Saturday, February 1, 2014

TVAAS at least being considered

This post in in response to the news today here  about the state BOE in Tennessee at least at the moment, apparently removing the requirement for a teacher to have a 3 on TVAAS for 2 out of 3 years for professional licensure.

I'll say in short that this is good news, for teachers, for teacher morale, and by extension, for students of good teachers throughout the state.

I know that the immediate response would be that we need ways to remove ineffective teachers from the classroom.  I've taught for 14 years now and have definitely seen/worked with a few teachers I wouldn't want my children to be taught by.  Emphasis on a few.  In that, I agree that bad teachers should be removed, but what I would disagree with is that you need a bogus number whose statistical validity for the use that Tennessee applies, i.e. hiring and firing of teachers, to it has been disputed by the very person that came up with the measure.  My solution for the problem of bad teachers is to not hide behind a bureaucratic numerical rationale, but rather to have administrators at the building level make the call with due process involved.

Let me throw a few things out there about why TVAAS is bad, from my own mouth, but also from others I've spoken to or been emailed as a result of being on the Teacher Working Group:


  • TVAAS is a black box.  I've heard this from a ton of folks at all levels.  I read through the powerpoint of a presentation given by someone behind TVAAS explaining how it wasn't a black box (in fact, that was a bullet point on a slide).  They gave a formula and explained the variables.  Except one, which had proprietary algorithms, so they couldn't be explained. That folks, is a black box.  Just like chemical companies with their proprietary surfactants, I understand protecting trade secrets, but don't tell that something isn't a black box but then don't let me see how it is calculated.  I'm no statistician, but I think I have enough math to figure it out if shown it all.  I suspect I have more math in my background than a few of the folks deciding if it is right for everyone.
  • If someone gets a 4 or 5 on their observations, but gets a 2 on the TVAAS, why do we assume that that it is the evaluator that needs calibrated, rather than the data.  I'd rather trust my principal to make an informed decision about my teaching as opposed to TVAAS data. (as would almost everyone I've talked to about it).
  • I don't know this for certain, but from all accounts I've heard, TVAAS counts outliers.  Again, I don't know a lot of statistics, but outliers aren't usually the basis for these types of decisions. I have a teacher that had one student of 67 account for 50% of their TVAAS score by blowing off the test.
  • Hand in hand with that is that TVAAS has an unfair effect on teachers, especially as you get to higher levels.  If you teach Biology and the teacher next to you teaches Physics, and you both had the same kid who blew off a test, it hurts you as a Bio teacher because that is a TVAAS course.  On the other hand, it doesn't do much of anything to the Physics teacher as that is not a TVAAS course.  Most courses in High school are not, thus providing a disincentive to teach those classes.
  • That same one student blowing off a test could then bomb another teacher the next year in the SAME CLASS as TVAAS doesn't take that into account.
  • If you teach HS and have 6 classes, but only one is a TVAAS course, that one class is 35% of your evaluation (more if it happens to be where your observations are, but who knows).
Ironically, despite this tirade, I'm ok with getting TVAAS data, I just wish it was explained more.   Tell us what the expected gains are.  In the case of a firing level score (1 or 2 the way it was) have a committee at the school sit down and see what happened.  Was it one student drawing a pot leaf on their assessment (happened to me one year btw)?  Did say, a school lose two weeks to snow the semester that exam was given?  Was the teacher pulled out for a lot of meetings?  Was their plan time constantly eaten up with meaningless PLC time? 

What is to keep teachers from fleeing schools where they fear a student not taking the test seriously and thereby harming their entire career?  What is to stop the inner circle around any principal from making sure they cherry pick the classes for themselves or make sure they only teach non-TVAAS classes?

Again, I understand the thought behind objective data behind decisions in education.  But I question whether the data in this case (TVAAS) is objective.  Trying to reduce education to numbers when it is both an art and a science is futile, like trying to assign numbers to art.  Train good principals (I have an awesome one), and empower them to shape the culture of their schools, trust them to let you know if their are teachers that don't fit that culture.

So I will applaud the state BOE for at least considering things, for having the wherewithal to change their opinions as warranted.  As for the flummox situation, speak to ordinary teachers in a variety of schools, ask them what they think, pros and cons.  Don't surround yourself with people only like yourself or make educational decisions based mainly on a political predisposition.  And recognize that things change as we learn more, that is what education is all about.