> Well, a good teacher could (and can) indeed see, what methods he use are effective or not. Any evidence? Doctors could figure out that not washing hands was

> Well, a good teacher could (and can) indeed see, what methods he use are effective or not. Any evidence? Doctors could figure out that not washing hands was killing patients, but didn't. > As he gets new data with new "fresh" kids who are individuals, but usual still average. So you could experiment. There will always be multiple things a teacher does differently so it will be, in practice, impossible to attribute success or failure to any particular practices. That's aside from random noise (which might be more systematic - what if some random good / bad kid happens to have a large influence on the whole class?). Remember how Skinner proved pigeons could develop superstitions? Or look at any other profession that deals with a little bit of randomness (human factors / biology) and see the stupid stuff they used to believe.