23 January 2011

Teaching to the test

I have never taught a course with a state-designated end-of-course test before.  I have prepared students for a writing test at the beginning of March (months after their course was over, or only a quarter into it).  I have taught to county-designed tests, copies of which I had even before the course started.

I have never taught English I before, so, for the sake of my students and my school's reputation, I have to learn.

So I'm trying to prepare students to feel confident and competent by the time test day rolls around.  I'm trying to diagnose problems now, do mini-lessons to appease the grammarian gods and impress the upper echelons of literary luminaries.  I figured for a midterm test, I'd go ahead and use a model test, with my own modifications for higher order thinking, time constraints, and diagnosis purposes.

Let me say this, first: do not try to do all of the adaptations the day you are giving the test.  You'll muddle questions and have to throw them out.  Granted, to some extent, this worked well for me, in that it evened numbers out to make calculation easier.  Still, it's embarrassing and probably makes life harder for the kiddos while they're taking the test.

Something I did right, however, is give students (in the end) only 30 questions and ask them to explain their answers.  I think it worked well to have the answer sheet include the questions as well as the space for explanation (no matter how time consuming it was in the adaptation process), because not only could they mark up the questions, but I could also see their thought processes more easily.  Yet it was more confusing to copy the texts they were to analyze on a separate page, and it was but a vain effort to save paper (since I printed twice, and my saintly colleague who stepped in to copy for my copied both copies as one test).  And if I'm going to be doing all of this flipping to answer a few questions at a time, why not a little more?

As for grading this test, it is not as simple as a quick zip through the scantron, as the EOC itself will be.  Plus some creative pointage is in order to make this an even 100-point affair.  So I decided students will get 1 point for answering the question (30 points possible) PLUS 1 point for getting the right answer (30 points possible) .  Then they had to explain 10 of their answers.  They get another point for explaining (10 points), and 3 more points for answers that are accurate, thorough, and complete.  I'm finding it hard to distinguish between thorough and complete on some of these, but mainly it comes down to are they missing something for their explanation to make sense (complete)?  Or are they missing something to confirm their answer another way (thorough)?

At least it's still possible to get the answer wrong and get 5 points for the question, right?  In theory, it's possible to get all of the questions wrong, and still pass, but the likelihood is slim, and "thorough and complete" is killing more of them than "right or wrong" is.  And though they only had to explain 10, there was a reward in explaining more (not advertised), as a few students who struggled still got 100 or close to it.

In the end, though, I have 18 out of 20 passing the test, which, according to my observations, represents the ones who have a shot at passing the EOC this year. After all, as my erstwhile fearless leader said, there are some "even God can't get to pass."

Be it noted also: it has taken me approximately 3 hours to grade these 19 tests, but that is including time for entering grades, grading a whole set of quizzes in between for variety, a lengthy calculation process, and, of course, blogging.  All in all, I think this will help me prepare them for the EOC while incorporating higher order thinking.

No comments:

Post a Comment