Add a page to explain the statistical test being used

[lukasvermeer]

We can use a Chi-squared test to test for statistical independence. In other words: we will test the hypothesis that there is no statistical interaction. If we can reject that hypothesis (i.e. the p-value is low) then we can say we are confident there is an interaction.

The same test is also used for detecting SRM, but in the case of interactions we have two experiments instead of one and we look at conversions instead of traffic.