In the rush among state's to announce that teachers will be evaluated on student achievement and on their own teaching performance, no one is particularly clear about exactly how this will be done and whether any model of evaluation is a tried-and-true method with rock solid validity and reliability.
Those last few words betray my research background, and also note the big problems with evaluation of teacher effectiveness. Validity and reliability mean we are sure that teacher performance is the cause of the student achievement we are measuring, and that we know exactly what kinds of behaviors and actions they teacher undertook to cause the learning.
That's a problem because in education things might appear related, even though they are not. So suppose one year, students in many classes in a district do very well on reading. Is it because of what the teachers suddenly did? Or is it because the new reading curriculum is well suited for students and teachers? It's like hemlines and the stock market. Does a rising market really "cause" people to wear shorter skirts?
Meanwhile, what contextual aspects of students, schools, and teachers should be included in the statistical growth models of student learning? There are a couple of models most commonly discussed, EVAAS eg. But some places are making it up as they go, with each district creating its own model.
I'm catching up with this post, getting in the habit of writing and starting build a body of work.
No comments:
Post a Comment