Tweaks required for Simpson's Paradox notebook

[drbenvincent]

Recently I updated the Simpon's Paradox notebook.

@Armavica pointed out that I missed an edge in the causal DAG from group to x

So I should really add some discussion of group being a confounding variable, and therefore it should be included in the adjustment set.

Also... Change the title to remove the "mixed models" terminology, and remove it from the notebook. This just causes confusion