Statalist


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

Re: st: deviance test


From   Steven Samuels <sjhsamuels@earthlink.net>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: deviance test
Date   Sun, 21 Dec 2008 15:38:56 -0500

Maarten is absolutely correct.

The deviance for model X is: (log-likelihood for saturated model) - (log likelihood for model X)

This makes the deviance for the saturated model equal to zero. The likelihood ratio test statistic for comparing models X & Y (nested in X) is identical to the Deviance test statistic for comparing the same 2 models:

Deviance for Y - Deviance for X = (log-likelihood for X) - (log- likelihood for Y)


-Steve



On Dec 21, 2008, at 9:03 AM, Maarten buis wrote:

--- Sebastián Daza <sebastian.daza@gmail.com> wrote:
I need get deviance-based hypothesis test (df) with stata (10) for a
logistic model:

Aren't deviance-based tests just a variation on likelihood ratio tests?
In that case, see -help lrtest-.


*
*   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