Add Durham to the list of school districts crowing about gains on state test results.
The press release on the district's web site credits teachers and principals for "these impressive gains." But the press release, posted today as part of the district's release of preliminary test results, doesn't directly mention the fact that retests were counted this year.
“The EOG and EOC preliminary information is fantastic news and true confirmation that our teachers, principals and students are working hard, with the support of parents, to ensure that good teaching and learning is happening in Durham Public Schools,” said Superintendent Carl Harris in the press release. “We will continue to focus on ensuring that we are using good data to make decisions around the delivery of instruction to our students.”
You will find out about the retests if you click on the attachments for the press release. They show a big boost from their inclusion.
For instance, the passing rate on the grades 3-5 reading tests was 45.8 percent, a slight gain from 42.3 percent the prior year. But with the retests included, the passing rate hits 54.3 percent
Barring any change, Wake hasn't planned on releasing the results before the state puts them out on Aug. 6.
Comments
This is a classic example of
Thu, 07/16/2009 - 08:39 — shearertwThis is a classic example of educrats doing what is best for them and not the students. The EOGs are already dumbed down in comparison to better, nationally recognized test and standardized test used by states with far superior education systems than in NC. Therefore, only the first score should count for school system reporting. Any student who fails or is on the borderline for failing the NC EOG would likely fail any of the more challenging standardized test of successful school systems. These kids are behind on a national level, period. If a school system wants to give them a 2nd try for promotion reasons, fine, but don't pat yourself on the back in the process because every kid that fails the NC EOG is a representative of the failed NC education system. Counting the re-test is just another way of hiding the facts.
different forms
Wed, 07/15/2009 - 19:11 — klanders65oops. wrong place
retests should count
Wed, 07/15/2009 - 15:08 — klanders65They should count. The tests have very low reliability at the student level. It makes mathematical sense to count them.
Huh?
Wed, 07/15/2009 - 15:48 — Bob_Sconce(1) The bigger issue is districts comparing with the previous year's numbers (without retests) then bragging about improvements. That's just deceptive. Any district that does not fully disclose the difference in the same press release is deceiptful, and providing an AWFUL example for the students. ("But, Johnny, your research report didn't mention all the studies that tended to disprove your conclusion" "It's OK teacher. The school district did the same thing.")
(2) Please define "low reliability" and state how you know that they have "low reliability". Does it only work one-way -- i.e. is it only the case that it fails students who really do know enough to pass? Or, is it also the case that sometimes it passes students who really should have failed?
The retests are nothing but a gimmick to make the schools appear to be better without actually having to be better.
The EOGs are not meant to
Wed, 07/15/2009 - 19:20 — klanders65The EOGs are not meant to give student specific information, but rather are for whole groups. There are different forms of the test and students will get questions that measure different objectives. That way, across a grade and a school, the results reliably tell us whether the students in a grade have mastered the objectives--both hard ones and easier ones.
They give very reliable information for grade levels in schools or for reasonably large groups of students (20 or so would have pretty good reliability.) They simply are not very reliable in the statistical sense at the student level.
Any individual student might score differently given a different form of the test. They might pass if given a test that measures specific objectives that their teacher covered really well. Some other student in the same class might get a form of the test that measures objectives that the teacher didn't get to yet or skipped, or maybe the kid didn't get those objectives. That kid might fail that form of the test. If he gets a different form, he might pass.
A student wouldn't be likely to pass a test if they didn't know the material on the test. There aren't false passes in that sense. But given two students in the same class who have mastered the content the teacher taught at the same level as each other might get different forms of the test. One form may have objectives these two students have mastered and the other form may include some they have not mastered. They both know the same things but one student passes and the other one fails. It is only fair to do retakes.
The class's test scores would show that some objectives are not mastered in this class. If we are going to make individual decisions using these scores, then we have to provide retakes. They never intended to use these tests for decisions about individuals. Now that they are doing so, they have to retest kids who fail. The statistical reliability at the student level is far too low to make decisions about individuals without getting more than one score.
You are right, they shouldn't compare different situations. They have last year's retake scores. They should compare apples to apples.
You ask how do I know? I read it somewhere on DPI's website in some paper explaining the reliability of the EOGs. I'll see if I can find it again. It was a technical document. I am probably the only one who ever read it.
So...
Wed, 07/15/2009 - 20:41 — Bob_SconceAs I understand how EOGs work (at least at the elementary school level), individual decisions are not based on the EOGs -- they're all about performance of the schools, not of the students. The EOGs do not determine whether or not a student moves on to the next grade level, nor do they play any part in that student's class grade. How could they? The 2008 EOG scores weren't known until half way through the following school year. What did they do, reach out to a 2nd semester 4th grader and say "You remember that test you took 7 months ago? well, you didn't do so well, so now you're going to have to retake 3rd grade"??
However, I don't think you answered the last part of my question, so I'll rephrase: re-testing assumes that, maybe, some level-2 kids just got unlucky with the few things that they didn't master showing up on their test. But, it doesn't assume that, maybe, some level-3 kids just got lucky with the few things that they did master showing up on their test. If your model of the test is correct, shouldn't you retest both those who were slightly below the cut-off AND those who were slightly above?
I understand the variability in individual scores on standardized tests. It's a basic problem with assuming that every person has a certain quantifiable ability level which can be measured. That's a bad assumption.
I may not understand this
Wed, 07/15/2009 - 21:11 — klanders65I may not understand this totally. I've tried to ask questions but no one seems to know. I found a reliability report again tonight on DPI's website. It talks about how there are different forms that cover different objectives more heavily, etc. And, that the error for an individual varies depending on where they fall on the bell curve. It is more extreme at the extremes.
They do the retests right away. They do the retests within days of the first test.
I believe at some grades, the EOGs do determine whether students move to the next grade level. I think that is part of NCLB. I don't see how they could legitimately use them for that but I think NCLB requires that they do. I think at grades 3 and 5 they are supposed to hold back kids who fail. I am not sure on the grades. I don't know how they handled the fact that the reading EOG scores weren't known until October or so. I guess they had to use math scores only for holding back.
You are right that the error goes in both directions. They don't retest the low level 3s. Your argument is right. They should, based on the same reasoning as retesting the high 2s. I guess I don't know that they don't. I assume they don't retest the low 3s.
I always want the kids in the most challenging rigorous situations because I see that they thrive there, so I don't want them to retest low 3s and put them back. And I do want them to retest high 2s and move them up.
It is not statistically correct. But, I think this is a situation where what politically makes the schools look good and what is best for the kids are the same. That is rarely the case.
I think they are doing the right thing with the retests, even if they are doing it for the wrong reasons. They do a lot of statistically unsound things that are bad for the kids. This is not really statistically sound but it is good for the kids so I am in favor of it. I want as many kids as possible in the most challenging situations because that is best for the kids.
Re: Moving up
Thu, 07/16/2009 - 07:42 — Bob_SconceSo, the reason I didn't believe the EOGs had any effect on promoting kids was the fiasco with the 2008 EOGs, where there was a new reading EOG test. Because of the new test, the schools held on to the scores for a good 6 months. I know of a (then) 3rd grader who got a 2 on his reading EOG, but only found out about it the following January (or maybe it was February), far too late to hold the students back.
My main complaint here is that retests will naturally improve pass rates, even if there's no actual improvement in performance. Yet, the districts are all crowing about this increase as if they were responsible for it. It's a horrible example for students by teaching them to take credit for things that they're not responsible for. I'm hoping WCPSS takes the moral high road here, but given past experience, I'm not betting on it.
I agree. Districts are
Thu, 07/16/2009 - 13:50 — NWRaleighMomI agree. Districts are taking credit they didn't deserve. Most importantly, some families won't be able to escape from failing schools, as they have right under NCLB.
Historically, the majority
Wed, 07/15/2009 - 23:48 — KeungHui (author)Historically, the majority of students who fail the EOGs have still been promoted anyway. Principals have invoked the right to look at the entire student's portfolio. There's also a reluctance to hold back a student too often and create an exessively overage students.
Glad to hear that. I teach
Wed, 07/15/2009 - 23:56 — klanders65Glad to hear that. I teach older kids so I wasn't sure what happened down in the lower grades.