Writing better multiple-choice questions

Multiple-choice questions are quick to grade, but they are surprisingly hard to write well. A weak question can be guessed without understanding the material, or it can confuse strong students who read too closely. A good one measures exactly what you intended. The difference usually comes down to a few habits.

Start with the stem – the part that asks the question. A good stem poses one clear problem and gives enough context to answer it without reading the options first. Avoid negatives like "which of the following is NOT" when you can, because they trip up students who understand the content but misread the wording. If you must use a negative, make it stand out so it is impossible to miss.

The options are where most questions succeed or fail. Keep them parallel in length, grammar, and structure so that no single answer stands out for the wrong reason. The classic giveaway is the correct answer being noticeably longer or more detailed than the rest. Distractors – the wrong options – should be plausible, ideally based on common mistakes or misconceptions, so a student who has not learned the material cannot rule them out by logic alone.

Test each question against a simple bar: could a student who knows the topic answer it confidently, and could a student who does not be genuinely tempted by a wrong option? If every distractor is obviously absurd, the question is too easy and measures nothing. If even you have to squint at two options, the question may be ambiguous and needs tightening. Aim for one clearly correct answer and three believable wrong ones.

This is where Portfelim helps. When you generate multiple-choice questions from your material, it drafts the stem, the correct answer, and a set of distractors drawn from the content itself – which gives you plausible wrong options instead of filler. You get a full set of questions in seconds, then refine them using the principles above: tighten a stem, swap a weak distractor, or balance the option lengths.

Treat the generated questions as a strong first draft, not a final exam. Read each one as a student would, fix anything ambiguous, and make sure the correct answer is genuinely the best one. You will still apply your own judgment – but you will spend your time editing good questions instead of inventing them from nothing.