o_0

[framefolly]

Thank you for the thoughtful comments I received when I asked about your reading/viewing experiences of To Kill a Mockingbird -- they are really helpful! I will reply to them, but for now, I can't resist directing your attention to something incredible:

Firestorm Erupts Over Virginia's Education Goals

Seriously? SERIOUSLY?!?! I thought Sue Sylvester

Comments

[zandperl]

Florida is already doing the same. What's worth noting in the Florida case (and perhaps in the VA, I haven't read through the whole article yet), is that while the absolute standard is higher for whites and lower for blacks/Latinos, that the improvement goals (called gains below) is higher for blacks/Latinos and lower for whites. Let me introduce you to a few concepts in assessment I've recently learned to show you why this is an important thing to consider, and my apologies if you're already familiar with these.

Let's say you give a final exam in a course (called a post-test in the jargon). You've got two students A and B in the class who both receive 75 points out of 100 on the final exam. At first glance, these students seem to be equal at the end, and based on traditional grading methods, they both have a 75 so they both have a grade of C on the final exam, case closed. But let's say you gave the same exam as a pretest, before the students had any instruction. If student A got a 25 on the pretest while student B got a 50 on the pretest, then you can see that student B improved a lot more than student A. We can put a number on that improvement by subtracting the pre- and post-test scores, and call that the gain = post-pre = final-initial. Student A has a gain of 50 (points), while student B has a gain of 25, so you can see that student A improved more than student B. I would say that student A learned more than student B here, b/c student B came into the course with more prior knowledge (or test-taking skills, or whatever) than student A.

Now what if we had a third student C, whose pretest score was 25 and post-test 50, and a fourth student D with a pretest of 70 and post-test of 95. Now three students, B, C and D, all have gains of 25, so you might be asking, which of those three students really did improve the most? That's where a last term comes into play, the normalized gain. Normalized gain says "out of the number of points which the student could have improved, how much did she actually improve?", which is calculated as (post-pre)/(100-pre) and turn that into a percent.

Student B = (75-50)/(100-50)=50%, so B gained half of the knowledge she could have.
Student C = (50-25)/(100-25)=33%, so C only gained a third of the knowledge she could have.
Student D = (95-70)/(100-70)=83%, so D gained most of the missing knowledge.

Another way you could look at this, is that if a student already has a score closer to 100, it's harder to bring up your score by 25 points than if you start off with a score closer to 0.

You could even have a student E, who goes from 75 pre-, to 80 post-test, this student would have a normalized gain of only 6.7%, so even though student E has a higher final exam score than students A, B, and C, student E had a worse normalized gain than any of those students. In fact, student A's normalized gain is 67%, ten times higher than the normalized gain of student E, even though student E has a final exam score of 80 while student A has a final exam score of 75. Student A clearly learned more than student E.

So how this is relevant to the current topic of FL/VA standardized testing scores, is that black/Latino students start off at a lower level than white students - black/Latino students are most like student A going from 25 to 75, while white students are most like student E going from 75 to 80. Even though the absolute score standard that black/Latino students are being held to is lower than that of white students, the gain (or normalized gain) standard they're being held to is higher, and (if we assume that standardized test scores actually measure knowledge and learning), the black/Latino students are actually being expected to learn more than the white students, because they have to make up for the unfortunate deficits they come into K-12 schools with.

In conclusion, I feel that gain (or normalized gain) standards are more relevant than absolute score standards, so Florida is doing things right IMO, though I have to check the article on VA to see if they're using gain standards.

[zandperl]

Reading the VA article you linked, and re-reading the NY Times FL article which I liked, it looks like both states are looking not at gains in score standards, but gains in overall pass-rate standards. The same argument holds for those, but IMO is not as strong an argument as if we looked at individual students score gains.

Also, the VA article doesn't explicitly state the "pre-" pass rates as does the NY Times FL article, but does refer oblique to the pre- in the quote "we're starting with black children where they are. We can't start them at the 82 percentile because they're not there."

[framefolly]

I must be brief because I really should be sleeping ;) .

Thank you for explaining a possible motive for these policies so clearly. I understand the importance of considering gains in drafting goals.

What I object to is conflating pre-test performance with race/ethnicity. Why not set separate goals for students who perform at different levels? Why go through the intermediate and inflammatory step of race? Doesn't this do a disservice to the high-performing Black, Latino, and disabled students as well as the low-performing Asian and White students?

Policy that is tied to skin color is problematic given this country's history. It also can promote racism in people who confuse correlation with causation.

[slaymesoftly]

Agreed - totally. Children are individuals and should be judged accordingly. If the students who have the most to gain are (mostly) black, so be it. That doesn't mean that all the black (or Latino, or Native American, or redneck, or ADD or whatever) should be evaluated based on their defining characteristics. If you want to evaluate how well the school/teacher/student is doing based on the gains made, that's fine. But don't measure those gains based on what you expect that student to do based on his/her color or ethnic background.

[curiouswombat]

I think, as an outsider, what surprises me is certainly not the possibility that they should consider what is often called in UK education the 'value added' - but that they are not actually interested in what the start scores are for individual students, or for individual schools - rather they are using racial stereotyping.

So an Asian kid who had been brought up in the care system, and attended what we have been known to refer to as a 'sink school,'* would be expected to do so much better in their exams than your president's children, just to get a pass...

* "People use sink school to refer to a school that is in a very poor area with few resources" (Collins dictionary).

[zandperl]

You (and framefolly and slaymesoftly) make good points. I think the biggest problem with VA and FL's method is that they're looking at pass rates, which cannot be tied to individuals (an individual either passes or doesn't pass) and thus must be tied to a group. A better way of looking at the pass rate could be something like "we want to take the group of students who failed the 4th grade exam, and have X percent of them pass this year," or "let's take the group of students who got a score in the range of A-B on their grade's exam last year, and have X percent of them pass in this year." But I still would rather see them looking at score gains rather than pass rate gains. The passing score is arbitrary, while looking at whether a student gained knowledge (or testing ability or whatever standardized tests actually test) does indicate a gain regardless of where you put that arbitrary passing cutoff.

As for whether it's racial stereotyping and your hypothetical Asian kid who'd attended an inner city school*, since these standards are being applied to the pass rate of the whole group of all Asian students, this one individual student wouldn't (ideally) be held to a higher standard than a black kid, just the group of all Asian kids. The purpose of doing this is that black/Latino students have historically underperformed and the states want to draw attention to this and to "close the achievement gap". Honestly, I think this puts more pressure on the teachers in the inner city schools than it does on the students of any race. But I do agree that it is problematic, and that most people can't separate in their heads between the individual and the group, and it'd be much better to treat each student as an individual (such as setting goals for everyone to have a normalized gain of 50%), or if you must group them then group them by the individual's previous performance.

* Inner city school is a US term referring to schools in urban low income areas. There isn't any equivalent term for rural low income schools - they have some similar problems to inner city schools, but they tend to get less media attention. Inner city schools tend to have high black/Latino populations, though sometimes they have large immigrant populations too, for example I student taught at a school in Springfield MA where there were nearly as many second generation Russian children as there were African American children.

Value added hasn't yet caught on here in the States, one of the reasons being that the standardized tests are widely regarded by educators as being worthless. If you're not really measuring anything useful with the test, you can't know if the teacher is actually adding any value.