August 22, 2008

Postal Service More Loved Than Public Schools

According to Lisa Snell:

An August 2008 poll conducted by Education Next and Harvard University finds that Americans think less of their schools than of their police departments and post offices. When asked to grade the post office, 70 percent of respondents gave an "A" or "B." In contrast, only 20 percent of Americans said public schools deserve an "A" or a "B." Twenty-six percent of the country actually gave their public schools a grade of "D" or "F." And African-Americans are even more down on public schools, 31 percent gave public schools a "D" or an "F."


I'm not surprised. The post office delivers my mail faithfully, albeit expensively and with a substandard tracking system, regardless of my social status, my ability to receive mail, or my mail receiving style.

August 21, 2008

Learning Styles Are Bunk

Dan Willingham has another video out on the non-existence of learning styles.



Willingham goes into much more detail on learning styles in his Summer 2005 American Educator article: Do Visual, Auditory, and Kinesthetic Learners Need Visual, Auditory, and Kinesthetic Instruction?

Vicki Snyder also makes a similar point in Myths and Misconceptions about Teaching: What Really Happens in the Classroom. Learning styles are presented as the fifth myth of teaching. (I reviewed the book here.):

Myth #5: the myth of learning styles refers to the popular idea that teaching methods should be matched to students' unique characteristics. Although individualization is desirable, learning style assumes that certain learner characteristics are intrinsic when they may in fact be the result of experiential factors that are amenable to instruction. As a result, teachers may inadvertently deny low-performing students opportunities to learn.


The myth of learning styles is based on three faulty premises: learning styles are intrinsic, learning styles can be assessed, learning styles can be matched to instructional styles. Snyder points out that all three premises are untrue.

In any event, as far as teaching goes, we only really care about the differences and similarities that influence learning and instruction. Of course, the vast majority of differences between children have little or nothing to do with how kids learn. Often these differences are expressed in terms of "learning styles and modalities," "multiple intelligences," and "differing interests." All of these so-called differences are similar in that none has any empirical support nor has any been shown to have an effect on learning or instruction.

This is because the content of instruction dictates about 90% of what has to be taught:

Content, and the nature of content, doesn't change according to the interests of children, nor according to any other characteristic of children. If we were trying to teach a gorilla to read, the nature of reading wouldn't change. Obviously, when it comes to the nature of content, differences among learners don't have much to do with anything.


Learning style differences are usually assessed informally through teacher observation. Teachers, however, often know little about inducing real learning. These learning styles are often expressed as superficial external traits like visual, auditory, tactile or kinesthetic which mask the underlying complex cognitive traits. For example, children cope with their inability to read in ways that might superficially seem like a learning style, but that actually reflect poor reading skills. It's easy to misinterpret certain behaviors.
Consider the following examples.

  • Sometimes elementary teachers say that poor readers are auditory learners because they can't track words with their fingers. It's more likely that they can't read the words. Usually these auditory learners can keep their eyes riveted to a television or video game screen for hours.
  • Sometimes elementary teachers say that poor readers are visual learners because they memorize and rely on picture clues rather than sounding out words. It's more likely that they revert to visual clues because they can't read the words. Without knowledge of the underlying sound structure of language, they have little choice but to rely on memorization and guessing.
  • Sometimes high school teachers say that poor readers are auditory learners because they need the text read aloud or explained to them.
    Sometimes high school teachers say that poor readers are visual learners because they need pictures, graphics, and visual displays to explain the text to them.
  • When students are labeled tactile/kinesthetic learners, they often need hands-on experience, group work, and activities to learn, not because of their learning style but because they need structure, assistance, and feedback on difficult or unfamiliar tasks.

In all of these examples, the source of the observed behavior is poor reading skills. To ignore the basic problem in no way benefits the students.

The point is that all kids (and humans) share some characteristics that are useful for learning, and, therefore, instruction has to accommodate those samenesses among learners, rather than the many differences among them. Learning styles and "intelligences" and student interests and modalities couldn't possibly have too much influence on learning, not when the nature of content doesn't vary among learners, and not when some of those things that make us all human are so central to learning.Efficient instructional programs make every effort to communicate the essential nature of content to all learners (because it is the same for all learners), and they make every effort to take full advantage of the ways all humans generalize more accurately and efficiently. What is the same about children is their innate capacity for language, to learn to read and think inductively and deductively. What is the same about all children is that they will learn if given appropriate instruction. They may learn at different rates and may need different amounts of structure and practice to master academic skills and concepts, but they can learn.

August 13, 2008

The IQ Conundrum for Broader, Bolder

Here are some charts from Gersten, R., Becker, W., Heiry, T., & White. (1984). Entry IQ and yearly academic growth in children in Direct Instruction programs: A longitudinal study of low SES children. Educational Evaluation and Policy Analysis, 6(2), 109-121. that show the gains made by the low-SES DI students in Project Follow Through for a range of IQ blocks from under 71 (2 sd below the mean) to above 130 (2 sd above the mean).

There are six IQ blocks shown on the chart. From left to right:

Block One: IQ below 71
Block Two: IQ between 71 and 90
Block Three: IQ between 91 and 100
Block Four: IQ between 101 and 110
Block Five: IQ between 111 and 130
Block Six: IQ above 130

For each IQ block the mean standard score has been graphed at the end of grades 1, 2, and 3.

There are arrows (<, <<, <<<) along the Y axis (mean standard score) that show the national median for each grade. I've (helpfully) drawn a blue line at the third grade national mean, as you can see, only the kids in blocks with IQs above 100 are performing above about the national median for math and only those above 110 for reading. (The blue line only has meaning with respect to the third grade scores (the top point). You could draw horizontal lines from the double arrow (second grade) and compare it to the middle point and from the single arrow and compare it to the bottom point.)

Click on each chart to enlarge.

This chart is for total reading for the Metropolitan Achievement Test.



This chart is for total math for the MAT.



Here is Becker's interpretation of the charts:

The data showed almost no contribution to "learning rate" (pretest to posttest gains) for IQ. If IQ were correlated with gains, lower-IQ children would make smaller gains and higher IQ children would make larger gains. This does not happen for Reading on the Wide Range Achievement Test (decoding) [Ed: Not shown.] or comprehension on the MAT, there is no IQ effect gains from the end of grade one to the end of grade two (most of the gains are about equal), but there is an effect for the gain from the end of grade two to the end of grade three. I believe this effect is due to the fact that the end of third grade test for Reading Comprehension on the Metropolitan uses an uncontrolled, adult-level vocabulary (as found in fourth grade texts). Since vocabulary instruction in school does not progress gradually to the adult level (but jumps from a carefully controlled vocabulary to an adult vocabulary after third grade), the test at this level is now measuring something not taught in school. Thus, students who score higher on a test of verbal skills (IQ) do better on a test of verbal skills (Reading Comprehension) when the content was not systematically taught in school. (A caution: The data may have imposed a ceiling effect on the brighter students; the program stressed preventing failures and thus teachers may have given more effort to teaching lower performers. Even if this is the case, however, the data are noteworthy in showing what can be done "gainwise" for lower-IQ children.)


Here is my observation. I understand Becker's comparable gains argument, but look at the mean percentile ranks for each IQ block:

Math End of Third Grade

Block One (IQ below 71): 24th
Block Two (IQ between 71 and 90): 39th
Block Three (IQ between 91 and 100): 47th
Block Four (IQ between 101 and 110): 61st
Block Five (IQ between 111 and 130): 69th
Block Six (IQ above 130): 88th

Reading End of Third Grade

Block One (IQ below 71): 11th
Block Two (IQ between 71 and 90): 29th
Block Three (IQ between 91 and 100): 34th
Block Four (IQ between 101 and 110): 44th
Block Five (IQ between 111 and 130): 58th
Block Six (IQ above 130): 81st

Also notice the gradual slippage from first to third grades in Reading even for the smartest kids. There is no slippage in math. Interesting.

I don't see how the lower IQ kids are going to be able to learn in a regular classroom given these percentiles. That would seem to foreclose a college education for these students and probably an academic high school education. Am I wrong?

And for the Broader, Bolder crowd, given that many low-SES students have lower IQs and that SES inerventions have not been able to to show a significant effect on IQ past about third grade, how exactly are your proposed SES interventions going to get around this IQ conundrum. Look the high-IQ, low-SES kids are performing well. The low-IQ ones aren't. I'd like to hear a rational argument that makes sense of this.

August 11, 2008

Day 17: Still Waiting


Two weeks after I first called for some evidence on the effectiveness of Broader, Bolder, I finally received a (sort-of) response from Big-Labor Fat-Cat Leo Casey.

Leo must have had a few of his underlings poring over the ERIC databases non-stop finding the requested evidence. Here is Leo's evidence. I am leaving in all the internal citations and footnotes.


Classroom teachers recognize immediately the educational value of providing a comprehensive array of services to students living in poverty. They have seen the effects of undiagnosed and untreated eye problems on a student’s ability to learn how to read, and of untreated ear infections on a student’s ability to hear what is being said in the classroom. They know that the lack of proper medical care heightens the severity of childhood illnesses and makes them last longer, leading to more absences from school for students who need every day of school they can get. They have seen asthma reach epidemic proportions among students living in poverty, and they know that the lack of preventive and prophylactic medical care leads to more frequent attacks of a more severe nature, and more absences from school. They understand that screening for lead poisoning happens least among children in poverty, even though their living conditions make them the most likely victims, with all of the negative effects on cognitive functions. They know that the stresses of life in poverty make mental health and social work services for students and their families all that more important, and yet they are least likely to receive them. They see how the transience that marks poverty disrupts the education of students again and again, as the families of students are constantly on the move. In short, teachers know that the students living in poverty lack the health and social services routinely available to middle class and upper class students, despite the fact that they need them even more. And they know that the absence of these services has a detrimental impact on the education, as well as the general well-being, of students living in poverty.


I emphasized Leo's evidentiary citations since they do not conform to the generally accepted norm. Leo's logic goes something like this: Leo knows best because Leo knows best. The circularity of this argument is surpassed only by its arrogance.

There is, of course, little actual research backing up Leo's claims. This is fortunate for Leo since in the few instances where there is research, it proves Leo wrong. Let's take a look at one of those claims.

They have seen asthma reach epidemic proportions among students living in poverty, and they know that the lack of preventive and prophylactic medical care leads to more frequent attacks of a more severe nature, and more absences from school.

As luck would have it, we actually have legitimate research on the efficacy of an asthma intervention. Here are the results.

  • An asthma self-management program incorporating health education and parental involvement increased academic grades for low-income minority children but not standardized test scores. (Evans et al.)

  • A subsequent study of the asthma self-management program was expanded to include health education for asthmatic children and their classmates, orientation for school principals and counselors, briefings for school custodians, school fairs including caretakers, and communication with clinicians demonstrated higher grades for science but not math or reading and fewer absences attributed to asthma as reported by parents but not fewer school-recorded absences. (Clark et al.)

Notice how the subjective measures (teachers' grades and parental reporting of grades) conflict with the objective measures (standardized test results and school-recorded absences).

Apparently, this isn't the sort of evidence that Leo is looking for. Leo isn't looking for any evidence:

Disingenuous calls for “evidence” that community schools work require a willful myopia on the effect on life in poverty on education — a blindness made possible by a complete unfamiliarity with the real world of the classroom.

If you ask Leo to provide support for his (expensive) opinions, you're being disingenuous. If you don't trust Leo that community schools work, you're being willfully myopic to poverty's effects on education. Of course, based on Leo's educational track record, if you're still foolish enough to be taking Leo at his word at this point, you'd have to be priapic.

I'll take disingenuous and myopic over priapic any day. I'm sufficiently hyperopic to know better than to take Leo at his word. Especially when that word calls for yet another bromide that gives more money and power to Leo.

August 8, 2008

Prediction Time

Following up on my last post on Charles Murray's new book, real education, it's time to see what Murray predicts will be the results from the grand experiment he proposes:

On measures involving interpersonal and intrapersonal ability. I expect statistically significant but substantively modest gains. On measures of actual knowledge, the experimental group will score dramatically higher than the members of the comparison group, perhaps 30-plus percentile points higher (technically more than a standard deviation). On measures of reading and math achievement, the differences will be no more than 15 to 20 percentile points (about half a standard deviation). Three years after the experiment ends, all of the differences will have shrunk. The differences in reading and math will be no more than 8 to 12 percentile points (no more than a third of a standard deviation) and may have disappeared altogether.

More formally, I predict that the magnitude of each academic effect will be a function of the g loading of the measure. Measures of retention of simple factual material have the lowest g loadings and will show the largest gains. For highly g-loaded measures such as reading comprehension and math, what has been accomplished by the last half-century of preschool and elementary school will be shown to be about as good as we can do, no matter how much money is spent.


This is a decent prediction. The I think that Murray overestimates the ease at which facts can be taught to and retained by low-IQ students and underestimates their ability with respect to math and reading comprehension.

Facts are difficult to learn because facts must be mostly learned on a case by case basis which is not readily amenable to acceleration. Math and reading (decoding and comprehension) are easier to teach because these skills, can be accelerated (even though teaching language and vocabulary remain problematic). But I knew that from the Follow through and the Baltimore Curriculum Project data. The data shows that we can get at least about three-quarters to a standard deviation improvement on average by the end of elementary school, better if we discount the schools that are so incompetent that they are unable to implement well-tested programs with fidelity.

Murray's point with respect to fade-out is well taken, but I'll leave that for another post.

August 7, 2008

Real Education: A Call for an Educational Experiment

I'm reading Charles Murray's latest book, Real Education: Four Simple Truths for Bringing America's Schools Back to Reality.

I agree with some of the points Murray makes and I disagree with others.

In any event Murray proposes a very good idea in Chapter 5:

Hence my second proposal, for a study that would be the most expensive educational demonstration project in history and would take as much as fifteen or twenty years from beginning to end. I state in the form of a challenge to everyone who is convinced that we can tach low-ability children far more than we are currently teaching them: Put up or shut up... Here is the proposal:

select children who test low in accademic ability but are not clinically retarded--say, children with measured IQs from 80 to 95, which demarcate the 10th to 37th percentiles. Make the number of the children in the study large enough that the results cannot be explained away as an accidents of small samples. Then provide these children with the best elementary education that anyone knows how to provide. Build new facilities or renovate existing ones. Hire the best teachers and create model curriculum. Measure how well the children are doing at the end of elementary school, and compare their progress with that of other children matched for IQ, family background, and whatever other variables are considered important.

...

The people who conduct the experiment should be free to use any teaching techniques, any class sizes, any amount of one-on-one tutoring, and type of technological aid. They shouldn't worry about making the program financially affordable for wider application, but instead bring to bear every resource that anyone can think of, at whatever cost that will maximize the education that these children acquire. Or to put it another way, their mission is to conduct the experiment in such a way, if it fails to produce success, there will be no excuses. Only three ground rules are nonnegotiable:


  • The organization that selects the experimental and control samples and tests the children must be completely independent of and isolated from the organization that conducts the experiment.
  • The design must protect against teaching to the test and test-practice effects.
  • The design must include a test for fadeout, conducted three years after the experimental education ends.


Great idea. Sound familiar?

That's what I thought too. So I dashed off an email to Murray informing him that we'd already done something very similar thirty years ago: Project Follow Through.

Murray wrote back that he thought something was out there (even though people kept telling him there wasn't) and hoped that Real Education would surface it. Sure enough it had and I gave him a crash course on PFT.

In the post I'll tell you what Murray predicted would be the results of this grand experiment and we'll see how well his predictions matched the results of PFT.

August 2, 2008

SES and Rotten Instruction

Deep in a comment thread over at Sherman Dorn's blog, Dick Schutz makes an excellent point:

The only thing that the [standardized] tests are sensitive to is SES and racial/ethnic characteristic. If those two variables were partialled out statistically, the results would show that schools are pretty feckless instructionally.

That's not "news." It's been around since the Coleman Report of the 1960's. But the popular conclusion is that we have to "change society." The "obvious conclusion" has been overlooked--change instruction. When you do, you find that the correlation with accomplishments and SES is near-zero. That's empirical reality, not a statistical manipulation.


At least at the elementary level and if we don't include comprehension with uncontrolled vocabulary, and if we teach the higher-SES kids like we do currently, but his larger point is valid. SES matters quite a bit as long as instruction is rotten.

August 1, 2008

More Visual Aids

I have two more charts related to my earlier post on school expenditures and student performance.

For the first chart I calculated the differential between total student pass rate and the total pass rate predicted from the regression between percentage of economically disadvantaged students and pass rates. (Basically this crudely controls for amount of economic disadvantage in a school district.) Then I divided the pass rate differential by the standard deviation to arrive at a z-score so you can more clearly interpret student achievement. Then I plotted this z-score against the differential between the total expenditures for a school and the median total expenditures ($10711.5). (Does that make sense? Let me know.)



For the next chart I did the same thing. this time, though, I used the pass rate for economically disadvantaged students only since this is a better indicator of how well these districts do with at-risk students. The results are similar.




It should be clear from these charts that at these funding rates student achievement is not affected by school expenditures. There are plenty of schools that perform well with low expenditures and plenty that fail even with high expenditures.

Bronze: Where the Least Motivated Find the Will to Succeed

A blog reader, Carol Glenn, has come up with a new proposed after-school program that incorporates DI and Core Knowledge. Heer idea is in contention over at ideablob where she can win $10,000 in seed money if you vote for her. Check out her business plan and cast your vote.

Here's a good visual representation of Carol's plan: