Dana Goldstein has a rather. . . inexplicably passive . . . piece on "A Decade of No Child Left Behind" at The Nation Online. I recognize that not all of the Nation online folks are prime time Nation writers, or that the material online is the groomed and researched matter that goes into our beloved, weekly tabloid of rile and revolution, but Ms. Goldstein has, so far, been more journalistic than advocate, more "there's this, but then there's that," than I could ever have expected under the masthead.
I am no one's judge, though. There are other delights on the plate, and I only mention her piece as a point of departure, not as something to pick on. Oh, it's true that she forgives its abuses by giving this glossy "but on the other hand," saying,
it is no longer possible for the media or political elite to ignore the inequities in our education system. Unfortunately, there has been very little acknowledgement of the fact that gaps in academic outcomes have multiple causes
but ending with a prettification, whereby conservatives have as much to say as progressives. (Good old "balance" in action.)
Since I am not a journalist and do not believe in discussing symptomatic conversations as if they were analyses, I want to talk about the deep questions behind NCLB and the system it aped and replaced: the generalized IQ and college entrance standardized test. I would like to play skeptic and strip bare the questions we need to answer before we go forward, if we act solely on reason, and then speculate on why we plunged forward anyway.
Here's a menu of what's to come, so you can decide whether or not you want to click over the speed bump:
I. Standardized testing of any sort, if it has determining value (i.e. if you say, "It means the student/school/teacher is good") has to replace another value.
II. Tests therefore must satisfy certain criteria before they have an advantage over the status quo
A. They must be comprehensive inventories of the things we consider necessary.
B. They must be reliable instruments that will return the same value with the same student at different times of day or location.
C. They must be accurate assessments, in that they must reflect, with as little noise as possible, the contents of the inventory of skills or knowledge we deem desirable.
D. They must be unique and solely competent measures, in that another form of assessment that is less disruptive must not be available.
E. They must be accurate across time, so that they reflect actual attainment, rather than strict performance (long term memory as opposed to "cramming"; concepts rather than memorization).
III. Inasmuch as these questions are "no" in many cases, the motivation for moving toward the test must be sought some place besides assessment.
If you part company now, that's ok. If nothing else, take that little list of questions with you and apply it to your next testing opportunity. Do not do so to deny the test, but to ask why, since no test can pass the challenge, we do it anyway.
First, NCLB must be the most disliked ten year old since Dennis the Menace.
That doesn't tell us anything, by itself. Impressionistically, the program is a disaster. In my own time teaching, I have yet to meet a parent, teacher, or administrator who has liked it. Past 2001 itself, I never met any supporters of NCLB. A Pew study in 2007 showed a variegated raspberry being blown by the public on the subject. The same study showed that Republicans in that year of defensiveness supported it more than anyone else, but college graduates, who themselves had been through much testing, had the greatest doubts about tests.
The fact that the NEA doesn't like NCLB, parents don't like it, students who went through it don't like it, and even partisans don't really like it does not mean that there is anything wrong with NCLB by itself. I worked at a place where everyone agreed that the current president of the institution was bad. However, each one thought so for a different reason, and each wanted to replace him with an ideal president of their own passion.
One criticism comes from the idea of any federal "intervention" in education. After all, that is supposed to be a sacred local issue. The problem with that is, well, Kansas. The Kansas Board of Education could decide that evolution is less scientific than a literal reading of one chapter of Genesis, and Texas could decide that the "Reagan Revolution" should bump the origin of the labor movement in the laissez-faire of the 1920's. The other, more relevant, criticism is that economic disparities between the wealthy and the poor simply repeat endlessly.
Without federal involvement in education, states like Mississippi have a dark horizon. With low achievement and low tax receipts, they cannot fund schools well. To attract business, they give credits (tax abatements) to businesses, and the next generation of students are poorly served again due to low tax receipts, etc. Jonathan Kozol's Savage Inequalities has demonstrated the fundamental problem with property tax funding. The "nice" places get nicer, and the "nasty" places are confirmed in their poverty.
Of course, when you hear from today's GOP candidates that they demand that the Department of Education be abolished, with their "DOE est delinda!", their motivations have nothing to do with taxes. They're looking for the ability to reinstate racial and religious priorities in education.
So, without getting to how NCLB got going in Rod Paige's fraudulent "Dallas Miracle," it depends on the test, the test that should pass the five bars I outlined above. Let's look at the questions, and why they must be answered:
1. Is the test a comprehensive inventory of the skills and knowledge we deem necessary?
A test that will fail a student, put a student into a "track," or deny a student a further opportunity (negative typing) or which will certify a student as adequate, excellent, or superior (affirmative typing) should not be random material. The things we ask for, especially when we say "failing school/student/teacher," must be the items we have determined are the necessities for the curricular division. This requires that we have an agreement ahead of time on what the proper subject of a curricular division is, and then what its minimum is, and be certain that all of what we regard as necessary for adult life is present.
Note that this, if we are going to say that a candidate is superior or a failure, must contain both inventory and skills. If we decide that all adults must understand basic algebra, then we have a series of propositions that will have to be known and understood (inventory). We also, though, will be asking for a skill and will need to be sure that the candidate can solve for X in a simple polynomial. If we decide that chemistry is vital, then we will have a large inventory of concepts relating to molecular structures and bonds, but we will also have formulae and possibly laboratory science skills.
2. Does the test return the same value with the same student at different times of day and at different location?
This is the "accurate test." We want to know that kids are going to be reliable and the test a reliable method of cracking those children.
This is the most fundamental error of standardized tests, or the most glaring fracture of the regimen. Humans are not reliable. Human knowledge improves and degrades with time. Teaching, similarly, does not yield its results in a predictable manner. The same words by the teacher will yield six or seven separate messages in a classroom. Some students will get "aha" moments later, others will put the class into the discard bin.
Because it is impossible to ask the same person the same question twice and get the same degree of difficulty, testing tests for this sort of reliability is nigh unto impossible.
3. Does the test reflect, with little or no noise from cultural contamination or linguistic interference, the inventory and skills we want to evaluate?
Cultural interference is real. Some tests are ludicrous. However, whether they have overt cultural interference that adds noise to the testing, they have linguistic noise.
Subordinate clauses vs. straight independent clauses, font sizes, vocabulary, extended prepositional phrases, multiple objects . . . all of these, as well as what we would call bad writing (ambiguous pronouns, dreadfully vague objects, implied subjects), are a presence in the test itself. Each part of language is noise. We can standardize it, and there is nothing wrong with seeking standard American English, but the test has to be scrutinized carefully. Every single piece of noise is a deficiency in assessing the student's attainment.
4. Is there another way to measure students that causes less disruption to the curriculum, inasmuch as the classroom was not developed around tests?
Ok, this one's a mug's question. However, there is something serious behind it.
For us to alter the classroom as it exists, there needs to be an advantage to the change. If there is another method that works inside the classroom method, then that would be superior to the standardized test day to the very degree that it fits into the curriculum. (There are hundreds of forms of assessment out there. Scan-Tron isn't the only one.)
5. Do the tests measure attainment across time and show knowledge as opposed to a performance?
This is hard, but what we need for a test that calls a school "failing" or a teacher failing or a school excellent is a test that shows what a student has moved into full memory. An actor may recite two hundred lines of Shakespeare, but that does not mean that he understands Elizabethan England, or even the play he's in. The performance of memory is not the capacity of memory. To find out,there is no way to assess except through demonstration of skills through application. That, however, would require what they call a "subjective" assessment.
Why do we do it, then?
I will return to the "performance gap" for a moment. Dana Goldstein speaks of how good it is that the media must speak of it. Per-for-mance gap...
That's not, perhaps, an education gap? That's not, perhaps, a funding gap? That's not a destructive society surrounding poverty? It's not demonized working poor parents?
You see, the performance gap is a set of numbers on a standardized test. It has no reality by itself. It is only a line derived from a test, whether the test is the SAT or one of NCLB's monsters. NCLB mandates that every school must improve its percentages every year, for example. Now, if your school is actually good, you're in trouble, because you'll be a "failing school" if you don't make your numbers go up. I worked briefly at Baltimore City College, and it's an extraordinary high school. They were good, had been good, and served the population that wasn't supposed to be good. NCLB told us that we had to get better, when we were best in the city.
Rod Paige's NCLB came from Rod Paige's "miracle" of moving numbers. He got percentages to improve, not students.
With national administration comes national disinterest.
When Washington funds, at least since the "Reagan Revolution," it means "accountability" for funds (unless guns are involved, or fusion x-ray lasers in space). The red-tape worms of Washington like numbers, and politicians who do not teach and cannot be taught understand percentages. Rod Paige got a great improvement in the percentage of graduation (by expelling anyone who would drop out), and his scores went way, way up (by reclassifying any student with low scores as Special Education), and that made him a great educator.
There is already a tendency in Education Schools to love a good number more than a good date, but NCLB was born out of stock-trader like number games. It was also, though, born out of a mistrust, if not hatred, of teachers. Teachers, by the right wing fantasy, are turning children into gay communists. Teachers, by the left wing fantasy, are turning kids into creationists with guns. Either way, from D.C., there is this idea that the grunt on the ground cannot be trusted, and a good test will fix things.
So, ten years of NCLB? It's time to look at is our students learning. They is thumbing on there cels 4reel. As for anything else, we should come up with a good test.