Sunday, October 21, 2007

Testing grammar

Sunday, October 21, 2007
Testing grammar

Topic: Assignment 2 discussion and queries Date: 30 January 2007 1:22 AM
Subject:Testing grammar Author: Mondy, Steven

I think that all speaking tests are quite difficult to administer. There are just so many factors to take into consideration. Our classmate has pointed out a very good point with the testing, or at the very least the acknowledgment of the test-taker’s ability to self-correct. We are often too much caught up in the timing and correctness of procedure to allow students enough time to self-correct (a skill that may become more apparent ‘quite a ways down the track’ in their lang. learning).
From my experience (in Japan), learners tend to do better on written tasks, and have great difficulty producing oral language. There have been those students who the reverse is true, but generally written tests seem to be less invasive, especially if enough time is given. As for the question about giving a written test on paper, well, I think that I read somewhere, maybe in Hughes, maybe Bachman, that we need to maintain face validity ~ testing speaking with a written test somewhat invalidates the procedure.
I would go more for some kind of rubric, that incorporates grammar into one of the sections, along with categories including the one Nico suggested, in that the ability to self-correct is a factor in the level of proficiency that the learner has achieved. Your test seems to be more discreet point testing, and for such a case, then yes, you would have to mark them down on knowing that particular grammar, as long as you have given them enough time to answer the question, and as long as that is your focus within your objectives. A more integrative test, based on a role-play would probably pose more difficulties with rater inconsistencies, but may make grammar less of a focus, and make communication the focal point.
Also remember that, the student’s score is dependent on many factors, and the interpretation of them is relative to the other people taking the same test. Cut off scores don’t remain fixed to a magical 50%.
Steven Mondy

No comments: