Teachers blast L.A. Times for releasing effectiveness rankings

[Francesca9]

Source: LA Times

National and local teachers unions sharply criticized The Times on Sunday when the newspaper published a database of about 6,000 third- through fifth-grade city school teachers ranked by their effectiveness in raising student test scores.

Read more: http://www.latimes.com/news/local/la-me-teacher-react-20100830,0,1507297.story

---

It would be hard to read a newspaper story about what a bad teacher I am.

[Flabbergasted]

1. Why are they bad teachers? nt

[EFerrari]

6. Isn't that the wrong question?

A think tank came up with some arbitrary measure. And for the icing, madfloridian just posted an OP about a lady who was "rated" by these people while she was out of the country on a Fullbright.

[Flabbergasted]

8. I was asking the OP why they thought the teachers were bad. nt

[MadHound]

11. Possibly because they have a class full of underpriveleged kids,

Who come from abusive homes that don't value education. Throw in a few learning or developmental disabilities and you have the makings of a class that is going to make little, if any progress.

Furthermore, this is all based on standardized tests, which a large number of children have problems taking. Not to mention that the kids have no stake in this test, so they feel free to blow it off just for the hell of it if they want.

There are a number of reasons why test scores turn out bad, none of them which have anything to do with the job the teacher is doing. Using standardized testing to determine how a teacher is performing is a highly inaccurate way of assessing teacher, or student for that matter, performance.

Better methods include observations, their portfolio, individual student progress, peer review, etc. Methods that used to be used, but now are falling out of favor.

And you never, ever publish these results in the paper. It is unfair to the teacher and student. How would you like it if you job evaluations were published for all to see?

[roody]

14. Even a writing sample! Oh, never mind, that's not on the test.

[Recursion]

17. If what parents care about is test scores...

...that's a stupid thing to care about. But it seems to be what a lot do care about.

Either some teachers are better at raising test scores than others, or not.

If so, the parents will want them teaching their kids rather than other teachers.

If not, from the test score perspective it doesn't really matter who's a teacher.

[CoffeeCat]

23. School administrators and principals are teaching...

...parents to view these standardized tests as the most important indicators
of their child's academic performance.

In my kids' elementary school--those standardized tests are fawned over. Parents
are told how important those scores are and a great deal of weight is put on them.

The gifted programs in our school--are based totally on those test scores. I have
a straight-A student who desperately wants to be in AELP. She wants the extra work
and creativity that the AELP kids get. I asked a mother about AELP and she said
that kids are "invited" and the invitation is extended only to students who get
a 95 percent overall score on those standardized tests. My daughter only has a 92
and was not even considered. I do a lot of outside enrichment with her, because she
just loves to learn and read, so it's not that big of a deal. However, this situation
underscores the point that these tests are considered the ultimate (and sometimes only)
measure of student ability.

When school funding is contingent upon those test scores--of course principals are
forced to focus on those tests. And teachers are going to be pressured into teaching
in ways that increase those test scores. Students become little trained monkeys,
churning out the high numbers for the administrators--so everyone can get their funding
and look good on paper.

In my opinion, this dehumanizes children and turns them into little test-score factories
who are supposed to meet quotas. Some children don't take tests well. Some children
are hungry on testing day. Some children need more time than the test allows. Some
children get test anxiety. Some kids are from violent homes and have other things
on their mind. Many children are on meds that may affect test scores. Yet--it all
comes down on the teacher when these scores aren't fantastic?

It's ridiculous.

[Recursion]

21. Mine are about to be

since I work for a Federal contractor.

http://washingtontechnology.com/articles/2010/08/25/the-opening-of-fapiis.aspx

[roody]

13. Maybe they are not. Maybe they are willing to teach poor kids.

[boppers]

2. Dupe.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x4522217

[KILL THE WISE ONE]

3. teacher should demand "effectiveness" data from all public employees be released

that would be fun wouldn't it DMV, Police, DSHS :+

[EFerrari]

7. I would like to see Duncan's "effectiveness" rating. n/t

[sandnsea]

4. Look at this elementary school

All the 3rd grade teachers suck. Or do they? Maybe there's a problem with the third grade curriculum, or even the second grade curriculum that has these kids entering this school behind. Why would you not want to have this information?

http://projects.latimes.com/value-added/school/los-angeles/raymond-avenue-elementary/

And the newspaper did this project. Just like they report on excessive deaths in hospitals or lost kids at CPS or any other failing public service institution.

[Hannah Bell]

10. no, bill gates funneled money through a non-profit to pay a RAND analyst to do it.

the newspaper didn't do the work, didn't pay for it -- they just published it.

it's a massive fraud.

bill gates = oligarch.

[bemildred]

5. What really annoys me about this is how dumb it is.

They assume that there is a cause and effect relation (good teachers cause better scores) when random variation will do the job nicely. Even in an orderly world where the normal distribution applies, some teachers will look much better than others as a result of random variation, and some much worse. If you want to claim some causal effect is at work, you FIRST have to prove that random variation won't do the job, you have to rule out the null hypothesis.

The "data" means squat about whether this or that teacher is doing a good job, even in the debased sense that they are effective in raising test scores.

[LWolf]

12. I've been trying to explain this since high-stakes testing first appeared

at the state level.

Nobody listens. :(

[bemildred]

15. "We don't have a clue" is never a popular option with the administrative classes.

Unfortunately, as events prove time and again, they really don't have a clue. Meanwhile everybody suffers with this drivel.

[Recursion]

16. OK, but there's still signal under the noise

I don't remotely know about every system, but I know IMPACT scores do have a year-to-year autocorrelation that is positive at x1 > x2. So it can't all be noise.

[bemildred]

18. Year to year autocorrelation just means it's not random.

Things don't stray too far from what they were last year, there is some inertia. It has nothing to do with cause and effect.

You still do not know anything in particular in the way of cause and effect about any particular case. Statistics NEVER give causal information about particular cases, and they should NEVER be used that way, and the default assumption when you have a good year or a bad year is always that you got lucky or you got unlucky.

[Recursion]

20. Well, statistics of a non-ergodic function never give causal information

Even so, batting averages still mean something in baseball.

Are you really saying teachers have no impact on their students' test scores?

[bemildred]

24. Batting averages don't give you causal information either.

They tell you so and so had a good year, or years, or lifetime, or the opposite, or he might be boring. He might really be good, he might use "performance-enhancing" drugs, he might have just been so-so but really lucky, you don't know. You seem the same ignorant shit about the stock market, it shows autocorrelation, therefore your money is safe and the market is a well-understood thing,and we know why it went up yesterday, and it will behave itself tomorrow.

I'm not saying teachers have no effect, I'm saying test scores and statistics based on test scores don't distinguish good teachers from bad teachers. It's tautological, you define the test to be a measure of teacher quality, and then you proclaim that teachers with bad statistics are bad teachers.

Test scores and statistics based on them are a POOR measure of teacher quality, even in theory. It is an UNPROVEN assumption that they are good for anything at all, even as measure of how the students are doing. It could be a shitty test. It could be a shitty test to give that batch of students. It could have been an unfortunate series of ten years when the bad test was given on a bad day to the wrong sort of students.

This is what I mean about dumb: autocorrelation is the default, the norm, the null hypothesis, it is precisely when you don't have autocorrelation that it is time to suspect there is some instrumental cause behind the sudden change. But you still have to rule out chance and chaotic collapse and similar scenarios, which is why in SCIENTIFIC work one wants repeatable observation or experiment to establish a presumption of cause and effect.

[Recursion]

25. You always have autocorrelation

It's just the inner product of the function with itself over its domain (integral of F(s)*F(t - s) ds). I think we may be using the word differently. White noise (which is what I thought you were saying test scores are) has a nonzero value only where s = t.

[bemildred]

26. So, like I said, it's tautological, just like all of mathematics?

There is a big area between random noise ("white noise") and strict causal order, and it is where most of the real world is.

"Autocorrelation is the cross-correlation of a signal with itself." That is: present values "tend" to correlate with recent past values. It's an information theory thing, handy in computer engineeering, etc. I've used it to justify using certain algorithms with some success.

Assuming that a function is ergodic is a stringent claim, invariant measures are hard to come by in the real world, like simple cause and effect relations. It is like asserting that a function is continuous, or a distribution is normal, and one ought to offer some credible evidence before pretending to reason about the real world on such a basis.

Now: I'm saying there in no "signal" in test scores that applies to teacher quality. None, zippo. That does not mean that in the case of a particular bad teacher that you won't see bad test scores, or that in the case of a good teacher you won't see good test scores. It just says the bad test scores don't allow you to know anything about why, just like Baseball scores or the changes in the stock market don't really give you any information about why. That question is much more difficult.

[DrDan]

9. where are the teacher's unions? If this is not addressed now,

all teachers across the U.S. may face the same treatment. After all, it seems to have the administration's support.

My wife is a member of the teacher's union. And for years, I have been amazed at their lack of ability to address the issues of their members . . . . not unlike the D-party.

[bemildred]

19. Top down unions are just like any other top-down organization, it's all about the boss. nt

[Smarmie Doofus]

22. Asleep. Or maybe just in bed. nt