[R] additive interaction for a dichotomous dependent variable (i.e. risk difference)

[David Winsemius]

On Nov 19, 2012, at 1:57 AM, wouterjohannes wrote:
>
> No, I never got any response. I used an article by Knol to solve the  
> issue. Ik hope this is useful for you too.
>
> Best regards,
>
> Wouter
>
>
> Knol, M.J., van, d.T., Grobbee, D.E., Numans, M.E., Geerlings, M.I.,  
> 2007. Estimating interaction on an additive scale between continuous  
> determinants in a logistic regression model. International Journal  
> of Epidemiology 36, 1111-1118.

I'm a bit surprised that you did not use the binomial "family" with a  
probit link. The help page for ?family mentions the probit link as  
appropriate for "normal CDF". It would seem possible to construct  
competing models both with binomial errors and compare the deviance.

If you wanted to do this via embedding each model in a higher order  
model where the degree of "multipicativity" was represented by a  
parameter to be estimated, I remember that Breslow offered GLIM code  
to perform that test in one of his two volumes. My copies are  
temporarily in storage or I would have given a more specific citation.

N.E. Breslow & N. E. Day (1980) Statistical Methods in Cancer  
Research. Volume I - The Analysis of Case-Control Studies
N.E. Breslow & N. E. Day (1987) Statistical Methods in Cancer  
Research. Volume II - The Analysis of Cohort Studies

Wait ... I can do better. The entire chapter that I was remembering  
appears to now be accessible online from the publisher's website:

http://www.iarc.fr/en/publications/pdfs-online/stat/sp82/SP82_vol2-4.pdf


>
> ________________________________
> Van: sarahw [via R] [mailto:ml-node+s789695n4649687h3 at n4.nabble.com]
> Verzonden: vrijdag 16 november 2012 2:39
> Aan: Peyrot, Wouter
> Onderwerp: Re: additive interaction for a dichotomous dependent  
> variable (i.e. risk difference)
>
> Did you ever get a response to this or resolve this yourself?
>
> Many thanks!
>
> ________________________________
> If you reply to this email, your message will be added to the  
> discussion below:
> http://r.789695.n4.nabble.com/additive-interaction-for-a-dichotomous-dependent-variable-i-e-risk-difference-tp4635842p4649687.html

David Winsemius, MD
Alameda, CA, USA