Why has the ‘learning styles’ theory persisted without any evidence? And what do we actually know about how we learn?