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

st: No model significance test result for the 2nd model equation in MLE?

From   Tao Xiao <[email protected]>
To   [email protected]
Subject   st: No model significance test result for the 2nd model equation in MLE?
Date   Wed, 16 Sep 2009 11:56:46 -0400

I am sorry to bother but I really cannot solve this problem of MLE.

I need to specify more than one model equations in "ml model" command.
However, after the MLE routine ran to reach the maximum likelihood
with the data, the "ml display" command can only display model
significance test result of the 1st model equation in the header above
the coefficient table - a Wald chi2 statistic value and the
corresponding p-value. Because in our research, the model significance
test result of the 2nd model equation is also important, I wonder if
there is any way to display the model significance test results of the
second model equation, or of more than two model equations. Or, can I
get model significance test result for all the model equations
combined together, not just for the 1st model equation?

This problem can also be seen in the book "Maximum Likelihood
Estimation with STATA, 3rd ed."  by William Gould, Jeffrey Pitblado
and William Scribney. On page 37 of the book, we can see an example
with 2 equations: "mu" and "sigma". 2 predictors ("weight" and
"displacement") are included in equation "mu" and 1 predictor is
included in equation "sigma". The "ml" commands in this example are:

sysuse auto, clear
ml model lf mynormal1_lf (mu: mpg=weight displacement) (sigma: price)
ml maximize, nolog

In the header of the coefficient table, we can see:

Wald chi2(2) = 126.71.
Prob > chi2 = 0.0000

The chi2 test above with 2 degree of freedom is corresponding to the
"mu" model equation with 2 predictors specified. However, the test
result corresponding to "sigma" equation is missing.

I also studied the options in "ml display" command (on page 163 of the
MLE book above) but none of them can display complete results I am

Does anybody have an idea? Thanks in advance!
*   For searches and help try:

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