Writing a good test
What makes a test good or bad? The most basic and obvious answer to that question is that good tests measure what you want to measure, and bad tests do not. However, other factors are important in making a good test. The following are some characteristics of tests that need to be considered in making, administering, and evaluating them.
Goals
In writing a test, you must always keep in mind the goal of the test. The goal of the test is what you want to measure. To the greatest extent possible, the test should measure just what you want it to measure and nothing else. (However, if you are trying to test English communicatively, it is necessary to test more than one skill at once.) If you are trying to test one skill or area of knowledge, you should avoid a method of testing that requires another skill or area of knowledge. For example, if you want to test your students' knowledge of vocabulary, you need to make sure that your test is testing vocabulary knowledge. If a knowledge 0f grammar or reading is required for the students to respond to the vocabulary items, then the test is testing something other than vocabulary knowledge. You will not know whether a student messes a question due
to lack of knowledge of the vocabulary item or lack of knowledge of the grammatical form. It is impossible to entirely avoid having elements of grammar or reading in a vocabulary test, but the test should be designed so that these elements can be minimized. For example, if a vocabulary item is presented in context, the grammar of the context sentence should be well within the proficiency of the testees.
Audience
You need to consider your audience carefully. In language testing, the level of difficulty of the language tested is very important. If you have a homogeneous group, you need a test that is the right level for that group. If it is too difficult for the group, all of them will do poorly, and you will not be able to determine anything about their knowledge or skill and how they differ in that knowledge or skill. If the group is heterogeneous, the problem is more difficult. You need to include some easy and some difficult questions, so that every testee has some questions within his/her ability level and every testee has some question that are challenging.
Weighting
In writing a test, you need to consider how much weight each
item, each skill, etc, is given. It seems obvious to say that you should give the greatest weight to the most important things you are testing, but this does not always happen. It is tempting to devote the largest part of the test to whatever is easiest to test, not to what is most important to test. Therefore, the test maker needs to carefully consider what should be tested, depending on the syllabus of the class or other bases for the test and make sure that the weighting
reflects these.
Clarity
|
|
|
|
Length
In theory, the more items a test has, the more reliable it is. That is, on a short test, a few wrong answers can have a great effect on the overall results, but in a long test, a few wrong answers won't influence the results as much. However, in practice, a long test has drawbacks. If a test is too long, and particularly if students are doing the same kind of item over and over, they may get tired and not respond accurately or seriously. If there are reasons why a test needs to be long, it is a good to divide the test into sections with different kinds of tasks, so that the variety of tasks will hell students maintain their concentration.
Simplicity
There are many ways to measure language-related skills, and obviously some are more complicated than others. If you choose a more complicated way to test English, there can be a greater likelihood of errors or a greater likelihood that the test is testing something other than it is indented to test. Thus, it is important to keep the method of testing as simple as possible, while still testing the skill you intend to test. For example, a logic problem might be used to test reading ability, and the teacher might conclude that if the student got the right answer on the logic problem, he/she had
good reading comprehension. However, such a test would also be testing the student's ability to work the logic problem, a skill unrelated to reading. Objectivity
Tests should be as objective as possible, that is, should use the same standard to measure all students. This does not mean that we can use only objective tests, but it does mean that when we use subjective tests, we need to establish objective ways to grade, and we have to decrease the difference of in standards as much as possible. When you test language in ways that cannot be tested completely objectively, you need to set up concrete grading scales and train graders to use them properly. It is also useful to use more than one grader for each test. For example, in testing writing, it is useful to make a scale, with separate for different aspects of writing, such as content, vocabulary, organization, and grammar, and a description of what would be expected at each level in each section. Such a scale helps the graders be as objective as possible about scoring compositions .
Practicality
This is similar to simplicity. We have to make tests which we can administer. For example, even if we do not have many video tape cameras, if we make the test to video tape many students' performance, it might not be possible to carry out or at least is difficult to do. When you make test questions, you have to remember all tests have to be administered.
Economy
Unfortunately whatever we do we always have certain budgets. We have to make and administer tests within certain cost. If we include color copying, it will be more expensive. If we need some recording, particularly video recording, it will be very expensive. However, if we make tests which will be very difficult to administer or grade, that will be more expensive. The most expensive part is probably labor these days. Thus, if the test needs many people to administer or grade, it will be very expensive. It is always important to think the cost of the cost.
|
|
This is related to practicality and economy of the test. In some cases, particularly when testing writing or speaking, it is necessary to make relatively complicated scoring systems in order to make scoring more objective. However, if the procedure is complicated, it will cost more, it will take longer, and it will include more errors and/or subjective judgments. Therefore, it
is necessary to consider the best way to score a test, to find a balance between a scoring system that is fair to the students and objective and one that is reasonably easy to administer.
Pretesting
|
|
Summary
Making economical, practical, scorable tests which meet your goals and which are of an appropriate level of difficulty is an important part of English instruction. These factors should always be considered in making tests.
Prepared by :
Zahra torabi