Friday, May 09, 2008

Evaluating Teachers Based On Student Performance

Suppose a swimming instructor told his 10-year-old students to swim the length of the pool to demonstrate what he'd taught them, and half of them nearly drowned? Would it be reasonable to make a judgment about his teaching ability?

Or suppose nearly all the 10-year-old students in a particular clarinet class learned to play five or six pieces well in a semester? Would it be reasonable to consider their achievement when deciding whether to rehire the music teacher?

These questions answer themselves. Only an idiot would overlook student performance, be it dismal or outstanding.

However, suppose test results indicated that most students in a particular class don't have a clue about how to multiply with fractions, or master other material in the curriculum? Should that be considered when the math teacher comes up for tenure?

Whoops, the obvious answer is wrong. That's because public education lives in an upside-down universe where student outcomes are not allowed to be connected to teaching.

And so we have the battle lines drawn in the battle to assess teacher performance. As EIA pointed out, the author's credentials cannot be easily dismissed:

Mr. Merrow, a former teacher in high school, college and federal prison, is education correspondent for the "NewsHour with Jim Lehrer" and president of Learning Matters, Inc.

So, how do we address his points? Hidden in his essay is the answer:

Of course, not every kid comes to class equally able to complete the day's assignment. Some are new immigrants, others are gifted, and still others might have a learning disability. These factors affect test scores as much as or more than who is teaching.

Still, students at whatever level of performance can also be evaluated on how much they've improved over a given period of time.

Tennessee pioneered "value-added assessment", which allows schools to determine how much value an individual teacher or school adds to student performance over time. A good explanation of how such an assessment is done can be found here.

I certainly don't think it's fair to place the power over a person's employment into the hands of students. However, if we can determine through scientific analysis that a particular teacher consistently underperforms, that their students do not grow a year's worth during a school year, then targeting additional effort to help that teacher improve his or her pedagogy would be a wise use of resources.

I understand the arguments against many of the methods proposed to evaluate teachers using student test scores. However, I don't see good arguments against a value-added system.


Anonymous said...

You're absolutely right. I find it interesting that, year after year, the students of what I consider to be "good" teachers score better on standardized tests than the students of "bad" teachers.

Edusoft also does a good job of breaking down student performance by teacher.

Its time for teachers to stop saying "We must stop teaching to the tests", when they mean, "Don't hold me accountable for teaching anything at all! Just leave me alone and let me show my movies."

Anonymous said...

I am all in favour of better teachers getting more money, after all, I am one of the better teachers!

But seriously, I really don't know how you can measure this. Value added is one stat, and obviously it is better than just plain old test scores. But even then, you have a problem because some kids are just easier to drag up the grades (levels in the UK) than others. One class I got this year are astoundingly stubborn I am really struggling to get them up the "level ladder", another class have flown up, making on average a whole level in just a year.

The system the UK uses is that you get a payrise every year, no matter what, but when you reach a certain level, you need to apply for threshold. This involves putting together a lot of evidence that you meet certain standards. Your headmaster then assesses your evidence and decides if you pass or not.

Anonymous said...

Trouble is, there's no incentive to excel as a teacher other then pride. That can certainly be a powerful motivator on an individual basis but organization-wide it's thin reed to lean on.

Worse then the lack of incentive to excel is that the absence of objective measures of performance opens the door to the edu-crap that's a blight on the system.

If you're not differentiating between good and bad teachers your not differentiating between good and bad methodologies and techniques. Like the lousy teachers then, the lousy methodologies are attractive *not* because they improve educational outcomes - who's measuring that, right? - but some other quality. It really doesn't matter what the other quality is since the only quality that ought to matter, the educational value, isn't being measured.

Anonymous said...

I want better teachers to be paid more. BUT . . .you can't, in any rational way, base it on student performance.

Students are an input upon which a teacher has no control. Give a bad teacher teak and a good teacher balsa, and the bad teacher will probably produce sturdier tables.

Student performance measures are untenable in at least four ways:

1) The teacher has, in most cases no say over his caseload. if he does, the best teachers will cherry-pick the best students whenever possible. I know, because I've done this--not for financial gain, but for student preference and personal satisfaction. So -- how could my performance with the 'good kids' be compared to someone else's?

2) Value added is a great idea -- but it will undoubtedly vary based on the input. a student not lacking any of the needed skills to go on will be exponentionally disadvantaged compared to those with some.

3) Standardized tests suck. All of them. And, they suck more when we place no value on the student's effort. Yet, they are consistently used as the sole measure of teacher's performance. I'm sorry, but if you place a $5000 bonus upon my student's performance, I will teach to the test, and give them all $5 to try hard.

4) Performance based pay has a strong bias towards the popular teachers. Some popular teachers are popular because they are well liked by their students, and teach the material well. On the other hand, some teachers are popular because they hand out As like candy. On the other hand, there are many hated teachers that do a great job. If I'm to be paid more based on student performance, why not hand out more As and Bs?


Anonymous said...

I consider myself a good teacher. I support merit pay. Give me any criteria and I will outperform most of my peers. Give me the worst students, value-added evaluation, and
I will outperform most of my peers.

Additionally, give me the worst students and I will raise them to a level above some of the teachers that have "good" students.

Anonymous said...

I'm was what is considered a very successful student. Coming from a family of teachers, my folks were always involved without helicoptering or browbeating teachers.

I learned both the subject matter and how to test well on it. (My brother jokes that I'm better at the tests and he's better at the subjects themselves, which is true a lot of the time) Tests were the reality for me, and learning how to take tests in general was more useful in my school career than having good handwriting.

What's a fair way for students and parents who are involved and observant to contribute to teacher evaluations without falling into the quid-pro-quo trap of A's = Good Evaluation (A++++ Would Learn From Again!!11!)

I've had teachers that I benefitted from greatly, and the best we could do for them at the time was to call and write to the principal and superintendent with our thoughts, but we never knew if these got back to the teachers with real recognition and appreciation from the district.


Mr. W said...

I hate the fact teachers are tied to standards tests (let me say my scores last year were top in one subject, but bottom of another). Teachers really don't have that much control over students achievements on tests.

I read in one study that socio-economic status and something like metanl stability are the top 2 factors for success on these tests. Tell me which one I can help out with as a teacher.