Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: AIC for goodness of fit and gllamm in stata?
From
Jessi Bishop-royse <[email protected]>
To
[email protected]
Subject
st: AIC for goodness of fit and gllamm in stata?
Date
Mon, 7 Mar 2011 23:17:54 -0800 (PST)
Hello statalisters!!
I am trying to determine goodness of fit using gllamm. I include an example
below, where I'm modeling the outcome (premsurv) on an individual level
indicator of race (black) which is binomial, along with 4 other level two
indicators which are also binomial (perblack, poor_poor, hispanic and rural). I
have seen some mentions of using AIC elsewhere, but wanted to get a sense of
whether this was appropriate or not?
. gllamm premsurv black perblack_1980 hispanic_1980 poor_poor1980 rural_1980,
i(countynumber) family(binom) link(logit)
Iteration 0: log likelihood = -3510.1953 (not concave)
Iteration 1: log likelihood = -3502.5874
Iteration 2: log likelihood = -3501.9328
Iteration 3: log likelihood = -3501.8699
Iteration 4: log likelihood = -3501.8698
number of level 1 units = 114954
number of level 2 units = 66
Condition Number = 8.2279657
gllamm model
log likelihood = -3501.8698
------------------------------------------------------------------------------
premsurv | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
black | .7542904 .0874572 8.62 0.000 .5828775 .9257034
perblack_1~0 | .1151921 .1426495 0.81 0.419 -.1643958 .3947801
hispani~1980 | -.0600987 .1453882 -0.41 0.679 -.3450544 .2248569
poor_po~1980 | -.1729206 .2809316 -0.62 0.538 -.7235365 .3776953
rural_1980 | -.0541847 .21655 -0.25 0.802 -.4786149 .3702454
_cons | -5.666236 .1031606 -54.93 0.000 -5.868427 -5.464045
------------------------------------------------------------------------------
Variances and covariances of random effects
------------------------------------------------------------------------------
***level 2 (countynumber)
var(1): .06861259 (.0426707)
------------------------------------------------------------------------------
Thanks so much in advance!
Best, Jessi Bishop-Royse
*
* 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/