Re: st: Interaction term in logistic regression

[Maarten Buis <maartenlbuis@gmail.com>]

I can see a couple of problems:

1) For a logistic regression the dependent variable needs to be the
first variable mentioned in the variable list and it should not
contain a prefix (i.).

2) The -xi:- prefix is outdated, you should use the factor variable
notation instead whenever possible, see -help fvvarlist-.

3) I would consider using a linear spline (1st order spline) instead
of age in categories (0th order spline). The linear spline is a more
flexible than age categories, but is still easily interpretable. For
even more flexible but still interpretable effects you can look at:
Roger B. Newson (2012) "Sensible parameters for univariate and
multivariate splines" The Stata Journal, 12(3): 479-504.

4) If you use income as a variable, then I would call it income and
not economic status. The latter could be interpreted as more general,
for example it could wealth, which you did not do in your model.

-- Maarten


On Tue, Apr 22, 2014 at 7:12 AM, Ashish Gupta <ashishatbhu@gmail.com> wrote:
> I am planning to employ  logistic regression with interactions to see
> the effect of socio-economic characteristics on age-specific
> mortality.I would like to know that how one can include interaction
> term (Age versus economic status) in adjusted as well as unadjusted
> model.
> Example of the model-
> xi : logit i. race  i. age_group i. income i.place_res  i.age_group ##i.income
>
> Any response will be appreciated.
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/



-- 
---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany

http://www.maartenbuis.nl
---------------------------------
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/