A treatment can work better than a rival in every single subgroup of a study and still look worse overall once you combine the data, purely because of how…
This is the stat trick that makes headlines lie without technically lying.
Visit When Every Group Disagrees with the Total