Week 8: Test Week

There's not much class stuff to reflect on this week as all I've been doing is giving tests and interviews.  But this may be a good opportunity to reflect on my testing and grading policies.

My classes are much smaller this year compared to last year which has afforded me the time to test my students in a way more relevant to my teaching philosophy.  In my classes, I try to focus on productive skills but part of the challenge of doing that is actually making a test that is relevant to those skills.  One of the problems with testing speaking, for example, is that it is not something that can easily be set up.  It is simply not something that can be done together as a class, unless you have a lab with recording equipment (which I don't) and a computer system setup that is easy to use and troubleshoot (which doesn't exist).  So if I want to test speaking, usually I have to go through the time consuming process of interviewing students one by one.  This is possible with smaller classes, which is great, but on the other hand there is not much time to test each student.  And with a smaller sample of speaking it inflates the importance of even the smallest of mistakes.  This part is not ideal as some of those mistakes are random in nature.  In fact, I'm sure it affects the true accuracy of my evaluation.  (For example, if I happen to ask a question a student doesn't understand well they would get a significantly lower score than if I happen to ask a question that the student student happens to have a good understanding of.)  So on the one hand I have a desire to test productive skills in line with my philosophy but on the other I have time limitations that make it quite difficult to execute well.

You might ask, why test that way at all if it has a higher degree of inaccuracy.  Part of the motivation for me is that without a production based test to back up my production based lessons, students won't move outside their comfort zones in terms of class participation and practice.  If they think that the test is simply a listening and reading test then they would do what they always do in English class... listen and read.  Good skills, but not the skills they need to develop at this point to communicate effectively.  The best compromise I've been able to come up with so far is giving a short 20 minute listening and grammar test in conjunction with the interview in order to more accurately gauge student level.  But my fear even with this is that students focus on this part more than the interview itself.  Even if the interview has a significant portion of the grade allocated to it.

Perhaps the best thing I can do for next semester is to throw out the written test completely and assign less points to midterm and final interviews.  Then I could assign more points to in class work and participation to help mitigate the inconsistencies inherent in the short interview samples.