September 12, 2007

Napa meet Radnor

Joanne Jacobs has a good post on the distinguished Napa High School's dirty little secret: a giant achievement gap between white and Hispanic students:

Although it’s rated by California as a “distinguished school,” Napa High School is in its second year of program improvement because it missed No Child Left Behind targets for bringing English Learners to grade level. The target was 22.3 percent; the category includes students who’ve been reclassified as fluent in English but haven’t tested as proficient for three years in a row.

Joanne then proceeds to use the data from School Matters to show just how large the achievement gap really is:

Despite its distinguished rating, Napa High has a large gap between white and Hispanic students (nearly all students are one or the other) and an even larger gap between poor and non-poor students.

I did pretty much the same thing when I wrote about the distinguished Radnor Middle School a few posts back which also sports a frightful achievement gap.

These schools are the rule, not the exception. Tony suburban schools are just as clueless about educating poor and minority kids as their decrepit, underfunded inner-city counterparts.


Anonymous said...

The poor and minority kids are at the same school, presumably being given the same instruction...


I wonder what accounts for the achievment gap? Something inherent in poverty? Race? IQ, you say?


KDeRosa said...

Holding instruction constant ...

There are some kids who will succeed without parental support.

There are some kids at the margin that will succeed once they get some parental support.

There are other kids at the margin that will not succeed even with parental support.

Then there are kids who won't succeed and who won't get parental support.

What do you think separates these kids?

Anonymous said...

I think we may agree, but are thinking about this topic in different ways.

I think the kids who won't succeed even with parental support are the ones who have parents that aren't eqipped financially or educationally to help them, OR -- their kid has a disability. And, if we are holding instruction constant, we must surmise that they are also receiving ineffective instruction.

I think that if you took a typical school (with typical constructivist curricula) where every student had an IQ of 100, but half the school had parents who were educated and affluent, and the other half was uneducated (no college, some or no high school) and poor, the poor kids would be failing at a drastically higher rate than the affluent kids. Of course this is just me hypothesizing, but I believe it is what would happen.

Smart people (educated) with means (money for tutoring, or know-how to help at home) ensure that their kids will perform.

Holding instruction constant, if it is effective instruction -- your scenario, I don't believe, would hold up. All the kids except those few (less than 1% of all children) with the more severe cognitive disabilities would succeed. The "gap" would not be there in terms of kids meeting grade-level benchmarks.

I believe the disparity in achievement between low and high IQ widens with poor instruction. It narrows with effective instruction. Engelmann's book "Give Your Child a Superior Mind" goes into these details. He chronicles how getting to a child early, and equipping him with the knowledge needed to succeed in today's constructivist schools can overcome in deficits in IQ, SES, etc., but moreover will actually raise a kid's IQ.

I guess ultimately, the philosophy I'm putting out there depends upon whether you believe IQ is made, not born. I believe it's mostly made, as it's not really a level of intelligence, but a descriptor of how the child performs intellectually compared against the norms of the population. I just think environment plays the big role with most people, but genetics plays the bigger role in kids with disabilities (which in turn amplifies environment for them compared to the non-disabled).

I don't know, that's just how I look at it.

KDeRosa said...

As I recall, there was a parental support model in Follow Through that performed poorly and failed to raise student achievement.

There is some evidence, however, that a high SES family can slightly boost the achievement of a low IQ child adoptee. But this is a far cry from asserting that parental support is a bigger factor than IQ when it comes to student achievement. That's why I say that parental support can help at the margin, but that IQ remains the determining factor as to how well a student will perform academically.

Even with effective instruction, the smart kids will outperform the dim ones. Zig, himself, has said this. In the DI testing, the dim kids received extra practice and the smart ones were often held back because there wasn't enough students to form a class. And, as you know, DI doesn't rely on parental support.

Effective instruction can temporarily raise IQ in children, but that's an artifact of how IQ is being measured, i.e., through academic performance tests. There will not be a rise in non-academic IQ tests, such as Raven's Progressive Matrices. In fact, even if IQ is boosted, it will revert back to it's pre-effective-instruction levels by adulthood.

Environment does play a role in academic performance and does affect IQ, but mostly negatively. harsh environments can depress IQ and prevent a person from reaching their genetic potential. Think malnutrition, lead exposure, poor prenatal care, being raised by low IQ parents, etc.

Anonymous said...

I've been working with a group of low income, primarily Hispanic students (17 ESL Hispanic, 3 African American). They all attend suburban schools in the California Bay Area.

I prepared them for the ACT last spring, primarily with content instruction, with a bit of test strategy thrown in. Most people think test prep is a waste of time, of course, but these results are hard to explain without test prep.

While their original scores on the diagnostic were well below average for all Hispanics, their actual ACT scores after instruction came in close to the national median in math and English. Their reading and science scores were lower, just off the overall Hispanic mean, although dramatically improved (reading by 50%).

I'm working with them now for the October re-test. They just took their first practice test on Wednesday, and I'm interested to see if they've retained their knowledge from last spring.

In short, I've had success in improving achievement levels in kids of color at suburban schools. Regardless of color, students in the lower half of the ability spectrum don't understand how standardized tests work. Once these students understand that the tests aren't rigged, that they aren't random, and most importantly that their scores aren't dependent on their teacher's opinion of them, they are far more vested in improving. I often explain that a test is a high performance event, far more analogous to a competitive athletic trial than it is to school. This puts the test in a context that makes sense--they can work, they can improve, they can see their improvement in the results.

My experience has thus far taught me that students in the lower achievement levels are probably far more capable than their school tests are showing. That doesn't mean they're all secretly brilliant, or that the IQ distribution you describe isn't accurate. However, they hold school in considerable contempt and, at the same time, feel a great deal of anxiety about their skills and their ability to perform. Take school out of the equation and given them a means of measuring their own abilities and they can surprise everyone.

My experience has also taught me that school tests and school grades are usually lousy assessment tools. I would never believe, for example, that a student had second grade reading ability until I'd assessed him myself.

I'm all in favor of testing, obviously, but the students have to be vested in the process, with positive results accruing to higher scores. These students don't care about their results on a test that has no impact on their lives, and the more that teachers stress their importance, the more contempt the students are likely to feel for the test. Students should be given a breakdown of their scores, a copy of the test, and the ability to see where they went wrong. They should have an understanding of what *they* need to do to improve, and a reason to demonstrate improvement by doing better on the test.

This is unlikely to happen, not because it's not good practice, but because the expense of creating tests and the fear of cheating offsets the students' best interests. But without this information, without giving students a reason to care about their scores *for themselves*, not for some abstract like real estate prices or a successful school rating, I don't see much hope of a meaningful reduction in the performance gap on the NCLB tests.

Anonymous said...


Loved your post. Seems as if you have hit the nail on the head with your assessment of your students and moved in the right direction to help them.

Not sure why we focus so much on IQ, environment etc. We cannot change that. Seems to me Cal has looked beyond the numbers to the kids.

Keep up the good work.

KDeRosa said...

Consider this.

With the ineffective instruction being doled out today, we're seeing lower-performers performing many grade levels behind the upper- performers.

With more effective instruction, we would expect to see an across the board increase in performance with perhaps a decrease in the amount that the lower performers are being the upper-perfromers. For example, maybe we could get the spread from six yeare down to two years.

However, upon testing the kids, we'd still see that the distribution would be the same. High IQ kids on top, low IQ kids on bottom.

This is what Cal's data may be showing.

Anonymous said...

DeRosa states:

"However, upon testing the kids, we'd still see that the distribution would be the same. High IQ kids on top, low IQ kids on bottom.

This is what Cal's data may be showing."

Are you saying that IQ trumps everything so why bother with effective or ineffective instruction? The low IQ kids will always be at the bottom no matter what we do.

Until schools teach reading correctly and give all kids a fighting chance from the start I will continue to believe that there is hope for all kids.

I just tested a group of first grade students who transferred into my school. Every single one of them was clueless about reading. Some knew a few sounds, some knew a few sight words but not one child had any idea how to read an unknown CVC word after one year of balanced literacy instruction. I showed them the title to a book I use, called I See Sam. Most of the kids could read the word /I/, a few could read the word /See/ but not one knew what to do with the word /Sam/. They all looked at the picture which did not have any clues for the word Sam. These kids are now behind the rest of their class. They have lost one year of correct reading instruction. I have no idea if they have high or low IQ's but we have already failed them so for me, IQ is meaningless.

btw, are you going to continue your reading blog on guided reading?

Anonymous said...

Be sure and read this article in today's Ed News email: It goes into all those reasons why achievment isn't happening for the minority and low SES kids.

See if one thing in that article makes a lick of sense to you.

Anonymous said...

Another thought just came to me...

If IQ is the biggest factor on low achievement, instead of SES/education of the parents, who are actually responsible for the kids' good performances instead of the school cruddy instructional practices....

Is there a suggestion that low IQ/high IQ is something inherent in race?

Because if the low kids with no parental support, and the low kids WITH parental support still don't perform, and we control for education being constant... we argue it is IQ. If in a given school, the kids performing that low are mostly minority... aren't we then sort of implying that race is a factor by virtue of genetics (the inherited IQ), whose IQs can only be raised "temporarily"?

If I'm leaping to logical conclusions, and am committing logical fallacy... please tell me. I dropped "Intro to Logic" in college because I had, er, other things to do with my time....

Anonymous said...

"Are you saying that IQ trumps everything so why bother with effective or ineffective instruction? The low IQ kids will always be at the bottom no matter what we do."

I think Ken's position is that with equivalent instruction the high IQ kids are going to perform better than the low IQ kids at academic subjects. This is largely what having a higher IQ means.

You still want effective instruction, though, because everyone moving up, say, two grades worth of skill is valuable even if the lower performing kids don't catch the higher performing kids.

It is, however, foolish to assume that given similarly effective instruction the lower IQ kids will somehow "close the gap" with the higher ID kids. They won't.

Ken, did I get this right?

-Mark Roulo

Anonymous said...

"With more effective instruction, we would expect to see an across the board increase in performance with perhaps a decrease in the amount that the lower performers are being the upper-perfromers."

Yes. I also suspect that what we call IQ is actually measuring our capacity for abstraction-- that is, fluid intelligence.

I don't think that all students in the lower half of the ability spectrum are of low abstract intelligence, but I begin with that assumption. Kids who can abstract and generalize but just haven't had the chance or desire to use their skills usually reveal themselves quickly by an astronomical increase in scores.

Unfortunately, we are making all the wrong decisions to better teach students in the bottom half. Rather than accepting that a student has weak abstraction skills, we insist that the way to academic success lies through algebra--a lousy math subject for a concrete thinker. Then we teach algebra as if everyone can easily grasp abstraction--again, exactly the wrong approach.

We should start them with geometry, and use proofs as a method of introducing abstract logic and generalization to these kids. Then we can work through algebra by teaching certain types of algorithms as *tasks*, not as open ended problem solving.

We should also accept that these kids will learn more slowly, and that most of them won't ever be interested in academics for their own sakes.

Unfortunately, any such finding would almost certainly show that students with more concrete than fluid intelligence would be disproportionately black and Hispanic, and thus there would be talk of educational ghettos.

One thing we could do to offset this is should also change our admissions tests to focus on achievement as much as aptitude--or perhaps have alternate tests. The SAT, GRE, LSAT, and GMAT are aptitude tests that reward abstract thinking almost to the exclusion of everything else. That's why I always use the ACT for lower ability students--it's just as hard a test, but it's very clear about what it wants. Colleges should consider demanding and using that test. For that matter, they could use the ACT, which (like the SAT) is as difficult as any of the grad school tests.

I suspect that higher intelligence translates to higher achievement no matter what. But Kathy's question--why bother?--seems odd. We bother because a student who knows how to read, calculate, and learn is far better off, to himself and society, than a student who doesn't. I want the top of our bottom half to be sales people, carpenters, small business owners, social workers or soldiers. I want them to have futures and not be stuck in a job ghetto just because they didn't want to go to college.

The idea that life is worth living only if you can understand--and *want* to understand--James Joyce is a notion that only the overeducated could have come up with.

BTW, if anyone is interested, I have the data for my class here:

Anonymous said...

Why bother? I got that feeling from Ken and all the talk of IQ. I bother every single day by trying to help failing reading students.

Cal, you sound like a thinking teacher who uses his data to guide instruction. I admire your work.

Anonymous said...

Mark said:
It is, however, foolish to assume that given similarly effective instruction the lower IQ kids will somehow "close the gap" with the higher ID kids. They won't..

I think we need to define what the "gap" is when we are talking about school performance.

If we mean the gap to be ANY difference in perfomance, then of course arguing that we can ever "close the achievement gap" is a silly argument. But we can, and have already demonstrated that we can, close the gap between those that make the standard cutoff and those that do not.

If the "gap" is those that are able to read at a third grade level by the end of the third grade (taking into account that SOME kids will be reading well beyond 3rd grade level by the end of third grade), it is absolutely possible, just like the military has pretty much "closed the gap" on getting all of their recruits to make it through boot camp/basic training.

I don't think anyone in K-12 educational circles, much less NCLB, is suggesting that we get to a point where all children learn equally and perform and achieve equally. What we are talking about is some definition of basic proficiency that meet an agreed on definition of what that basic proficiency entails. It doesn't have to be a college prep level by the time all of our kids get through high school, but it should fall somewhere around being able to competently understand a newspaper, fill out a job application, read and understand directions, understand basic grammar, have a basic knowledge of history, geography and general science, etc. (academics) and being able to write a basic report-type document that everyone can understand.

We need to define "gap" before we start comparing apples to oranges, IMHO.

Anonymous said...

"One thing we could do to offset this is should also change our admissions tests to focus on achievement as much as aptitude--or perhaps have alternate tests."

We could do this. And I wouldn't complain. But part of the reason that the SAT is an *aptitude* test rather than an *achievement* test was that way back in the 19xxs (I don't know how far back), someone noticed that smart kids with lousy instruction didn't do well on achievement tests. Measuring something that was *less* influenced by lousy instruction was considered progressive for the time.

So we *could* go back (and I wouldn't object). But we probably ought to acknowledge that we are returning to an old status quo, and not trying something new.

The pessimist in me suggests that we won't see a tremendous change in the populations that do well and poorly even if we do switch.

-Mark Roulo

Anonymous said...

"If we mean the gap to be ANY difference in perfomance, then of course arguing that we can ever "close the achievement gap" is a silly argument. But we can, and have already demonstrated that we can, close the gap between those that make the standard cutoff and those that do not."

No argument from me. If we achieved this, the result would be that the group differences for various grade level tests would tend to vanish -- most kids in each grade level would pass (a good thing unless we did this by reducing the cut line). But then at college admission time, we'd still see the current distribution (because the SAT isn't a pass/fail test).

This would be an improvement. I'd take it because more educated kids across the board is a good thing. But it wouldn't "solve" all the problems with the current race/SES distribution of achievement/aptitude scores.

-Mark R.

KDeRosa said...

Lots of good comments.

Mark is doing a great job clarifying the debate.

Yes, I will be finishing the Tale of Two Reading Program posts.

Anonymous said...


I'm not suggesting that the score distribution will change dramatically, or that high intelligence people will do worse on achievement tests. I do think achievement tests are a better measurement tool for more "concrete" thinkers (for lack of a better word).

At the time the SAT was instituted, admissions tests were in Latin, trig, physics, and so on. A bright kid with no instruction couldn't compete. The SAT was designed to identify bright abstract thinkers with exceptional verbal skills, whether or not they had a good education. In the years since, we've realized that bright abstract thinkers with exceptional verbal skills are overwhelmingly white. At the time, it was quite possible to be poor, bright, and white, so the SAT became an outstanding equalizer for this population.

The SAT is still an excellent tool for identifying bright abstract thinkers with great verbal skills. It's not nearly as strong at identifying the abilities of average thinkers with weak abstraction preferences.

General purpose achievement tests would give this latter group a clearer target. Most achievement tests, btw, are subject specific--for example, the SAT Subject tests. The ACT is the only general achievement test I know of and it's pretty well known that some students show a strong preference (and higher scores) on the ACT than the SAT. Three of the four major grad school tests are general ability tests, with no more than tenth grade math required, if that.

Miracles wouldn't happen. The ACT and the SAT have similar score distribution, with much lower scores for blacks and Hispanics. The ACT or other general achievement based tests would make a difference to kids with high concrete intelligence and few abstraction skills, but the difference would be on the margin. However, the margin would be a good start for grad school.

Remember, I'm suggesting an education sorting method that would have distribute races disproportionately. I brought up the grad school achievement test option as a way to assuage the naysayers in the mythical land that adopted my proposal.