I am running a fixed effect survival model and am having a couple of unexpected results.
1. STATA drops several of the FE variables, indicating collinearity. I'm not sure how this is the case, given that the fixed effect variables, vsa2, are area dummies generated from the xi command. The minimum number of observation in an area is 130 and there are 1000 "areas".
2. Quite a few of the FE dummies have "missing" upper bounds for the confidence interval. Should I be worried about this?
Below is an excerpt from the log file. I appreciate any suggestions and insights you can give.
. xi: streg frmrsmokr crrntsmokr drink12peryr_90 lttlstress1yr medstress1yr alotstress1yr exercise $racegroup $indiv_contrl i.vsa2, dis(gompertz) fr(gamma)
note: _Ivsa2_58 dropped due to collinearity
note: _Ivsa2_197 dropped due to collinearity
note: _Ivsa2_348 dropped due to collinearity
note: _Ivsa2_426 dropped due to collinearity
note: _Ivsa2_469 dropped due to collinearity
note: _Ivsa2_510 dropped due to collinearity
No. of subjects = 5537 Number of obs = 5537
No. of failures = 150
Time at risk = 487048
LR chi2(1018) = 815.22
Log likelihood = -392.14229 Prob > chi2 = 1.0000
------------------------------------------------------------------------------
_t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
frmrsmokr | 1.877704 .56903 2.08 0.038 1.03675 3.400792
crrntsmokr | 3.235288 .9527995 3.99 0.000 1.816486 5.762271
drink12pe~90 | 1.72386 .5578645 1.68 0.092 .9142004 3.250591
********
_Ivsa2_984 | 2.27e+10 3.87e+10 13.96 0.000 7.97e+08 6.44e+11
_Ivsa2_985 | .0001834 1353.748 -0.00 1.000 0 .
_Ivsa2_986 | 1.03e+09 1.78e+09 12.06 0.000 3.54e+07 3.02e+10
_Ivsa2_998 | .0001937 1403.955 -0.00 1.000 0 .
_Ivsa2_999 | .0021168 12262.63 -0.00 1.000 0 .
_Ivsa2_1000 | .0004532 2977.283 -0.00 1.000 0 .
-------------+----------------------------------------------------------------
gamma | .0189637 .0034703 5.46 0.000 .012162 .0257654
/ln_the | -14.6198 426.525 -0.03 0.973 -850.5934 821.3538
-------------+----------------------------------------------------------------
theta | 4.47e-07 .0001908 0 .
------------------------------------------------------------------------------
Likelihood-ratio test of theta=0: chibar2(01) = 0.00 Prob>=chibar2 = 1.000
Thanks,
Phoenix
--------------------
This email message is for the sole use of the intended recipient(s) and
may contain privileged information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply email and destroy all copies
of the original message.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
