value is negative.Is it normal? What does it mean?
Restricted model better or unrestricted model better? If
model one LR chi2()=-1, model LR chi2()= -10, which one to
choose?

This is not normal, so it means you have done something
wrong. Are you sure that you are comparing comparable models?

3. Is loglikelihood and loglikelihood ratio different?

yes, log liklihood is the log liklihood of a single model,
while the log liklihood ratio is the ratio of log likelihoods
of two different models.

-- Maarten

-----------------------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://home.fsw.vu.nl/m.buis/
-----------------------------------------







*
*   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   |   What's new   |   Site index