Thanks Graham, I couldn't agree ore on the last point. There is a lot to learn when it comes to causal inference, so we need as many contributors as we can get.

Regarding your initial question, I have not tried the do function before, so I would also default to doing it by hand. In case it is helpful, I have a write up and video on the subject here: https://medium.com/towards-data-science/causal-effects-via-dags-801df31da794