Is "all of the above" a bad answer option?

"All of the above" feels efficient — one item that credits knowing several facts at once — but it is one of the most consistently flagged item-writing flaws in the measurement literature.

Why it fails

It rewards partial knowledge. If an examinee is sure that options A and B are both correct, and only one answer can be selected, "all of the above" must be the key. The examinee gets full credit while knowing nothing about option C. The item's logic, not the examinee's knowledge, gives the answer away.

It rewards elimination. The flaw cuts the other way too: an examinee who is confident that any single option is wrong can eliminate "all of the above" immediately, improving their guessing odds among the rest.

Writers overuse it as the key. Because "all of the above" items are usually built by listing true statements, the inclusive option is disproportionately often correct — and students learn this. "When in doubt, pick all of the above" is standard testwiseness advice precisely because it works.

It breaks the single-best-answer logic. In a well-formed one-best-answer item, options sit on a spectrum and the examinee picks the best. "All of the above" instead asks the examinee to judge each option independently as true/false — a different, cruder task that Case and Swanson's NBME guide steers item writers away from.

For these reasons, Haladyna, Downing and Rodriguez (2002) include avoiding "all of the above" among their evidence-based item-writing guidelines, and Case and Swanson's NBME guide excludes it from acceptable one-best-answer formats. Audits of faculty-written exams — such as Tarrant, Knierim, Hayes and Ware (2006), who found writing flaws in roughly 46% of the nursing exam items they reviewed — routinely list it among the recurring flaws. "None of the above" draws similar criticism, with the added problem that examinees who select it correctly have demonstrated only what the answer isn't.

What to do instead

If the learning objective is that several things are all true (say, three features of a syndrome), you have better options:

• Split it. Write separate focused items, each asking about one feature in a way that satisfies the cover-the-options rule.
• Invert it. Ask which finding is most characteristic, most specific, or the best next step — forcing discrimination among plausible options rather than true/false verification.
• Check the data. If legacy items with "all of the above" are still in your bank, their item analysis will often show the pattern: high correct-response rates driven by cueing rather than mastery. That's a signal to rewrite before the next administration.

"All of the above" is one entry on a longer checklist of option-level flaws — see the most common MCQ item-writing flaws for the complete set.