How many distractors should a multiple-choice question have?

The traditional answer — four or five options, so three or four distractors — is a convention, not an evidence-based requirement. The research points the other way.

What the evidence says

Rodriguez (2005) conducted a meta-analysis spanning roughly 80 years of studies on the optimal number of options and concluded that three options (one correct answer plus two distractors) is generally sufficient. Three-option items showed reliability and discrimination comparable to four- and five-option versions of the same questions, while taking less time to answer and — critically — less time to write.

The reason is practical: item writers usually cannot produce more than two genuinely plausible distractors. The third and fourth distractors added to satisfy a format rule tend to be implausible, and an implausible distractor is functionally invisible. If almost no examinee selects an option, the item behaves as if that option didn't exist — a five-option item with two dead distractors is really a three-option item with extra reading load.

Haladyna, Downing and Rodriguez (2002) made the same point in their item-writing guidelines: write as many good distractors as the content supports, rather than a fixed number. Case and Swanson's NBME guide likewise emphasizes that every distractor should be plausible to an examinee who has not mastered the objective — often a common misconception, a related-but-wrong entity, or a plausible next step that's premature.

What about guessing?

A common objection is that fewer options raise the guessing probability (33% for three options versus 20% for five). In practice this matters less than it seems: guessing odds apply per item, and over a full-length exam the effect on total scores is small — especially since the "extra" options on longer items were often eliminable anyway. Reliability at the test level is preserved, and the writing time saved can go into producing more items, which improves reliability far more than extra distractors do.

How to tell whether your distractors are working

The empirical check is distractor analysis: after the exam, look at how many examinees chose each option, and whether low-performing examinees chose distractors more often than high performers. A distractor selected by under about 5% of examinees is a candidate for revision or removal. See what is distractor analysis for how to read those numbers, and how to write flaw-free MCQs for techniques for generating plausible distractors in the first place.

Practical guidance

• Write the stem and correct answer first, then add distractors until you run out of plausible ones — typically two or three.
• Never pad with throwaway options, "all of the above," or "none of the above" just to reach a house-standard count.
• Review distractor statistics after each administration and retire options nobody picks.