November 20, 2006

NYT perplexed over achievement gap

The NYT is perplexed over the achievement gaps between whites and blacks and hispanics.

Despite concerted efforts by educators, the test-score gaps are so large that, on average, African-American and Hispanic students in high school can read and do arithmetic at only the average level of whites in junior high school.

"The gaps between African-Americans and whites are showing very few signs of closing," Michael T. Nettles, a senior vice president at the Educational Testing Service, said in a paper he presented recently at Columbia University. One ethnic minority, Asians, generally fares as well as or better than whites.

The reports and their authors, in interviews, portrayed an educational landscape in which test-score gaps between black or Hispanic students and whites appear in kindergarten and worsen through 12 years of public education.

Some researchers based their conclusions on federal test results, while others have cited state exams, the SATs and other widely administered standardized assessments. Still, the studies have all concurred: The achievement gaps remain, perplexing and persistent.

Why is this so perplexing?

Because of political correctness, that's why.

The achievement gap mirrors the IQ gap. Of IQ gaps we cannot speak amongst polite company, such as those that still read the NYT.

That's the elephant in the room.

Even the legislators behind NCLB were not so crazy to think that NCLB was going to magically erase the achievement gap. You can search all day long in the legislation, but you won't find anything that suggests that the real performance of all the racial groups will be the same in 2014.

What you will find is achievement measured by defining a cut score over which even our dullest students will be able to jump over into proficiency land. However, once they're in proficiency land there will still be an achievement gap. And, that gap will be the same as it is now due to the persistence of the IQ gap.

The aim of NCLB is to set a minimum competency threshold and get all students over it. The theory goes something like this. States set a proficiency level as low as is politically feasible, then schools improve teaching until 99% of students can meet or exceed the goal. If the goal were set much below 99%, we'd still see an achievement gap due to the IQ gap, so don't expect to see that level reduced below about a mid ninetieth percentile.

The best NCLB can accomplish is masking the achievement gap.

Insta-update: Guest blogger Jal Mehta at Eduwonk either toes the political correctness line like a good think tanker that wants to keep his job or gets it wrong:

[F]rom a policy perspective, I think (and here is that Rorschach test) it reveals the weakness of external accountability as a lever to accomplish what is one of the most difficult of our social policy objectives -- breaking the link between family poverty and school outcomes. In a sense, setting standards and measuring progress through assessments are the low-hanging fruit -- what they leave undone is the much more difficult task of actually helping kids do tomorrow what they couldn't do today. These questions--of which accountability systems are one part but a relatively small part--is where the policy dialogue needs to go. Let's hope today's story sounds the bell and starts that conversation.

If this is the typical substitution of "poverty" for "IQ" wink-and-a-nod that we see when talking about racial achievement gaps, we have the problem that "
breaking the link between [IQ] and school outcomes" is all but impossible. Smart kids learn more per unit of instruction and that's that. Otherwise if Jal really means that poverty is somehow responsible for diminished school outcomes, he has a wet streets cause rain problem. He's got the cause and effect thing backwards. And, he's ignoring the effects of the IQ gap. An immigrant family from one of the Northeast Asian countries who might be living in poverty today will likely have children who perform well in school, whereas your native inner city family who just won the powerball lottery will likely still have children who won't do well in school. Poverty has little to do with school performance, especially the way we define poverty in this country which includes many families who would be considered filthy rich in most of the rest of the world.

And, as far as the "
what [accountability systems] leave undone is the much more difficult task of actually helping kids do tomorrow what they couldn't do today" meme goes, I view this as a good thing. NCLB is basically saying, you've been taking our federal money for a long time now with little to show for it. We don't like it. If you want to keep on taking our money, you need to start showing results. Here's what we want to see in 14 years. You set the standards as you see fit and you, as the experts in education, find a way to meet those goals, in whatever way you want to. You want to teach to the test? fine. You want to expand the school day? go ahead. You want to adopt a sunshine and lollipop curriculum? Great. Just don't come crying crying to us, if you don't meet your goals. Little did they know, educators would be crying in less than five years.

Super update Two: Joanne Jacobs writes that:

[W]e could cut the gap considerably by giving disadvantaged children more experienced and expert teachers, a well-designed curriculum, frequent diagnostic testing and after-school tutors who could provide the support their parents can't. I also think we could work more closely with parents to tell them what they could be doing to help their children do well in school. Maybe the gap wouldn't close completely but it can be narrowed.

Sure, we can raise student achievement of low performers by improving their curriculum. I'm on board with that. But, I don't think, it'll diminish the achievement gap. How long do you think it would take for middle-class families to start demanding the same improved curriculum that the poor kids are getting? Once they get it too, their achievement will also rise. Back to steady state, i.e., achievement gap.

And guestblogger Jal Mehta at Eduwonk begins construction on what appears to be a strawman when he characterizes my position as "assuming that racial gaps are [] immutably given by nature." I see where he's going with the "immutable" language. We've already squeezed out most of the immutability about as far as we're going to take it, now that the poor are super-nutriated. Closing the gap further by reducing "poverty effects" have proven elusive and starts to bring us into the land of Kozol.


Anonymous said...

Dude you are going to be in trouble with the PC police.

I have been debating all morning whether to post the same basic thing.

KDeRosa said...

lol. one can only hope.

Anonymous said...

So you don't think much of Hirsch's work about the value of directly teaching his core knowledge series when it comes to lessening performance gaps?

KDeRosa said...

Hirsch may have the theory right, but what he has not been ablw to do is find a methodology to get all that information into a lower performing child's head. engelmann, who knows a thing or two about inducing learning, doesn't think that the average kid, let alone a lower performing kid can learn everything set forth in the CK curriculum.

Onwe problem is that all those concepts and vocabulary are hard to teach efficiently. See my recent post on this. Learning vocabulary, for example, remains largely a linear additive set which is not amendable to generalized instruction.

Even if we were able to make this kind of learning within the ability range of low performers an achievement gap will likely persist. This is because, if group A is capable of learning faster on average than group B, group A will always have an advantage over group B.

MassParent said...

Connecting the dots, suggested reading "proficiency for all - an oxymoron".

Any threshold based test that requires something to approach 100% of population crossing the threshold has to, by definition, be something that the 98th percentile of the distribution can achieve - and then you fudge the data by allowing an "error band" of a couple percent. That standard would still be very difficult of a school that happens to be in the bottom 25% of distributions, but an average school could achieve it.

And as best I can tell, none of the states have done that.

As a practical matter, most states set a proficient standard following the lead of states like Massachusetts, where proficient is somewhere between the 20th 50th percentile of the distribution. A couple or three standard deviations in the distribution away from the threshold.

So, as the study concludes, it would be much easier for the US to lead the world in math than it would be for the US to achieve 100% NAEP proficiency in math.

If you agree that IQ determines ability of students to learn, then any standard that is even moderately challenging for an average school population is practically impossible for schools where the population is a standard deviation below the average (that is, at the 15th percentile).

KDeRosa said...

But, Massparent, there are interventions out there that on average boost performance by a standard deviation. And, for schools that implement the intervention with fidelity, the boost in achievement can be as high as 2 to 3 standard deviations. If we do the math, we see that a typical title I school (20th percentile) that boosts its perforamnce by 2 standard deviations, can get about 85% of its students to passa legitimate test like the SAT-10, CTBS, ITBS. make the test a little easier and you'd be able to get almost all students over the hump.

Of course, I'd rather have a sliding pass-rate scale coupled with a more rigorous test that recognizes that low performers are not evenly distributed in the population.

Anonymous said...

Ken, I see your point about the lack of evidence that all kids can be taught the core knowledge series.

I think though if they could, the future is a bit brighter, if Hirsch's theories are right (never having tested them myself, I'll retain some tentativeness here). To the extent that the black/white achievement and IQ gaps are caused by a lack of cultural knowledge that could be explicitly taught, and if Hirsch has identified the right set of knowledge, then the students with a poor background will be gaining more value from being taught what luckier kids already know.

Also, I think the DI approach of making sure students have a very low error rate, and a high success rate, could improve the attitudes of poorer-performing students, and have an impact that way, in a way that is not so open to high-achieving students.

So, even if you're right about IQ being basically immutable, I think there is still the possibility that a good curriculum could have an impact on the achievement gap.

KDeRosa said...

Thing is, what's causing a large part of the lack of cultural knowledge is the kids' and their parents' lower IQs. They are not learning as fast as their higher IQ peers. So even if Hirsch has properly identified what they need to know, the problem remains how to get all that information into their heads. DI works wonders with knowledge that can be accelerated, but vocabulary acquisistion is notoriouslt difficult to accelerate. DI has made inroad here, but teaching all the background knowledge that lower performers need to know remains a problem.

Anonymous said...

RE: the immutability of IQ check out the work of Reuven Feuerstein and Instrumental Enrichment. His results in Israel refute the immutability of IQ and sees IQ as rather what a child has been taught rather than what a child can absorb. Feuerstein has data to prove this.

MassParent said...

Ken, of the interventions that boost performance that much (still not nearly enough to make NCLB mandate work through more than a couple or four years, except in schools that are already above average), if there are any that replicate well, hopefully we'll find out soon.

Replication is hard, though. Lots of programs over the years have worked well as pilot programs, or where you've got above-average staffers. I can see it now, NCLB 2.0; 100% above average teachers.

Instructivist said...

How do you explain the success of the same IQ-afflicted students in schools like KIPP and Amistad? I think many avenues need to be explored before resorting to the immutable IQ explanation.

KDeRosa said...

Massparent, you're point is well taken but good whole school interventions like SfA and DI have had many thousands of kids go through them successfully.

KDeRosa said...

Charles, KIPP and Amistad are getting those results becausethey are providing exemplary teaching and increasing instructional time. Put a bunch of high perfomers in these programs and crank up the pace and you'll get the same kind of gap.

My anecdote for today is that I just completed the entire first grade curriculum of Connecting Math Concepts (Level B) with my son in less than three months with about 1/2 hour of instruction a night.

Anonymous said...

I think we've been through this before, perhaps on the KTM site. The problem I have is that many think the goal is to close the gap. The real problem is to raise the results by quie a bit, whatever the gap might be.

I see schools that can't seem to get kids to master anything. They don't take any responsibility for learning or results. Our public schools finally try to get kids to learn their adds and subtracts to 20 in third grade. Even my son's private school has kids in 5th grade still figuring out 15-7 with their fingers. They just go through the motions.

Look at any state test and their minimal proficiency cutoff. Look at the questions and look at the results. These tests are so simple and the cutoffs so low that there should be no "gap" in results based on anything. IQ is really a non-issue at this point. The problems are bad teaching, low expectations, and no accountability. If you focus only on a relative gap, you miss the bigger picture of absolute or externally (world-class) defined expectations.

Think of a 3D graph where the independent X and Y coordinates are IQ and quality of teaching (or curriculum). The vertical Z, or dependent axis, is performance, perhaps meeting state proficiency requirements. As the teaching (or curriculum) is improved, the dependent variable should go up dramatically - the level is so low. In the IQ direction, the gap might decrease, or it might increase. If the dependent axis is only to get the kids over a minimal cutoff, then the gap would close because it says nothing about the overall difficulty of the curriculum; only about meeting the low cutoff.

If the dependent axis measured average performance on more difficult tests (say, based on Hirsch's Core Knowledge series), then you might see a wider gap based on IQ as the quality of teaching or curriculum is improved. This isn't necessarily bad or wrong.

Good teaching is good teaching. Good curricula are good curricula. It doesn't matter what the IQ or gap is. I like to look at individual students. Set high expectations. Use good curricula and teaching methods. If the student can handle the work, it doesn't matter what their IQ is or where they are from. The real gap problem is more about an opportunity gap, rather than a results gap.

Those who focus only on a relative results gap are the ones who bring IQ or SES into the debate.

Instructivist said...

"Use good curricula and teaching methods."


The obstacles are formidable since the dominant ed creed is well entrenched.

However, even with good curricula and teaching methods you are still facing pupils with bad attitudes, poor motivation and atrocious behavior. There is something very wrong with the school culture. I am more and more convinced that decades of progressive, child-centered education (and the consequent demotion of the teacher) fosters this bad culture.

Anonymous said...

" are still facing pupils with bad attitudes, poor motivation and atrocious behavior."

Unfortunately, this is too easy an excuse for not trying. In many schools, there are few of these student issues. Also, many of these issues don't really show up until the kids are older.

However, if schools cannot or will not control or separate these students, then they must allow parents to take the money and go elsewhere. Of course, my position is that there is no consensus or common ground in education. This, in itself, demands full choice.

"I am more and more convinced that decades of progressive, child-centered education (and the consequent demotion of the teacher) fosters this bad culture."

I think that many more kids slide along year-to-year without meeting any specific education goals. Teachers don't expect much from them. The kids don't expect much from themselves. School slowly becomes a real drag for kids. Many parents I talk to might not know much about pedagogy, but they sure know about low expectations. It seems to me that to do well in school nowadays REQUIRES more parental invlovement. When I was growing up, my parents hardly ever looked at my homework.

In the "old days", the curriculum or pedagogy might not have been great, but there was the threat of flunking. This also kept the teacher in line. Grading meant something real and the teacher had to be able to defend those hard numbers. Nowadays, there are rubrics, which give all sorts of information, but are, in the end, meaningless and have no teeth. The kids don't have to try and the teachers don't have to try.

The other bad influence I see is all of the child-centered group learning. You never see desks lined up in rows in classrooms. It's forbidden. The kids never really learn self control or the ability to focus, for that matter.

I see this more now that I am in charge of one of the school's First Lego League robotics teams - which is all about child-centered learning. The kids are bouncing off the wall. It's not attitude or not caring. I'm convinced that it has to do with the fact that they never learn to sit still and listen to a teacher. They think that all learning has to be fun.

They all know how to break up into groups, but that doesn't mean that anything gets done. One of the teachers helps out and is constantly telling (yelling) kids to get back to work. The most commonly heard phrase is: "What are you supposed to be doing."

Even when I try to do a little bit of carefully-led discovery learning, many of the kids are off in their own worlds fooling around. I asked my son if this is normal and he said yes. It's a wonder that they learn anything.

Instructivist said...

"Also, many of these issues don't really show up until the kids are older."

You'd be surprised. This refusal to learn can be encountered in the early grades among the disadvantaged.

"I'm convinced that it has to do with the fact that they never learn to sit still and listen to a teacher."

That's exactly what I meant. Progressive education (now called constructivism, inquiry, discovery, collaborative learning or what have you) fosters this bad culture. In this environment a teacher committed to knowledge-based instruction is hopelessly out of place. It's like swimming upstream in a river of molasses.

Anonymous said...

Funny thing about IQ tests. Get 2 groups of black kids to take an IQ test. Tell the first that they're taking an IQ test. Tell the second that they're taking a test of pattern rhythm (or whatever - something that doesn't sound like a test of intelligence).

Guess which group will score better?

Anonymous said...

>The other bad influence I see is all of the child-centered group learning.

That's a curious observation. I've seen good implementations and bad implementations of child-centered learning (it is not always group based). The Japanese are spanking us on international exams, and their early education is heavy on child-centered learning, often done in groups.

So I wonder why you would think that child-centered learning is bad...

>The kids never really learn self control or the ability to focus, for that matter.

Ah - so that's why. Why would you think that this is the fault of curriculum, and not teachers or culture? The Japanese students far surpass our own when it comes to "self control or the ability to focus."

I realize that child-centered teaching is high on everyone's bashing list (not without reason), but perhaps we should be fair. There is no one way of good teaching.

National Mathematics Advisory Panel, (appointed by the Bush administration) found "no basis in research for favoring teacher-based or student-centered instruction." This finding is echoed by other think thanks, some of whom are conservative (RAND).

The perception of student-centered learning being inadequate is based on the abysmal manner in which this teaching method was implemented. For the most part, teachers were given poorly designed curriculum and very little support, and weren't even taught why student-based teaching might be a good idea (or when it might be a bad idea). If anything, it'd be a miracle if student-centered learning succeeded in this environment.