Wednesday, April 23, 2014

PARCC Discovers Unicorn Farm!

Over at the Thomas B. Fordham Institute, PARCC CEO Laura McGiffert Slover has breathlessly announced the discovery of a special farm where unicorns dance and sing and give rides to little winged cherubs while singing sonnets in Latin.

Okay, not really, but I would have found that more credible than Slover's actual announcement, which is that PARCC heralds the end of test prep.

PARCC states are creating tests worth taking, made up of texts worth reading and problems worth solving. They are designed to give teachers information and tools they can use to customize teaching and learning for each student, and give students test questions and tasks that are meaningful –the kind that great teachers routinely ask students.

This kind of ad copy creates an existential dilemma for me. Should I accept that Slover actually believes this stuff, and thereby conclude that the head of one of the largest testing companies in the nation is so woefully ignorant of how standardized tests actually work? Or should I conclude that Slover is so astonishingly cynical that she can shovel out baldfaced bovine fecal matter and expect the public to eat it up like caviar?

She of course tosses in "as a former teacher," but looking at her bio, it's hard to tell when. Slover was a board member in DC during the Rhee era, a honcho at Achieve, and sat on the committee that wrote the CCSS math standards. Slover is a fine example of a person who has figured out how to make the revolving doors of corporate government work for her. She wrote the rules with Achieve and the CCSS; now she's cashing in with PARCC.

None of that really matters as long she's willing to write such dumb things. Let's look.

The PARCC assessments mark the end of “test prep.” Good instruction will be the only way to truly prepare students for the assessments. Memorization, drill and test-taking strategies will no longer siphon time from instruction.

My existential dilemma is compounded here, because I find it hard to believe that anybody could be either that stupid or that big a liar. But let's pretend for a moment that somewhere, somehow, there is a person who could actually seriously consider the above statement.

All standardized tests are susceptible to, and therefor encouraging of, test prep. All of them. Always. Forever.

The whole guiding principle of a standardized test-- particularly one that is designed to be administered by computer AND to generate crunchable large-scale data-- is to force students to choose from a certain portion of the broad world of possible responses to any problem. As Slover writes, "Results will also finally make student performance comparable across states..." The person who is writing the question-- the live breathing subjective human being-- will embed certain values in that question and the only acceptable answer to it.

Let's pretend that the problem is to find a way to get from my house here in Western PA to downtown Cleveland. Let's go to the Arcade, because I like the Arcade.

Every parameter I put on a correct answer represents a value judgment. Did I say you have to get there within three hours? That means you have to travel by car, and you can't go by way of Pittsburgh. Should you travel through Amish country, or use an Interstate? Did you end up at the right place-- because if you don't know Cleveland, you probably don't know what I mean by the Arcade. Of course, if you know Cleveland well, you know there are several arcades and I might be asking a trick question.

It doesn't matter how sophisticated or simple the problem is. The problem has nothing to do with test prep. It's the solution. As long as your test model involves saying, "Out of all the possible solutions, we are only considering these four, and out of these four only one is correct," then your test is preppable. I can study what kinds of answers the testmakers like. I can study what kind of false answers they favor.

Standardized tests always reflect the values of the test makers. Learning to reflect and mimic those values is what test prep is all about.

Standardized tests are also eternally preppable because they are such hugely artificial tasks. For my "drive to Cleveland" problem, the most authentic task would be to actually drive to Cleveland. But we can't do that, so we are going to create an artificial task that, the test-makers believe/hope, will measure the same skills and knowledge.

Except that it won't. It never does. I will end up with a test question that involves choosing between four maps, or identifying landmarks that I would see on the way, or doing some other in-authentic act that is not exactly like driving to Cleveland at all. And all of that requires tests prep; my best chauffeur may be a terrible map reader.

We already know that the New Test Regime likes items that purport to rank high on the Webb Depth of Knowledge scale, which means we'll see lots of "Here are two things. Make a connection between them," or as the test preppers will say, "Spot the connection that they want you to find between these two selections."

This is what all standardized testing comes down to. "They" want you to come up with a particular answer. The Tests measure one skill-- can you figure out what "they" want you to say, and then say it? Because the PARCC doesn't change that fact of testing life in the slightest, it can be test-prepped. Because the PARCC is tied to such high stakes, teachers will do test prep for it.

Let me frame it another way. Any test you can train an uneducated minimum wage temp to score is a test that can be gamed, and any test that can be gamed is a test that students can be coached to take. So, test prep.

When I want to assess my students' understanding beyond basic recall and simple skills, I use some sort of open-ended assessment. If I want them, say, to compare and contrast two works of literature, I typically assign a paper, and there is no answer key. There is no prescribed format or organization. It is not the student's job to present the answer I want in the format that I want it. It's my job to take the student's essay on terms set by the student, and then to assess if she managed to successfully make her point to me. But key to this process is that it be student centered-- not teacher centered and certainly not test centered. If I want to know if the student really understands, really sees, really has something to say and can say it, there is no other way.

I can't bring myself to slog through Slover's baloney about how teachers are providing valuable feedback and the tests are being created by a coalition of thousands of educationny folks. Also, blah blah blah the test results for questions you will never see for the students you had last year will really help you with your instruction. How does her keyboard not simply melt in shame? But Slover's Big Finish underlines the fundamental problem at the heart of the PARCC test.

So, yes, the numbers matter, but what really matters is that these students—and a million others participating in the field tests, their teachers and administrators, and the PARCC states that are developing and field testing the new assessments—are making history.

Got that? What's really important here is that the students and teachers across the country are doing something great for PARCC.

Tests like the PARCC do not serve students or teachers. Students and teachers serve the PARCC. The only kind of assessment that can be prepared only with good instruction are those assessments that are student-centered. Not only can you do test prep for any standardized test, but that will always be the best way to get better results.

No comments:

Post a Comment