Of professors and physicians

doctorThere is an in-built confirmation bias in medicine. If the patient gets better, doctor assumes s/he healed the patient. If the patient dies, doctor assumes the patient was really sick and beyond help. This bias means doctors have gone on with useless or even harmful practices for years (or even centuries).

professor3Pondering the SRTE‘s, I suddenly realized I was thinking just the same way: when students said they learned things, I attributed that to my fantastic pedagogy. When students said they learned nothing — or worse, showed me that they learned stuff that was wrong — I assumed those students hadn’t come to class, or didn’t listen, or were unteachable…

Mmmm…. Well, Andrew, follow the logic you taught in class…

Ok. For sure, customer satisfaction surveys could not get medicine out if its confirmation bias. So SRTEs can not get we professors out of ours. What’s needed instead is the equivalent of the randomized control trials which frequently save medicine from itself (or more correctly, save patients from the practitioners). I need to enroll a class and then at the last minute, refuse entry to a randomly chosen half of the students so they go off and do another course. Which group would go on to best achieve the course objectives?

The endpoint of such a study is not easily defined, let alone measured. Performance on the final exam is irrelevant (and is in any case a softer-than-soft endpoint). The real test is whether a difference could be detected years later, long after graduation.

Good thing we can’t do that experiment.

Leave a Reply

Your email address will not be published. Required fields are marked *