Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Assessment of fit output from -clogit; missing dx2 and dbeta values


From   "C.Chetkiewicz" <c.chetkiewicz@comcast.net>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Assessment of fit output from -clogit; missing dx2 and dbeta values
Date   Tue, 6 May 2008 14:21:17 -0700

Hello,

I am running Stata 9.2.

I am assessing the fit of a -clogit model running on a 1:20 matched study.  I am following Hosmer and Lemeshow to assess the fit of the model and look at the various diagnostics statistics.  When I list out the statistics for the poorly fit strata, I frequently get what looks like missing values for the _dx2 (lack of fit diagnostic) and _dbeta (delta-beta influence statistic).  I haven't played around with the statistics enough to clearly understand why I am getting this output.  Should I be concerned about the model when I see these types of missing values or just these specific data records?  Perhaps I also need to examine the changes in the model when I delete the stratum to decide or do the missing values by themselves warrant dropping this stratum?  

Thank you kindly for any feedback/advice or suggested reading.

list uidstep sampled_path  p _dbeta _dx2 _hat if uidstep==1180

       +-------------------------------------------------------------------+
       | uidstep   sample~h          p      _dbeta        _dx2        _hat |
       |-------------------------------------------------------------------|
 8317. |    1180   observed   .1488741   .01837156   4.8842606   .00374729 |
 8318. |    1180     random   .0162666           .           .   .00290703 |
 8319. |    1180     random   .0131587           .           .   .00458869 |
 8320. |    1180     random   .0768851    .0000795   .07696454   .00103183 |
 8321. |    1180     random   .0124258           .           .   .00431216 |
       |-------------------------------------------------------------------|
 8322. |    1180     random   .0053696           .           .   .00248764 |
 8323. |    1180     random   .0338684   .00004183   .03391016   .00123196 |
 8324. |    1180     random    .034001   .00007216   .03407298   .00211323 |
 8325. |    1180     random   .0122176           .           .   .00414418 |
 8326. |    1180     random   .0124258           .           .   .00431216 |
       |-------------------------------------------------------------------|
 8327. |    1180     random   .0818624   .00016331   .08202539     .001987 |
 8328. |    1180     random   .0052728           .           .   .00379471 |
 8329. |    1180     random   .0061193           .           .   .00246949 |
 8330. |    1180     random   .0052728           .           .   .00379471 |
 8331. |    1180     random   .0986101   .00010536   .09871531   .00106622 |
       |-------------------------------------------------------------------|
 8332. |    1180     random   .1431424   .00064988   .14378941   .00449935 |
 8333. |    1180     random   .1228505   .00041656   .12326569   .00336799 |
 8334. |    1180     random   .0451796   .00003575    .0452153   .00078993 |
 8335. |    1180     random   .0052728           .           .   .00379471 |
 8336. |    1180     random   .0124258           .           .   .00431216 |
       |-------------------------------------------------------------------|
 8337. |    1180     random   .1084989   .00012592   .10862466   .00115786 |
       +-------------------------------------------------------------------+





***********************************************
Cheryl Chetkiewicz
P.O. Box 60208
Palo Alto, CA, 94306
USA
Telephone: 650-331-0069
E-Mail: c.chetkiewicz@comcast.net
 please consider the environment before printing this email



*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index