20 June 2013

Don't Game the Test -- each year's AP rubrics are different!

I've always been annoyed by Big Test Prep, and not just because of the competition for my AP prep book.  The entire industry promotes the attitude that standardized tests -- and by extension, all tests -- are games to be won via trickery, rather than authentic evaluations of knowledge and skill.  It's all BOOST YOUR SAT BY 100 POINTS USING THIS ONE WEIRD TRICK THAT THE COLLEGE BOARD DOESN'T WANT YOU TO KNOW!!

Folks, the AP physics exams are excellent exams, well-constructed by experts in physics, experts in physics teaching, and experts in test design.  The free response portions are graded to rubrics by savvy, intelligent physicists.  There's only one approach that will lead to success on the exam: know physics, and know how to communicate that physics knowledge.  Nevertheless, as an AP reader and consultant, I see teachers and students trying to game the exam.  

The biggest misconception about the AP exam is that there are hard and fast year-to-year rules behind assignments of points on rubrics.  

Folks, the rubric for any particular exam question originates with the original author of the item draft.  The development committee revises the question and its rubric.  After the exam is administered, table leaders are assigned to each question.  These table leaders show up a few days early to the Reading so that they can page through hundreds of sample exams.  They adjust this rubric again so that it can be applied consistently and usefully for the range of actual observed responses to the question.  

And there's more!  The table leaders have to present and defend their rubric to the entire group of exam leadership.  Suggestions and advice flows from problem to problem.  Now with the perspective gained from seeing all the rubrics from all the problems, each set of table leaders revises and finalizes their rubric.

The upshot of all this rubric design is that a "trick" that would have worked on a question on this year's exam is unlikely to work on a future exam.  

For example:  Take a look at the rubric for problem 3 on the 2011 AP Physics B exam.  The first part of this experimental question asks the student to check from a list the items he plans to use in the experiment.  The rubric awards two points for checking boxes: one point for checking a ruler or meterstick, the other for checking the light source.  Then part (b) asks for a diagram of an experiment; as long as every item that was checked in part (a) shows up somewhere, this part earned the full two points.

The unfortunate instinct of so many teachers and students is to exclaim, "Oh, if I just check every box and draw a labeled picture of each item, I get four points."  Well, that was true in this case.  No one, not even the development committee, could have really known that ahead of time in order to make use of that information, but sure, it's true.

Then the even more unfortunate logic says, "So, whenever you see an experimental problem, check every box and draw all the equipment, even if you're not sure you need it."  Bzzzz.  We who grade AP exams are not stupid, despite the hopes of Big Test Prep.  In other experimental questions, credit has been awarded only for a diagram that would work; or for only one specific aspect of the diagram; or credit may have been lost for silly checkmarks.  In future years, who knows what such a question's rubric might look like.  I talked with a group of readers who agreed with me that one good approach might be not to award credit for checkboxes at all!  

It's NOT true that you always get points for putting units on the answers.  Incorrect arrows on energy level diagrams or on free-body diagrams often (but not always) lose credit. 

So teach your students how not to game a test.

In your own class, use authentic and therefore unpredictable AP rubrics for authentic AP questions.  If you give your own questions on quizzes, don't set up a game-able rubric.  Don't allow students to ask questions on a test or quiz.  Don't allow the student to give two answers, hoping that one is right!  AP readers are instructed that if they see two answers* to grade the one that earns fewer points.  

* For example, "the acceleration could be to the right because of the net force, or the acceleration could be to the left because of the speeding up."

Above all, please don't even imply that the exam can be gamed.  Just encourage students to communicate how well they know physics... which is pretty darned well, because they have you to teach them, right?  There's no need for tricks!


No comments:

Post a Comment