Is the Matching Test a Good or Bad Test? An Evaluation Based on Gronlund and Waugh's Principles

Look at the sample test.

Match the words in Column A with the premises in Column B.

a. What kind of test is the sample above?

b. Decide whether it is a good or bad sample test! Give at least 3 arguments for your answer based on the principles for constructing test by Gronlund and Waugh (2009)!

Answers 1:

a. The kind of test on the sample

The kind of test on the sample is Matching-type. According to Fachrurrazy & Tresnadewi (2019), this test is basically like the multiple-choice test, but the format is made in two columns like as the sample, Column A (left column), which is called premise, and column B (right column) which is called response. On the sample, the list of words in column A must be searched for the opposite (antonym) words in column B.

b. Decide whether it is a good or bad sample test! With at least 3 arguments based on the principle for constructing tests by Gronlund and Waugh (2009)!

In the type matching test, each answer choice is usually only used once, but actually can be used more than once. If each choice is only used once, then the list in the answer should be made longer by adding 1-3 additional choices, it doesn't have to be too many. In the format above, the number of items in the premise is 5, and the number of choices in the answers is also 5, meaning the same as the list in the premise.

Fachrurrazzy and Tresnadewi (2019) state that don’t make the same number in both column A and column B. If each option in the answer can be used more than once, the number of options can be made the same or even smaller than the number of options. The number of items on premises. However, if the latter is chosen, be sure to inform the examinee about the possibility of using the answer more than once in the test instructions, to avoid misunderstandings. This matching type test is appropriate when the items in the premise are homogeneous; if not, it is better to use another type of question, such as multiple choice or binary choice (Fachrurrazy and Tresnadewi, 2019).

Based on Gronlund and Waugh in Fachrurrazy and Tresnadewi (2019), the strengths of the matching type test are as follows:

A concise and efficient form is provided in which the same set of responses corresponds to a series of stem items (i.e. premise)
Short reading and response time
This item type is easy to construct if changed from a multiple-choice item that has the same set of alternatives.
Easy, objective, and reliable scoring

And the weaknesses of this test according to Gronlund and Waugh in Fachrurrazy and Tresnadewi (2019) are as follows:

This type of item is mostly limited to simple knowledge output by association.
It is difficult to create items that contain a sufficient number of homogeneous responses.
Greater susceptibility to irrelevant clues than other types.

However, based on the principle in constructing the test by Gronlund and Waugh (in Fahrurrazy and Tresnadewi, 2019), the sample of the test, there are several things that are not suitable as the principles stated by Gronlund and Waugh as follows:

Employ homogeneous lists in a matching-type test. The example shows that the items in the premise and the options in the response are respectively homogeneous. But from the example is not homogeneous. And we can conclude that the type of test on the sample is not a good test.
Use a larger, or smaller, number of responses than premises. When each option in the response is used once, make the number of options in the response larger. But the example is not shown like that, because the number of premises and the number of responses are the same (5 in column A, and also 5 in column B). The sample of the test is not shown as larger or smaller in the number of responses (Column B) than the premise (Column A), as stated also by Fachrurrazzy and Tresnadewi (2019) state that avoid making the same number of items in both column A and column B. So, it is a bad sample test.
Describe the basis for matching and the number of times a response may be used. But in the example is no clear instruction (for example Match the words in Column A with their antonyms in Column B), and there is one item in the premise where there is no answer in the response, so this is a bad sample test.

And we can conclude that the example of a test in the sample is not a good test type (bad sample test) because of 3 reasons above that are not suitable with the principles of building a good test as stated by Gronlund and Waugh.

Reference:

Fachrurrazy and Tresnadewi, Sintha. 2019. Assessment in Language Teaching. Tangerang Selatan, Banten: Universitas Terbuka.

Answer 2:

a. The sample test above is a matching test.

b. Whether the sample test is good or bad can be decided based on the following principles for constructing tests by Gronlund and Waugh (2009):

Arguments for the sample test being good:

The test is aligned with the learning objectives. The learning objectives of a matching test are typically to assess students' ability to match related concepts, terms, or definitions. The sample test above matches the words in Column A with the premises in Column B, which suggests that the test is aligned with the learning objectives.

The test items are clear and unambiguous. The items in the sample test are short and to the point, and they do not contain any ambiguous or confusing language.The test items are of appropriate difficulty. The items in the sample test range in difficulty from easy (e.g., matching "long" with "short") to more challenging (e.g., matching "come" with "go"). This suggests that the test is appropriate for a wide range of students.

Arguments for the sample test being bad:

The test is too short. Matching tests can be used to assess a wide range of knowledge and skills, but they are most effective when used in conjunction with other types of test items, such as multiple choice, short answer, and essay items. The sample test above is only five items long, which may not be sufficient to assess students' knowledge and skills comprehensively.

The test does not include a variety of item types. As mentioned above, matching tests are most effective when used in conjunction with other types of test items. The sample test above only includes one type of item, which may make it less effective at assessing students' knowledge and skills.

The test items are not scored consistently. Some items in the sample test may be more difficult to score than others. For example, the item "match 'come' with 'go'" may be more difficult to score than the item "match 'long' with 'short'." This inconsistency could lead to inaccurate scores for students.

Overall, the sample test has both strengths and weaknesses. It is aligned with the learning objectives, the test items are clear and unambiguous, and the test items are of appropriate difficulty. However, the test is too short, it does not include a variety of item types, and the test items are not scored consistently.

Reference:

Gronlund, N. E., & Waugh, C. K. (2009). Assessment of student achievement. Pearson Education