Writing Better Objective Tests

Joseph Ryan, Department of Education

The objective test is only one of many ways in which students can be evaluated. Tests can be formal or informal, oral or written; and no one form of testing is necessarily better or worse than another. Objective tests, however, do offer some advantages over other forms of testing. By definition these testing procedures are more objective than other procedures. That is, they are less dependent on personal opinion than some other forms of testing. Objective tests also tend to be more reliable than other types of testing; and the objective format allows instructors to test a large number of students on a wide range of topics in a relatively brief period of time.

Before an appropriate test can be written, the knowledge or skills taught in class need to be defined with some care. Instructors must determine the content or material that has been taught and the type of skill a student should be able to demonstrate with respect to the content or material. This is illustrated by the three test items below.

  • Item 1: At which temperature does water freeze?
    1. 0 degrees Fahrenheit
    2. 30 degrees Fahrenheit
    3. -273 degrees Fahrenheit
    4. 212 degrees Fahrenheit

  • Item 2: At which temperature does water freeze?

    _______________degrees F.

  • Item 3: The lowest temperatures on Friday, Saturday, and Sunday nights were 33 degrees F., 47 degrees F., and 22 degrees F. On which morning would you expect to find water frozen in a bucket that was left out over night?
    1. Friday
    2. Saturday
    3. Sunday

Each of these three items deals with the same content, namely the freezing temperature of water. The items differ, however, in the intellectual skills that must be applied to that content. In Item 1 the student must recognize the freezing temperature of water. In Item 2 the student must recall this temperature. In Item 3 the student must demonstrate comprehension of the freezing temperature of water. When writing the objective tests, review each item after it has been written to judge whether the content and skill it requires have, in fact, been taught in class.

Multiple Choice Items

Multiple choice questions contain two major parts, the stem which presents the problem and several alter-native answers. The following checklist can be used to create or evaluate multiple choice questions.

  • The stem, not the responses, should introduce what is expected of the student.

  • The stem should be free of irrelevant material.

  • All the options should be plausible and homogenous.

  • All the options should be grammatically consistent with the stem.

  • Obvious verbal associations between the stem and the correct answer should be eliminated.

  • Overlapping options should be eliminated.

  • All options should be approximately the same length.

True-False Items

A true-false test item is written in the form of a declarative sentence. The student must judge whether the sentence is a true or a false statement. Some instructors prefer to use the true-false format with the additional requirement that students indicate how the false items can be changed to make them true. This adaptation requires that the instructor provide very clear standards for scoring these answers.

Use the following checklist to create or evaluate true/false items.

  1. The language of the items should be simple and clear.

  2. The statement should be specific enough to allow a judgement to be made.

  3. The statement should be clearly true or false.

  4. Specific determiners (e.g., always, never, sometimes, ever) should be avoided.

  5. Use only a single idea in each statement.

  6. The number of true statements and false statements should be approximately equal.

Matching Items

The matching item is a modification of the multiple choice question. In a matching test item, a list of words or phrases is presented in a column, generally on the left side of the page. These words or phrases are called the premises of the item. A second column, generally on the right side of the page, contains words or phrases called responses that are to be matched with the premises.

When there are exactly as many premises as there are responses and when each response is used once and only once in the matching process, the test item is said to have perfect matching. When some of the responses are used more than once or not at all, the item is said to have imperfect matching. Imperfect matching makes guessing more difficult.

Following are suggestions for writing matching test items.

  1. Clearly explain the basis on which the matching is to be made in the directions.

  2. Make sure that the directions make clear whether each response can be used only once or not at all. It is usually better to have more responses than premises and to state that each response may be used more than once and that some responses may not be used at all.

  3. Keep the lists of premises and responses short (5 or 6). If the lists are too long, the items will be testing the students memory and reading skills.

  4. Keep the lists of premises and responses relatively homogeneous.

  5. Write the responses in the form of short phrases, single words, numbers, or symbols and arrange them in an obvious order--alphabetical, chronological, etc.

Concluding Suggestions:

The preceding sections offer specific recommendations for improving the writing of three types of objective test questions. In addition, the following general guidelines may be useful when preparing any type of objective test item.

  1. Design each item to measure an important learning outcome as defined by course objectives.

  2. Include only one central idea in each test item.

  3. Write the stem and options for each item in simple, clear language.

  4. Do not make items more difficult through use of tricks of ambiguity. Increase the difficulty level by changing the stems or options.
  5. Make each test item independent of other items on the test.

  6. If negatives are used, the negative should be emphasized by capitalization or underlining, e. g., NOT, none, NEVER.

    *Excerpted from:
    Ryan, J. , Lackey, G. & Bell, M., (1981). Improving your classroom tests: Writing better objective questions. University of South Carolina, Department of Educational Research and Psychology.