???? ???? <renana.zeckbach <at> mail.huji.ac.il> writes:
>
> if I have binomial Ancova like:
> model<-glm(y~treatment*x, binomial)
> summary(model)
> how should I understand the output?
> I want to know if I have interaction between the treatments and if the
> regression is significant?
> should I write:
> summary(aov(model))
> or:
> summary(glm(model))?
> every summary give different answers.
> thanks!
>
summary(model) is probably what you want. Neither summary(aov(model))
nor summary(glm(model)) makes much sense -- 'model' already contains
the results of fitting the statistical model.
In general, giving a more detailed example will get you better
answers.
Ben Bolker