I have an experiment looking at the effect of rearing condition on development of the number of neural networks (ensembles) in mice. I therefore have two factors (Age 3, 5, 7 dpf) and Rearing Condition (GR and NR). I ran a two way anova in R :
my_anova <- aov(numb_ensembles ~ age * Rearing_conditions, data = dev) Anova(my_anova, type = "II")
Anova Table (Type II tests) Response: numb_ensembles Sum Sq Df F value Pr(>F) age 1230.9 2 4.9170 0.01352 * Rearing_conditions 0.7 1 0.0057 0.94053 age:Rearing_conditions 452.2 2 1.8062 0.18014 Residuals 4130.5 33 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
This revealed a significant main effect of age. I am now unsure which post hoc test to use. I'm guessing looking at all pairwise comparisons (such as Tukey's HSD) would be unsuitable based on the fact that their is no interaction. However, can i do planned pair wise comparisons (t.test with bonferrori) between ages conditional to the rearing condition? ie for the GR group is there a difference between each time point.