Re: st: Logistic regression with a continuous variable

[Phil Schumm <pschumm@uchicago.edu>]

On Mar 6, 2009, at 8:50 AM, Tomas Lind wrote:
> I´m doing logistic regression and one of the predictors is age. How  
> can I check that age is logit-linear? (I´ve seen a lowess plot but I  
> don´t know how it was done or if it was a good method.)


A good method for doing this is with a partial residual plot -- see  
the chapter in McCullagh and Nelder (1989) on diagnostics for  
details.  Joseph Hilbe wrote a Stata command called -lpartr- to do  
this in the case of logistic regression (-findit lpartr-).


-- Phil

P. McCullagh and J. A. Nelder. Generalized linear models. Chapman and  
Hall, London, 2nd edition, 1989.


*
*   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/