[R] interpreting interactions in a model

Dennis Murphy djmuser at gmail.com
Wed Aug 17 20:01:25 CEST 2011


Hi:

On Wed, Aug 17, 2011 at 1:56 AM, gaiarrido <gaiarrido at usal.es> wrote:
> Hi,
> I´ve got this model
>> model<-glm(prevalence~agesex+agesex:month,binomial)
>
> and the output of anova is like that
>
>> anova(model,test="Chisq")
>                          Df Deviance Resid. Df Resid. Dev P(>|Chi|)
> NULL                            524     206.97
> agesex                 2   9.9165       522     197.05  0.007025 **
> agesex:month        9  18.0899       513     178.96  0.034145 *

Where is the month main effect? Even if it's 'not significant', it
belongs in the model (principle of marginality).
>
> I don´t know how to interpret the interaction "agesex:month", my mind doubt
> between 2 options:
> a) For a giving group of "agesex" there are differences between months
> b)There are differences between groups of "agesex" in some months but not in
> others.
>
> Which option is correct?

In the absence of a reproducible example, no one can say, but it's
within the realm of possibility that either or both could be correct.
The significance test for interaction indicates *whether* an
interaction effect exists - it doesn't tell you *what type* of
interaction exists. Conditional on the inference that an interaction
effect is present, the next step of the analysis is to investigate the
nature of the interaction. This could involve planned contrasts,
multiple comparisons, graphics, etc. The car, effects, multcomp and HH
packages can be useful for these types of investigations.

HTH,
Dennis

> Thanks very much
>
> -----
> Mario Garrido Escudero
> PhD student
> Dpto. de Biología Animal, Ecología, Parasitología, Edafología y Qca. Agrícola
> Universidad de Salamanca
> --
> View this message in context: http://r.789695.n4.nabble.com/interpreting-interactions-in-a-model-tp3749430p3749430.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>



More information about the R-help mailing list