How Can Schools Fix the ‘Mania’ of Achievement Testing?

A conversation with James Heckman, Tim Kautz, and John Eric Humphries about how testing took over American education, and how to bring life skills back without losing accountability.

James J. Heckman   Photo: University of Chicago

Recently I’ve written a fair amount about early-childhood education and vocational education, both of which, so to speak, are having a moment. The former is being pushed, in various forms, by Barack Obama, Hillary Clinton, Pat Quinn, and Bill de Blasio, among others; the latter is at the heart of Chicago Public Schools’ new emphasis on STEM education.

Chicago’s a good place to follow these new initiatives, which in some ways run counter to the emphasis on standardized achievement and cognitive testing that’s intensified in the wake of No Child Left Behind. There’s the Thirty Million Words program, based off the work of the University of Chicago’s Dana Suskind; the Ounce of Prevention Fund, led by Diana Rauner, wife of potential future governor Bruce Rauner; and the Heckman Equation, the project of Nobel-winning University of Chicago economist James Heckman, whose research is a cornerstone of both movements—and who makes our Power 100 list this year.

Now Heckman has a new book out, edited with two U. of C. Ph.D candidates in Economics, John Eric Humphries and Tim Kautz: The Myth of Achievement Tests: The GED and the Role of Character in American Life. A collection of essays by the editors and a host of researchers, it essentially tracks Heckman’s path from investigating the use and abuse of the GED into the greater role of character (or non-cognitive) education and evaulation, and its significance as a predictor of life outcomes.

Heckman has been studying the GED for years, and his research tracks with Chris Rock’s observation that GED stands for “good enough diploma.” (Rock would have made a good economist, I suspect.) How so? According to Heckman’s findings, GED recipients are as smart, on a cognitive level, as high-school graduates, or at least ones who don’t go on to college—and smarter than high school dropouts.

But in terms of life outcomes, they end up a lot more like college dropouts. Worse yet, the GED encourages students to drop out, by offering a quicker, easier equivalent of a high-school diploma. Because it’s not, however, an actual equivalent in the real world, it lures students with the raw brainpower to graduate high school out of high school, an effect that’s strongest on minorities. Heckman’s found that, if you count GEDs as high-school dropouts instead of graduates, the progress that America has made in narrowing the black-white graduation-rate gap disappears.

This led Heckman to investigate why the GED doesn’t provide the returns it promises. In part, it’s because the initial intent of the GED has been distorted. As John Eric Humphries describes, it was developed to give World War II veterans an alternate means to demonstrate to colleges that they had the academic preparation for higher education. A high-school diploma, on the other hand, signals not only that a student has gained a certain body of knowledge, but had the persistence to gut out all four years.

What the GED wasn’t designed to measure is life skills—like character, determination, and motivation, skills those veterans were assumed to learn in the armed forces. For Heckman, Humphries, and Kautz, those skills are critical, and the GED ignores them. In fact, it’s a signal that a recipient lacks those skills, and much of the book is devoted to proving this hypothesis.

From there, it expands into the rise of achievement testing, which drove the acceptance of the GED and its application to younger and younger students over time, and overturned the foundation of American education in character development. “It just corrupts what schooling’s all about,” Heckman says.

I spoke to Heckman, Humphries, and Kautz (mildly edited for length and clarity) about their new book, their ideas for integrating character-based education into public education, and how character education can be integrated into the testing-and-data oriented public-policy world we live in.

How did you begin to study the GEDs?

James Heckman: I had done a lot of work on evaluating job training programs in the United States, and I found consistently negative evidence, or just very discouraging evidence, on remediating disadvantaged people when they were an adolescent or young adult, much less later. I stress “disadvantaged.”

I was struck by [a GED program in Corpus Christi, Texas] that I happened to visit that was part of my study on the job training program. I thought it would be useful to see what was happening on the ground, so I went to Corpus Christi. I visited the site for several days and saw a lot of things. But it was the GED program that struck me the most. I’d frankly never heard about it before, or I had and didn’t make much note of it. The locals were extremely excited by it.

It was something that was transformative from their standpoint. They were taking people who were mostly Hispanic, longstanding American citizens who were basically at six years of school, and they had elevated them to high-school graduation status within about, oh, maybe six weeks to two months of an intensive GED course. I said, well, this is something to study.

Around that time was also the time when [Richard] Herrnstein and [Charles] Murray were writing their book The Bell Curve. I got involved in that book because Murray came frequently to Chicago at that time and interacted with me, and I read a lot of chapters of that book. I was intrigued by the argument [that intelligence is measurable, heritable, and highly predictive of life outcomes].

I started thinking a little bit like Arthur Jensen… there was nothing you could do, there were those low ability people, and they just were duds. But then I began to realize that Murray… had a great piece of work in the sense that showing abilities are partially predictive of life success, like an achievement test score. Very predictive tests of certain life outcomes—crime, earnings, employment.

I was also looking at issues of race gaps, ethnic gaps and earnings. I thought, “well, these abilities are playing a really important role.” And one thing led to another, a slow evolution.

I realized that the GED was not really that effective. It took me awhile to understand what was going on. The GED [recipients] really were smart, as smart as high school graduates who didn’t go on to go to college. But they were performing at the level of dropouts.

About 10-12 years ago, I thought, well, there should be some measure of these traits, which we call non-cognitive traits, which were really traits that weren’t captured by the test score. And we realized how important they were. And the rest just led to the book, I think.

And the idea that these abilities were genetically determined, which is what I thought and I think a lot of people thought 20 years ago. I began to question it as I looked at the evidence on intervention. It was a series of things. It started with Herrnstein and Murray and the trip to the GED training center in Corpus Christi.

The GED has shown an ability to signal a certain type of student—as smart as a high-school graduate, but without the non-cognitive skills. Is it possible to take that finding and redirect it into job training?

JH: We do. The point is that there are multiple skills that we emphasize in the book. And we also emphasize that those skills can be acquired. We examine a series of programs and interventions—there are various ways that parents and other players in society intervene to shape these abilities. So we talk a lot about the job-training programs, we talk about second-chance programs, and what might actually be a more effective way to go forward.

Tim Kautz: That’s the point: we found a way to identify students who are relatively smart but lack other skills.

The other way to look at it is that a high school degree is actually a really good measure of non-cognitive skills. For people who don’t have a high-school degree that are relatively smart are good candidates for a number of other programs that could potentially build non-cognitive skills. And one promising thing is that it turns out that non-cognitive skills are more valuable than are cognitive skills. Which suggests that programs targeted at adolescents that attempt to build non-cognitive skills could be effective.

In our review of the literature we did find several programs that seemed to be effective in doing that. And they tend to be the ones that combine some form of education or non-cognitive skill training with the context to apply it—specifically programs that basically bring kids, adolescents, or dropouts into the workplace, and that’s a place where they can learn both cognitive and non-cognitive.

John Eric Humphries: The GED does a pretty good job of measuring the minimum competencies in reading and writing and mathematics. If you’re asking it can be potentially paired with a larger program to make sure those minimum competencies are met, I think that’s a real possibility.

But it’s also what the original GED was for in many ways. As we had World War II veterans returning, it was really that we believed these veterans had matured and learned a lot about life through the wartime rigor, and had developed discipline and types of non-cognitive skills we think might be developed by being in the Army. But we needed to make sure that before we issued wartime diplomas, that these veterans could at least read and do math.

The GED was much easier then, but that’s almost exactly what the GED was originally for. It was saying that, we think these young men had developed into adults, but we want to make sure that they meet minimum competencies.

Why is there a skepticism about vocational training in American education?

JH: I think the recent history is tracking. The reason why it had such an ill repute recently is that the vocational track is typically one that was used to put in students that weren’t doing very well, and to be honest I don’t think it was that well supported.

And the training facilities were obsolete. I remember that one day I was visiting one training center in the 1990s that was teaching people how to fix Volkswagen engines from the 1960s, which were no longer sold. So you were training people on a skill that had zero value. The reason is that they hadn’t received any new equipment in 20 years. It was a remedial activity, more or less a holding tank. There was no serious effort.

I think that if you look at programs that essentially integrate school and work, that can be very effective. People frequently talk about the Austrian, or German, or Swiss apprenticeship programs, we talk a little bit about those in the book.

But if you go back to John Dewey, and you think about the curriculum that he devised for [Francis] Parker and for the Lab school, it was involving not just academic education, but involving the structure of work, gardening—not just getting your hands dirty, but also the broader notion of what education was. I think that was something that got lost, and I think it got lost because society placed increasing emphasis on cognitive skills.

That was viewed as the great revolution 50 years ago. IQ tests were really important, and we really needed to promote cognition… The fact of the matter is, that there is this great faith [in cognitive education].

[Vocational education] fell out of favor simply because what was being provided was a mediocre form of integrating school and work. But I do think there’s some evidence that shows the benefit of integrating school and work.

It does seem like the pendulum is swinging back a little bit, for example here in Chicago, with the new emphasis on STEM education and computer science.

TK: I would add a couple other Chicago programs, like the Career Academies—those are another way that high schools can basically integrate school and work, and expose students to a workplace environment early on. And those evaluations that we’ve seen have been very successful, along with the Year Up program.

Basically they provide people with training and, simultaneously, an internship where they can immediately apply the skills that they’re learning, and get feedback on how they’re doing at work. And a lot of it seems to be through non-cognitive skills. So I think there is some swinging back.

Are there ways to teach character in a more traditional cognitive-skills based program?

JH: Yes. If you think about what the curriculum was. Don’t think about it as the curriculum per se. Think about what’s going on in school.

This is something that every major educator we ever found would say: you’re molding character. We have this great quote from Horace Mann, that what school is really doing is molding character, much more than just teaching reading, writing, and arithmetic.

These were lessons; you could read the McGuffey Readers. McGuffey Readers were filled with morality fables, and stories, and there was a sense in which children were being molded, literally the shaping [of character]. We have a character—there was a person here at Chicago, named [Franklin] Bobbitt, who we’ve all fallen in love with. He thought of the school as a factory, and it was shaping, like a factory producing steel rails, this was producing character and producing knowledge.

The goal, I think, of American education, for decades, and across many, many scholars, was basically to teach people broad lessons in how to live life, how to engage life, how to essentially be effective citizens and effective people.

In fact, I think what led to its downfall was that it was organized around a Protestant Christian version of what a good person was. The virtues of manliness, the virtues of citizenship, basically the kind of things that Charles Murray talks about in his book Coming Apart. He talks about the fact that we no longer talk about these traits. It’s considered quaint and old-fashioned.

He has a view, which I share, actually, that we’re getting an emerging underclass, which rejects all these values. There was a sense—cultural relativism was such that we shouldn’t impress these values. Many people say, oh, that’s very Victorian. Some people would take that as a compliment.

I think structure of the argument was moved away from teaching these lessons in school. Deportment was a grade that people used to get: how well they performed. We document in the book how school discipline declined, how a number of factors led to changing the curriculum.

And then the inexorable drive [to achievement testing]—and I think this is where maybe economics went off the rails, to be honest. Education became so focused on test scores, and in particular test scores associated with achievement. Suddenly, when we started thinking about how to evaluate schools, how to measure schools, how to be efficient, how to measure the quality of schools, it became increasingly focused on exactly this notion. It just drove out something that I think is extremely important, these character skills. Just because it was believed you couldn’t measure them.

There were a number of factors—partly this notion of a cognitive psychology, that the IQ test and the achievement test would summarize everything that schools were supposed to teach. There was always the refrain that it was too difficult to measure these other traits. Those things, the drive to measurement, really caused the refocus of the curriculum, and the curriculum started teaching to the test. Literally now we see it to this day.

JEH: I agree strongly with your point at the end. You see the emergence in the history of the education curriculum and how education developed, you see the emergence of the standardized test, and the idea that you could move to meritocracy, even though kind of every one of the developers of the tests says they know it only measures half the pie.

You see that, in the curriculum, as soon as you start measuring, quantifying half the pie, it really becomes the whole thing that matters. It’s a what gets measured, what gets taught outcome. You really see that in the last 40 years.

TK: I’d add a couple quick things. Now schools are starting to look for ways to measure non-cognitive skills. They’re hesitant to adopt measures that psychologists have developed based on self-reporting, basically because they don’t think it will be comparable across schools, or they’re worried that people will be faking answers.

A whole group of school districts in California that represents over a million kids just recently applied for a waiver to the No Child Left Behind Act, partly in order to incorporate measures of non-cognitive skills into their evaluations.

The interesting thing is that if you back to what Ralph Tyler was saying—you could actually use the types of data that schools are already collecting, like their activities and extracurriculars as measures of non-cognitive skills—we’re finding that’s actually a very feasible approach.

It’s exactly the intuition that Ralph Tyler was suggesting, which is that attendance represents a non-cognitive skill. Whether students show up or not says something about persistence. And it’s more predictive, actually, than the test scores. It’s completely possible for schools to start doing this now with very tangible measures.

Kautz adds, via e-mail, that in his study of national databases, “9th grade attendance is just as predictive of high school graduation as the achievement test scores.”

JH: The driving motive behind the book, and I think the important lesson that’s still not fully appreciated in American society—you see well-meaning politicians, Secretaries of Education, cabinet ministers in England, everywhere you go, always thinking we really want to improve the schools, and they’re always measuring the schools by these test scores. And there’s a belief that the test score somehow is very predictive of many life outcomes.

When No Child Left Behind was introduced, I think that was the high point of the madness. That was the point where you had, like in the French Revolution, Thermidor, where the Jacobins were beheading people. The PISA scores, or the No Child Left Behind test came around, and then the whole notion of education became corrupted to focus on this very narrow aspect of what everybody said was a piece of the story.

To be honest with you, I think one of the worst influences on education is schools of education. They get into these trends, these fads, there’s this fake science that comes out of those schools. Over the years they’ve consistently mislead the American public. In this case, with the emphasis on the test score, and the quote, scientific evaluation of schools through test scores and so forth, it’s completely corrupted the process of American education.

And many people have said this. But I don’t think they’ve produced a lot of evidence that these other factors matter, and they matter a lot. And you can do something about them. The whole mythology about what scientific education evaluation is about, what school should be about, I think is really just wrong.

And there are some good people in ed schools; I don’t want to attack all ed schools. But there’s this mentality that it has to be an experiment to be really good. We analyze a lot of experiments. But a lot of experiments have trouble; there’s no question that experiments have real flaws and produce very flawed knowledge at times.

Secondly, the main point of the book is that, the idea that we should boil the measure of a student down to a few test scores is insane. And corrupting. And it leads to corruption across the whole society.

By “corruption,” I mean it corrupts the statistics, it just corrupts what schooling’s all about. It misdirects it. Many educators have said this over the years, but they never had the evidence.

There’s always this view—I routinely ask people, what fraction of the variance, of the variability, in outcomes is explained by IQ. And most people say 50 percent, or 25 percent, or 30 percent. What do you think it is?

I’d say 20 percent.

Right, that’s what people say. It’s closer to four percent. Three or four percent. And these non-cognitive traits will play a comparable or even greater role; it depends on the path you’re talking about.

If I were to argue with what the average educated, well-meaning person would say, coming out of the ed school logic, which has permeated the press—I would bet President Obama would guess more or less you guessed—there’s this belief that cognitive traits are really paramount, that without them we fail, and these other traits either can’t be measured, or they’re not very important, both.

There’s another whole line that says those traits are not even stable traits. You are what you need to be in a given situation. There’s a structure out there where there’s a tremendous amount of, I think, misinformation about exactly what are the important traits for predictive success in a number of different aspects of life. I think that’s the part that’s really not understood.

It’s like the old joke about the drunk searching for his keys—he lost them somewhere else, but he can search closer under the streetlight, so he goes to look for his keys under the streetlight. That’s kind of what we’re doing with these tests. They’re very far away, but they’re clear—we understand them, we tabulate them, we compute percentiles. We can compare them over time. There are a whole group of people, including many economists, who think that this is the be-all and end-all of educational evaluation.

And yet, the fact is, they’re only a part of the story. And in some cases not even the most important part, not by a long shot.

What we’re trying to do is re-educate people from the education-school mentality of blind belief in test scores. And also the sense, as we have several chapters showing that if we take the GEDs as the high school equivalent, we delude ourselves in thinking that American society is doing better than it is. We ignore the fact that the introduction of the GED actually will create problems by inducing students to drop out. It’s the easy way out, so they take it.

It’s a belief that runs throughout American society that these tests are really good things. They’re just piddly predictors of what’s really important out there. They play a role, I’m not saying they should be abolished or abandoned completely, but they have to be supplemented. Then we’re back to Tyler, we’re back to Horace Mann, actually, and James Coleman, and just about everybody else that ever looked at education seriously, outside of the mania of the test.

How does the evaluation of character education—and schools getting away from their reliance on testing—fit into the increasing use of data to evaluate public policy?

JH: We’re not suggesting that schools shouldn’t evaluate. We’re suggesting broader notions of evaluation. They’re feasible. Like Tim was saying—with his data, he’ll actually have a lot of information that will allow him to look at these behaviors, teacher ratings of these other traits, and even grades themselves are communicating a lot of information on the character and motivation of students.

We’re also saying you can change the architecture of the school to promote these character skills. The first step is recognizing they’re important.

There’s a sense about having an objective social science. That was the dream with achievement tests, when the first screening tests were put out there—to have an objective summary of the factors that make people succeed, how to sort them in schools, how to assess the educational progress. It was all an honorable motive.

But it just happened that it became very specialized and very focused. It just missed some dimensions that are important.

In fact, with big data, and all the new evaluative schemes, that’s even better, feeding into what we’re suggesting. You can use administrative records that school districts have on hand to assess how character is evolving, and how effective schools are in shaping it.

There are inventories—preschools measure this all the time. If you go to a preschool program, they’ll talk about how kids are behaving with respect to each other, they’ll observe these interactions. There’s a very strong interaction pattern that’s there, and you can see where the children are conscientious, staying on task, and on and on. They will construct these inventories that go with this. At some levels of education this is done routinely. There’s no reason why it can’t be done in high schools, in junior high schools, and elementary schools.

JEH: What Tim’s doing with this should be quickly becoming possible, almost statewide, in any state. Part of the No Child Left Behind Act was that all states had to set up these longitudinal student tracking systems, and most states are in compliance with that now.

Looking at who’s dropping out, and how that’s related to their first quarter or first semester attendance records, for example, should be very easy for school districts to start doing, because of some of the infrastructure we set up under No Child Left Behind.

 

Share

Advertisement

comments
5 months ago
Posted by BSluther

As a teacher, I completely agree that schools focus too much on tests. Great article.

Submit your comment