Job Discussion Forums

[dmb]

What format do you use? Is it skills based? Etc. THe type of test I have in mind would be time consuming to sit. As it is for a private language school this might put potential learners off.

[Kent F. Kruhoeffer]

Hi dmb:

I'd vote for a multiple-choice test, especially since general office staff (with questionable proficiency in English) often end up administering and correcting them.

They're easy to write, relatively short to take, and can be corrected by anyone with an answer key.

Plus, it does give a fairly good indication of a student's level if it's written properly so as to make random guessing more difficult; i.e., variants a, b, c & d are all similar, but only one of them is correct:

example:

John _______ in Japan since 1995.

a) lives
b) will live
c) is living
d) has been living

I use 40 questions, with 10 questions at the beginner's level, 10 at the pre-intermediate level, 10 at the advanced level, and 10 that really separate the men from the boys. (or the women from the girls) Those final 10 questions could be a combination of idiomatic expressions and a few advanced grammar/vocabulary questions thrown in for good measure. (the future perfect, the subjunctive and a few TOEFL-level words)

When the scores are in, you can break them down and place your students sort of like this:

0 - 10 correct answers - Beginner Level

11 - 20 correct answers - Pre-Intermediate Level

21 - 30 correct answers - Intermediate Level

31 - 39 - Advanced Level

a "perfect" 40/40 - interview him/her for a job as teacher!

For borderline scores, I like to add a face-to-face teacher/student interview, to decide mutually if we go up a level or down a level. It's not a perfect system, but no written test will ever be perfect. The goal here is to get 'em in a group with peers of roughly equal ability as quickly and with as little stress as possible, and this seems to do the trick.

[Teacher in Rome]

I'd also add an element of production in the test. Students have to put the right form of the verb into the blanks - that sort of thing. Otherwise, you can end up with students correctly guessing the answers to the multiple choice, but not able to produce accurate English.

I agree with Kent in having a spoken test as well, although I'd have an interview / needs analysis with the student regardless of whether they are borderline or not. In Italy, for example, students often score highly on grammar, but much lower on speaking ability. It's important to have a spoken element in the testing so that they don't get put in too high a level group.

[Deconstructor]

Written multiple tests are not a good idea for the simple reason that they omit oral production, which is the most important aspect of learning English.

When I do placement tests, I simply listen for fluency and clarity; that is to say, I want to understand the level of coherency of the student's message and the level s/he understands mine. I rarely listen for specific grammar errors, rather the coherency of the production of the message.

I also want to know why s/he is learning English and what his/her goals are. Of course, at this point I get the same dumbass answer from almost every student: "English important, for job" or they acquire an idiotic look on their faces as if I just asked them to drop their pants and give me a bj.

Ultimately, it doesn't matter where you place them as long as they are in the vicinity of where they�re supposed to be since a vast majority of students shall forever remain on an intermediate level. They will always be like bad trapeze artists barely maintaining balance and never getting to the other side.

[Kent F. Kruhoeffer]

I agree wholeheartedly with both Deconstructor and Teacher in Rome regarding the desire to meet with each student on a one-to-one basis to conduct a spoken interview.

Problem is, in most of the schools I've worked, teachers simply don't have the time to meet with each student individually. Prospective students walk in the door at any and all times of the day, sometimes 3 or 4 at a time, and expect to do their placement tests right away.

"I'm sorry; all of our teachers are teaching at the moment; can you come back at 9pm?" just doesn't work, especially from the manager's point of view, whose job is to "sell" lessons.

In the end, many language schools with high student turnover find themselves in need of a placement test that doesn't require the teacher to be present, and therefore feel compelled to rely on a multiple-choice type exam, even though they know it isn't the best solution.

[ls650]

[dyak]

[Deconstructor]

[zaneth]

Phew! Somebody needs a vacation. Good thing summer's coming.

[been_there]

Back to the testing subject:

If you do multiple choice, they will be able to guess and get 25% correct, right?

Go for fill-in-the blank, or reading for comprehension. (short passage, some wh- questions at the end).

Would it be possible for you to record an audio tape of you talking about something so you can test their listening? Make 2 or 3 casette tapes and some questions to go with it. You could even record yourself describing a picture ("The ball is on the chair. There is a one-eyed penguin behind the ball." etc) and they would have to draw the picture.


Good luck.

[dmb]

Thanks for your input. I especially like

[dmb]

Another quick question. How do you eliminate subjectivity from testing? I once worked at a school that had something like 50 odd levels. each level corresponded to a particular unit in a book. When you had X amount of teachers doing the placement testing how could you get continuity? One person's level 10 is another person's level 14

[valley_girl]

We use an in-house CAEL test (a real, but unofficial, version of the actual CAEL - www.cael.ca). It is skills-based and we have found that student placements so far have been quite accurate using this test. It does require the test administrators to have some training in testing procedures and marking, however. It is also extremely time-consuming (the written part of the test is in the morning. and an interview with each student is conducted in the afternoon) so this might not be an appropriate test for a language school.

@dyak...That was too funny.

[been_there]

dmb,

I made one when I was in China, but it was a conversation between two people that the testees (testes?) had to listen to and then answer questions about. The questions started easy, multiple choice, then moved on to open-ended, "What do you think they will do next?" type.

The drawing one is a modification of a gap activity that I thought would be easy to grade by the reception staff (they just compare the pictures), but no, I've never used it as a test, although I've done it in class as a listening exercize.

[tesol1]

I agree with the multiple choice option above. It is quick and easy to both create and score. It is important, however, to have at least 2-3 examples of each target language at each level. If a student gets the correct answer on a given TL once it may be random. If, however, the same type of question is answered correctly 2-3 times you can be fairly sure it is based on true knowledge and ability rather than guesswork.