What is the best kind of language testing?
For English learners.
If the person being tested is going into an English language school or English language program I think it is unnecessary to test the learner. I believe that language programs should be constructed in such a way that any learner can easily find content to learn from that is interesting and at his or her level. The learner should be able to start any time and start at any level.
The responsibility of the school or teaching program should be to make a full range of learning opportunities available to the learner and then encourage the learner to explore the language and learn from it. Language learning is an individual task. You do most of the learning on your own, reading and listening and absorbing vocabulary.
The teachers are there to explain where necessary, provide feedback and encouragement, but not to set the learning agenda. They have to stimulate the learners to set their own agenda. Most of the learning has to be done by the learners. If the learners do not become autonomous, they will not learn very well.
It should not be necessary to do an evaluation of the learner’s level in order to place them in an appropriate class. In any case, any attempt to classify the learners will be in vain. Different learners will learn at different speeds. All have different interests, regardless of how they do on a placement test. As long as they pursue their interests in a systematic way, they will learn.
It may be advisable to have learners of similar levels join in the same conversation class. Even here, however, some weaker learners may benefit from joining an advanced discussion group. In any case the interests of the learners should be the main determining factor. If discussion sessions are available on different subjects, the learners should have the freedom to attend the sessions that interest them. If some learners are simply unable to follow the conversation they can easily be asked to (also) attend other groups that are closer to their level. In any case, prior testing is not necessary.
For students attending university in English.
Universities have a need to decide if non-native speakers have the ability to follow university courses in English. However, the students’ ability to pass TOEFL is not proving to be a good measure of their abilities. Some students simply master the techniques of passing these impersonal tests. They spend a lot of effort on this activity, which is a waste of their time. Other students may not do as well on the tests but may improve more quickly once they are in an English speaking environment. Should they be passed over in favour of a student who has mastered the techniques of multiple choice tests?
I think any test of English ability should be dynamic and cover a period of many months. All of the efforts of the students to prepare for their evaluation should be captured. Students being tested should have to follow a relevant course of study on line and regularly answer written and oral questions. The accuracy of their syntax, the richness of their vocabulary and their ability to understand English and to express themselves can be measured quite accurately over time. This can be represented in a statistical profile of the learner. We do this at The Linguist.
The learner thus builds up a profile of his or her English ability while improving at the same time. The student’s effort is thus directed towards learning English rather than preparing for the TOEFL test.
At any time the learner should be able to submit to a testing agency to confirm that this longer term profile accurately reflects the learner’s present situation. In other words the profile is confirmed on a pass/fail basis. These results and the whole profile are made available to any university the learner is applying to enter.
For employees.
A similar language profile system could easily be adopted to provide an employer with an accurate, statistics-based profile of an employee or potential employee’s language level.
In this way the learner’s preparation for language evaluation is not separated from the process of learning and using the language as is now the case. The profile that is developed is easily viewed, in terms of statistics and in terms of examples of the learners writing and recorded oral expression. When required this profile only needs to be confirmed.
This above is the kind of testing process that we are developing at The Linguist.
