Another set of tweets god me thinking:

How do we help school administrators help and evaluate CS teachers? asks @lsudol of @CSForAll #csta2017 Many don't understand CS

— Alfred Thompson (@alfredtwo) July 10, 2017

If the current system is any indication, checklists are not the answer

— Mike Zamansky (@zamansky) July 15, 2017

I get the intent. CS teachers should be evaluated by people who have some clue about the subject. Leigh Ann's reporting that some people are working on checklists got me wondering if it really matters? Sure, it matters if teacher observations were actually set up to improve instruction but given what we have in NY does it really?

In New York City, teacher observations are a major part of a teacher's annual rating and the other part is still that standardized test nonsense. Teachers are observed at least 4 and commonly 6 times or more per year. There might be pre and post observation conferences.

Sounds good but the system is amazingly flawed.

To start, the DOE uses the Danielson Framework which provides a basis for teacher evaluation. The framework is at best flawed with even Danielson coming out to say it's being misused. Of course I'm sure she protested all the way to the bank. A teacher can be a "Danielson style" teacher and be great but a teacher can also check all the Danielson boxes and be mediocre at best. Likewise, some of the best teachers I know break the Danielson mold and if a supervisor rigidly adheres to the framework these great teachers would be found ineffective.

To make matters worse, the rubric is so large that the DOE just focuses on small parts which means even if the framework was comprehensive and correct, teachers are only evaluated on a small part.

When a supervisor evaluates a teacher, they have a checklist with the rubric where they can rate a teacher 1 (ineffective), 2 (developing), 3 (effective), and 4 (highly effective) and write down some notes.

I guess the hope was to have some consistency but the truth is the process is very subjective. I know a supervisor who gave a teacher a ratings of 2 (which is basically failing) for the professional development category even though the teacher fulfilled all contractual obligations and then some. There are also supervisors who won't give a rating of 4 out of principle. On the other side, I've met supervisors loathe to give low scores.

The tweets at the top of this post refer to the current situation where CS teachers are never observed and evaluated by supervisors who are knowledgeable about CS. Even if someone designs a checklist - something I'm skeptical about given the fact that there are so few people strong in CS, experienced as teachers and with an eye for teacher evaluation out there, if the application of the Danielson framework is any indication, a CS checklist won't help.

Besides, CS won't be any different from any other subject. Since the destruction of the comprehensive neighborhood high school we've had a proliferation of small schools. Some people think that small schools are universally better. They're wrong. You need a mix. One of the problems with small schools is that you don't have subject area supervisors. You have a principal and an assistant principal. The result is that most teachers are evaluated by supervisors with no subject area expertise and no subject area pedagogical expertise.

The old system, even though it was also subject to abuses was much better - at least in the hands of a good supervisor. The observation system is inherently flawed. Observations are snapshots out of context and the mere presence of an observer changes the tone of the room. That said, if a supervisor knows the heartbeat of their school and has ongoing conversations with their staff, they can both ensure teachers are doing an honest job and also help them to improve.

The method of observation also doesn't have to be complicated. My first supervisor made it very simple. He said he looked for one basic thing - "was learning going on?" The conversation would then go from there.

Sometimes you have to look beyond the lesson. In my first year, I frequently observed a friend's history class. One day the class seemed to do very little. It seemed like all talk and play. I talked to the teacher about this. He said "that was what they needed today." He was a master and he was right. You might drop in on his class to see amazing instruction or you might come in and wonder why he wasn't fired long ago but at the end of the year when you looked at how much his kid learned and loved his class you realized that he was an amazing teacher. No rubric or framework needed. Just the eyeball test.

They also used to have more peer observations.They had to be coordinated by the department supervisor and he or she had to be in the loop but the peer observation model helped cross pollinate ideas in a department, build camaraderie and it also eased the workload of the supervisor.

To answer Alfred's question from his tweet, how do we help administrators help CS teachers? Get other teachers involved, keep an open mind and look at the big picture.