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 on April 23, and its replacement, statalist.org is already up and running.


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

st: Logistic Regression Coefficients Using MIM and SVY


From   Steven Frenk <smf7@duke.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: Logistic Regression Coefficients Using MIM and SVY
Date   Wed, 31 Mar 2010 11:25:00 -0400

  I am estimating logistic regression models using STATA/SE 10.1 (Windows) and
the MIM and SVY commands.  The models include my dependent variable and one of
my independent variables [mim: svy: logit DV IV, or].

  In four of my models, I get weird parameter estimates?the odds ratio and
CI?s are whole numbers and the standard error is zero. For example:

. mim: svy: logit z_tot_alldisorders z_totprogs06, or

Multiple-imputation estimates (svy: logit)               Imputations =       5
Survey: Logistic regression                              Minimum obs =    1499
                                                         Minimum dof =  1000.0

------------------------------------------------------------------------------
z_tot_alld~s | Odds Rat.  Std. Err.     t    P>|t|    [95% Conf. Int.]     FMI
-------------+----------------------------------------------------------------
z_totprogs06 |         1         0    7.53   0.000          1        2       .
------------------------------------------------------------------------------

The problem does not exist when I just use the SVY command:

. svy: logit z_tot_alldisorders z_totprogs06, or
(running logit on estimation sample)

Survey: Logistic regression

Number of strata   =         1                  Number of obs      =      7495
Number of PSUs     =      7495                  Population size    = 7485.1806
                                                Design df          =      7494
                                                F(   1,   7494)    =    283.75
                                                Prob > F           =    0.0000

------------------------------------------------------------------------------
             |             Linearized
z_tot_alld~s | Odds Ratio   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
z_totprogs06 |   1.469695   .0335955    16.84   0.000     1.405292     1.53705
------------------------------------------------------------------------------

 This is only occurring with four of my thirteen independent variables. 
Here?s some information about the four variables with the strange estimates:
--Two of them are continuous variables and two are dummy variables.
--Three had missing values (imputed with ice), and one did not have any missing
values.
--When I run logistic models with each imputed data set and use the svy: logit
command, the parameter estimates produced are logical and valid (so I don't
think it is an imputation problem).

  I've tried looking on google and statalist for information regarding this
problem, but I have not found any answers.  Does anyone know why this is
occurring and how to fix it?

Thank you for your time.



************************
Steven M. Frenk
PhD Candidate
Department of Sociology
Duke University
steven.frenk@duke.edu
(919) 672-7495




*
*   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