Exercise

# Anova: Multiple comparisons (2)

In R a handy function to follow up an Anova with pairwise comparisons is the `pairwise.t.test()`

function. pairwise.t.test() takes an argument `x`

that is the name of your response variable, followed by the argument `g =`

where you tell the function your grouping variable. Furthermore, you can choose an adjusment method for the p value by specifying the p.adj parameter. For instance, if you want to do a Bonferroni correction or a Holm correction, you can specify set the p.adj argument to either `p.adj = "bonf"`

or `p.adj = "holm"`

. So all in all the usage of the pairwise t test should look something like this:

```
pairwise.t.test(dependent_variable, g = grouping_variable, p.adj = "bonf")
```

Another often used correction for multiple testing is the Tukey method. To use this in R, you can use the `TukeyHSD()`

function. You can then assign your Anova object to the x parameter and you can assign your grouping variable to the which argument. The which argument needs to be the exact same grouping variable as specified in your Anova object.

Instructions

**100 XP**

- Use the
`pairwise.t.test()`

function to follow up your Anova analyses and print the output to the console. Make sure to specify the duration variable to x argument and the genre variable to the g argument. Also make sure to use a Bonferroni correction. All variables are available in the`song_data`

dataframe. - Do a Tukey to follow up your Anova. Your Anova object is availabe under the name
`fit_aov`

. Set the which argument to the character argument "song_data$genre" and print the output to the console.