[R-sig-eco] Marrying Tukey's HSD and ANOVA results

[Ben Bolker]

Lara R. Appleby 04 <Lara.R.Appleby.04 at ...> writes:

>  I've done a standard two way ANOVA using glm on the dependent
> variable "clutchsize" with the two factors "treatment" (which has 3
> levels called 1, 2, and 3) and "species" (which has two levels
> called 1 and 2). Apparently there is no significant interaction
> term. Then I did Tukey's HSD and found that there were significant
> differences between species at only one of the three treatment
> levels, treatment level 1.

> 
> ##ANOVA RESULTS
> >  summary(aov((clutchsize~treatment*species)))
>                     Df Sum Sq Mean Sq F value   Pr(>F)
> treatment           1  29.26  29.264  7.0230  0.00884 **
> species             1 138.14 138.143 33.1526 4.13e-08 ***
> treatment:species   1   8.11   8.110  1.9464  0.16487
> Residuals         163 679.20   4.167

  I think you failed to tell R that `treatment' was a factor
(i.e. a categorical variable).  The Df in the output above
suggests that you are actually (accidentally) running a linear
regression on treatment number, rather than an ANOVA on
treatment.  The Df should be 2,1,2 for the design you
described above.

  That doesn't answer your other question, but it would
be better to sort out the more fundamental issue first.