Diagnostic Testing
What do your students know? What don't they know? How are you going to find out? John Hunt provides some initial guidance.
What do you want to know?
As with a doctor and patient, a teacher must diagnose students' language problems before offering a course of treatment. This would typically occur at the start of the course, but may be repeated at intervals to check on individual needs and to assess the effectiveness of the course.
The first job in designing a diagnostic test is to decide what exactly you want to know about your students' language skills, why you need that information and how you are going to use it.
1. WHAT DO I NEED TO KNOW?
- What they know
- What they need to revise
- What they do not know
2. WHY DO I NEED TO KNOW?
- To choose course material
- To fit my course to their needs
- To fit them into an existing course structure
3. HOW AM I GOING TO USE THE RESULTS?
- To add to my syllabus
- To delete from my syllabus
- To prepare individual profiles for remedial self-study
The more specific these questions are, the easier it is to design the items and get more useable results.
Diagnostic testing is about how much English individuals know, so there is no need to rank-order students from best to worst (although this information will appear through the results) and no need to design test items that will discriminate between students. In other words, a diagnostic test is usually criterion-referenced, the criteria being the individual aspects of English that the teacher wants to know about. The criteria, whether grammatical knowledge or communicative competence, will be decided by the teacher and the focus of the curriculum. Recent research suggests that knowledge of grammar is a reasonable indicator of successful communication, and so discrete-point grammar testing has gained a little more credibility.
Do's and don'ts
Test or item design is a creative activity, tricky even, since a test item should be designed so that it tests one specific item of knowledge only. It must be scored objectively as right or wrong with the understanding that the testee has not been disadvantaged by requiring additional information s/he may not have. The point of testing must be discrete.
In a test of an item of grammatical knowledge, for example, the task should not involve an unknown item of vocabulary; the teacher must be sure that a testee got the answer wrong because s/he didn't have the grammatical knowledge, and not because s/he didn't understand the meaning of a word used in the question. Similarly, a test of vocabulary should not depend on understanding the grammar of the context.
The same precision and clarity is also important in the rubric which tells the student what to do in a particular task. Poorly worded rubrics can sometimes require a level of understanding of English that is far higher than the actual task to be performed. A situation to be avoided! Weir gives examples of good and bad test items.
Three types of tasks
There are basically three kinds of tasks that can be used for all tests: Multiple Choice, Gap-filling, and the 'Open-ended' question that requires a two or three word answer. The choice of which task-type to use to elicit a specific piece of information is decided by the creativity of the test setter, as long as the item remains within the parameters of what is a good test question: clarity and precision. When a test has been drafted, it is important that a colleague does the test before students see it. This is to ensure that no unexpected acceptable alternative answers are possible and that the test contains no ambiguities.
Which test type?
The following characteristics of the three test types will help you to decide which type to use when you design a diagnostic test.
Multiple Choice Questions
- They can be marked clerically but theoretically students could 'pass' by guessing each answer
下一篇：计算机辅助英语口试
