Room 20: EFL Thoughts

I’ve just finished an interesting book: Glenn Fulcher, Testing Second Language Speaking (Pearson-Longman, 2003). I was very interested in reading it since I am often asked to help schools develop speaking sections for course assessment exams. When I saw the title I was thrilled since I expected it would explain all the theoretical issues I have with developing relatively valid oral exams.

I was disappointed though…just like so many facets of language teaching and assessment, very little has been done theoretically in oral testing. I’d previously found the same issue when investigating communicative testing. Everyone seems to think they have the best way to do things, but no one can justify why.

Apparently a lot was done in the 80s on oral testing and most of the specific references in the book refer to these articles. However, from what I read I gather many of the results are questionable.

This leaves me, a non-theoretician, no place to turn for suggestions on how to include valid techniques in my exams.

The exams I recommend are done as pair-work and between the teacher and the student. This apparently isn’t bad (whew…) . However, I find it difficult to relate the needs and financial possibilities of a school looking for an assessment instrument that is not only easy to write and administer to a large number of students, but that is also inexpensive to create. The Fulcher book has an interesting discussion of the financial burden incurred when devising an oral testing instrument for a high-stakes, proficiency exam at a school.  Incredibly he estimates (and supports the estimate) it would cost almost 63,000 pounds to do one cycle of the exam. This, of course, is beyond the possibilities of most language schools.

An oral exam, when developed by a school for use for its students must include the following:
1. It must be transparent and have face validity both for the teachers and for the students taking it. It must seem to reflect what has been done in class.
2. It must be possible for the teachers (often untrained in test writing techniques) to be able to produce alternate versions of the exam so it can be used more than once.
3. It must be easy to give and involve a minimum of class time. The teacher, who will often administer the exam and rate it alone, has to be able to test up to 30 students in a 50-minute period.
4. The rating scale has to be transparent and easy to use. It also must be compatible with the grading system used by the institution.
5. Finally, the total financial investment must be reasonable. An oral exam is almost always just one aspect of the student’s final grade and usually represents a small percentage of it. Schools are not willing to dedicate either excessive time or money to the development of an oral exam.

Now, this doesn’t mean the schools don’t want a valid test instrument. It only means they can’t spend either the time or money to develop one and adapting an instrument that was developed for a different purpose is disastrous.

So, what am I recommending. I am asking researchers to investigate realistic, inexpensive testing methods that are valid and that can be used in small-scale situations. Most investigation is dedicated to large-scale testing situations which in reality represent but a small part of all the exams that are given every day.