RCTs can have major external validity problems too. Two RCTs with the exact same trial protocol in the exact same population can have *completely different* & *equally internally valid* ITT effects if their adherence patterns happen to differ.