Everyone in the US hates standardized testing. There’s a very good reason for that. It’s not used correctly.
Here’s what we have now.
Student starts school. Teachers tell students what kinds of things are going to be on the test. Teacher pushes those things really hard (generally to the exclusion of everything else). Long about mid-January, students who the administration feels may be in danger of failing said test are put into a special class. Students are pre-tested. Students are sick of testing. The actual test takes about an hour for good students and all day for the students who sleep during the test. Several months later (mid-summer) scores are sent out. Student enters next grade. The end.
That’s it. In many states in the US standardized testing is wasted. That’s the way it was when I was teaching.
What is the purpose behind standardized testing?
These tests do not measure teacher or school effectiveness. They are not for determining intelligence or anything like that. They have one purpose. To see how well a student has learned the curriculum. Not if they have life skills or are ready for college, but how well they have learned the curriculum. That’s all.
The curriculum is mandated by the state board of education (depending on the state) and contains all the things that each student at each grade and for each subject area are expected to know. These are large documents, but they are also freely available on your state’s department of education website.
Usually the curriculum is developed with the help of professional educators, education researchers, teachers, and even parents and is generally approved by the governor or state legislature as well.
That’s why there is such a huge emphasis on teaching the things on the test. If one ‘teaches to the test’, then one is teaching the curriculum of that one should be teaching anyway. Quite a few savvy teachers have examined past tests and with a little pattern recognition skills can figure out the types of questions that are likely to be on the test.
This is because every standardized test has a ‘blueprint’. For example, when I was teaching (and I’m going by memory here, so it’s probably not 100% accurate), there were about 40 ‘standards’ for life science. The test was about 70 questions. Some areas were prioritized. For example, there were always several questions on DNA and few questions on evolution (in spite of the standards). So teachers would tend to emphasize DNA and de-emphasize evolution.
Why don’t we just ask the teachers to test the students?
That’s a very good question and I’ll give you the best answer there is. And it’s simply because every teacher is different. At one particular school I taught at, all of the life science teachers had roughly comparable end of course grades. However, on the state test, some teachers were averaging a 50% pass rate and some had less than a 10% pass rate.
But wait, I thought all the students were averaging the same. They were and that’s the problem. When a teacher brings an X-box to his classroom, there’s no way that any actual teaching is being done (and I’ve seen this multiple times). Think about it this way. Let’s say one of the standards for instruction is “the student is to know and understand the Krebs Cycle“.
Teacher A writes this question for the end of year test.
Draw a diagram of the Krebs Cycle. Label all the molecules and energy inputs and outputs. Describe the purpose of the Krebs Cycle.
There’s some stuff wrong with the question itself, but there’s no doubt that a student who can answer this question correctly “knows and understands the Krebs Cycle”.
Compare that to teacher B.
What is the product of the Krebs Cycle?
And before you ask, yes, I’ve seen teacher written tests that were even worse than this. But the question is, does the student really have to shown knowledge of the Krebs Cycle to answer this question? Not a bit. We can eliminate option 4 instantly and then we’re down to a 33% chance that the student gets the right answer.
Now, let’s say these two teachers are in the same school… and this test is what is required for graduation. Suddenly knowledge isn’t the main issue here… it’s who has the second teacher. A really smart, well prepared student may not be able to do everything teacher 1 asks (I couldn’t right off the top of my head) and even a big chunk of students who slept through class would be able to get the answer for teacher 2’s test.
This is fundamentally unfair to everyone (teachers, students, parents, colleges, etc). Even if tests are written by the school districts, we would run into the same problem. So we have to have a “standard test” that every student takes. This is the only way that’s fair for everyone.
Beyond that, how can these tests be used?
A well crafted test can do much more than just generate a score and whether the student has apparently mastered some material. A well crafted test can identify the specific areas that need improvement (indeed, teacher level data could potentially show where teachers need some extra training too). A well crafted test item can actually show WHY the student got the question wrong. Not just “it’s wrong”, but “it’s wrong because the student’s answer indicates that they don’t have a good grasp of the final stages of the Krebs cycle.”
But even if the test is well crafted, very few schools look at that data. At least the ones I’ve taught at haven’t.
Ideally, the results page of the test should include a suggested remedial program for the student. The people who make these tests have the ability to do this. It’s never done because it’s too expensive.
A final note. A lot of professions have standardized tests. Nurses and teachers just to name two that I’m most familiar with. No one much complains about them. But these tests do what I think that all standardized tests should.
- Provide evidence that students possess certain knowledge, skills, or abilities.
- Provide information as to the students areas of strength and weakness.
- Allow comparison of the student to a specific level of skill or knowledge.
And that’s what this is really all about.
Testing and education is a science. There is a significant amount of research being done in these areas (I know, I’m doing some of it). We can make tests that are really good and provide a lot of really useful information. But if that information is ignored or a careful, deep analysis is too expensive, then we can’t do much with it.
I’d like to talk about some specific aspects of test design, but that’ll have to wait for later.
If anyone has questions, please ask in the comments. There’s a lot to this.