[R] Strange results with anova.glm()

Hi, I have been struggling with this problem for some time now. Internet, books haven't been able to help me. ## I have factorial design with counts (fruits) as response variable. str(stubb) 'data.frame': 334 obs. of 5 variables: $ id : int 6 23 24 25 26 27 28 29 31 34 ... $ infl.treat :

Re: [R] Strange results with anova.glm()

Dear Gustaf, I can think of two reasons why the two tests can disagree. First, the t-test from the summary() output is based on the covariance matrix of the coefficients, while the F-test in the anova() output is based of fitting alternative models. The two are not in general the same.