Multiple-Choice Tests

Why Use Multiple-Choice Tests?

A multitude of scholars, including Thomas Haladyna and Steven Downing, have indicated a number of strengths of multiple-choice testing. Sampling of content can be excellent, supporting content-valid test interpretations. Reliability of scores can be high. Items [Page 710]can be easily pretested, scored, used, and reused. Objective, high-speed scoring is possible. Diagnostic subscores are easily obtained. Adaptive testing can be done well with this format. Most content can be tested with this format, including many types of higher-level thinking, even though many writers limit themselves to pedestrian, knowledge-level items. It is a flexible format, and guessing is seldom a problem. If a test-taker is troubled by an item, another item, and opportunity, awaits.

Formal and Informal Uses of Multiple-Choice Tests

Multiple-choice tests are the typical choice to provide big summative evaluations such as whether someone passes a course or earns a professional license. Yet there are also instructional or informal times when a multiple-choice format can encourage student participation, provide corrective feedback to the instructor and students, focus everyone's attention on a topic, give a break from the instructor's droning, preview and prepare for a test, and even allow a microbit of humor.

A Few Goals for Test Developers

Multiple-choice tests should ask important questions consistent with educational objectives or with the projected uses for the test, and ask them clearly. An early National Assessment guideline essentially stated that an item be important and clear enough to be published in the newspaper. Roger Boothroyd concluded that test-takers should either answer correctly for the right reason(s) or, if answering incorrectly, do so for the right reason(s). He is consistent with Lee J. Cronbach's counsel to minimize irrelevant difficulties. Test items are perhaps our most closely read literature and deserve to be based on vital questions that are cleanly stated.

Guidelines for Selecting, Revising, and Writing Good Items

There are three ways to develop test items: selecting existing items, revising existing items, and writing new ones. Educators should evaluate their testing method to be sure that the test items correlate with educational objectives, as some of these methods may prove to be problematic. For example, David Frisbie and others found that tests accompanying textbooks are likely to include

• 1.
  Trivial items, even based merely on phrases lifted from the text,
• 2.
  Incomplete coverage,
• 3.
  Irrelevant items,
• 4.
  Little testing of application, understanding, and critical thinking, and
• 5.
  Item faults.

That is, developing a test by appropriating items from a test accompanying the textbook would likely provide a poor sample of items relative to the objectives of the course, with many trivial, irrelevant, low-level items; few items testing higher-level thinking; and big gaps in coverage. Additionally, items may contain errors in item writing, especially item-writing faults that either lead a test-taker to answer correctly for the wrong reason or answer incorrectly for the wrong reason, as cautioned against by Boothroyd and illustrated in Item 5 below.

...