Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

st: Omit Constant from Count Models


From   "Habiger, Matt" <MHabiger@ndbh.com>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: Omit Constant from Count Models
Date   Fri, 14 Sep 2012 12:58:54 +0000

I'm hoping somebody can inform me of what impact(s) omitting a constant term from count models, such as poisson or negative binomial, have? Does it impact t-statistics or the validity of coefficient estimates? I've scoured the internet and textbooks but haven't found any conclusive answers. Is it similar to OLS? Any references would be much appreciated.

I'm modeling the number of days a patient spends in a hospital for a given year and the constant is causing the predicted visits distribution to start at ~4 days (exp(1.46)). In the actual data, roughly 25% of days are below 4 (only those with visits are being modeled). When I drop the constant my estimates are much closer to resembling the actual distribution. Below are the outputs from two models for reference.


Truncated negative binomial regression            Number of obs   =       1334
Truncation point: 0                               LR chi2(5)      =      99.50
Dispersion     = mean                             Prob > chi2     =     0.0000
Log likelihood = -3639.7431                       Pseudo R2       =     0.0135

------------------------------------------------------------------------------
    inpunits |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  claims2009 |   .0047266   .0020829     2.27   0.023     .0006442    .0088091
previnpunits |   .0290032    .007004     4.14   0.000     .0152755    .0427308
       age09 |   .0041179   .0016468     2.50   0.012     .0008903    .0073455
  inplow_ind |   .1627803    .080974     2.01   0.044     .0040741    .3214865
inphigh_ind |   .2904038    .078893     3.68   0.000     .1357764    .4450312
       _cons |   1.469507   .0654664    22.45   0.000     1.341195    1.597818
-------------+----------------------------------------------------------------
    /lnalpha |   -.418825   .0693098                     -.5546696   -.2829804
-------------+----------------------------------------------------------------
       alpha |   .6578193   .0455933                       .574262    .7535346
------------------------------------------------------------------------------
Likelihood-ratio test of alpha=0:  chibar2(01) = 2745.11 Prob>=chibar2 = 0.000


Truncated negative binomial regression
Truncation point: 0                               Number of obs   =       1334
Dispersion     = mean                             Wald chi2(9)    =    1040.04
Log likelihood = -3781.6517                       Prob > chi2     =     0.0000

------------------------------------------------------------------------------
    inpunits |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  claims2009 |   .0102505   .0031548     3.25   0.001     .0040672    .0164338
previnpunits |   .0196405   .0113036     1.74   0.082    -.0025141    .0417951
       age09 |   .0290971   .0015435    18.85   0.000      .026072    .0321222
  inpdep_ind |  -1.145573   .7489945    -1.53   0.126    -2.613575    .3224292
inpschi_ind |   .4475055   .1869523     2.39   0.017     .0810857    .8139254
  inplow_ind |   .5862683   .1092474     5.37   0.000     .3721473    .8003893
inphigh_ind |   .4966423   .1118703     4.44   0.000     .2773807     .715904
bluecareind |   .1664171   .0713599     2.33   0.020     .0265542      .30628
     sex_ind |   .3234388   .0651693     4.96   0.000     .1957093    .4511683
-------------+----------------------------------------------------------------
    /lnalpha |   .4669258   .0887536                      .2929719    .6408797
-------------+----------------------------------------------------------------
       alpha |   1.595083   .1415694                      1.340405     1.89815
------------------------------------------------------------------------------
Likelihood-ratio test of alpha=0:  chibar2(01) = 4621.10 Prob>=chibar2 = 0.000

Thanks,

Matt Habiger 




---
CONFIDENTIALITY NOTICE: This email message and any attachments are for the sole use of the intended recipient(s) and may contain proprietary, confidential, trade secret or privileged information. Any unauthorized review, use, disclosure or distribution is prohibited and may be a violation of law.  If you are not the intended recipient or a person responsible for delivering this message to an 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/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   |   Site index