# Definition

ANOVA is the acronyme of** Analysis of Variances**. It is a statistical test used to compare the **differences of means** between two groups or, usually, **more than two groups**. The analysis of variance compares the dispersion between the means to estimates if the variability observed could be the effect of the intervetion or treatment.

To compute an ANOVA test, each group is needed to follow a normal distribution, to have the same variance and to be independant from each other.

# How to perform an ANOVA?

- Go to "Test variables"
- Select a Numeric variable you would like to compare between your different groups
- Select a List variable
- Click on the panel name "Compare values of ... between groups of patients"

# The ANOVA is not performed

An ANOVA requires some conditions and assumptions. If an ANOVA is not proposed, please ensure that:

- Your List variable contains at least 3 groups (modalities / values)
- Your Numeric variable is known for at least 3 patients in each group
- Your data follows the assumptions of normality and homoskedasticity. Otherwise, the Kruskal-Wallis test will be proposed