LunaPromptsThe candidate knows the concept but misses the question because two answers could be correct under different assumptions. The score now says more about the wording than the skill the team wanted to assess.
A useful technical MCQ makes its context explicit. Its wrong answers reveal plausible misunderstandings, and its right answer leads naturally into a practical task.
LunaPrompts multiple-choice question assessments help hiring teams evaluate technical knowledge alongside practical tasks. Use MCQs to check relevant concepts, identify gaps, and guide deeper assessment in coding, AI, or system design.
Check evidence-grounded answering
Assessment authorIllustrative scenario01 / 03The source does not contain the answer
The task requires factual answers to be supported by the supplied documents.
Start with the scenario in the interactive question below: an assistant must support factual answers with the supplied documents, but those documents do not answer the user’s question. The best response for this scenario acknowledges insufficient evidence.
Each distractor should represent a different misunderstanding. Answering from general knowledge confuses likely truth with support from the required source. Citing the closest document confuses relevance with evidence. Repeating the question avoids the task without explaining the limitation. These alternatives give the reviewer something more useful than random wrong answers.
Now examine whether the wording introduces ambiguity. If the task allowed general knowledge, the intended answer might need a different explanation. If it required a specific fallback format, say so. The candidate should be assessed against the stated context, not the author’s unstated preferences.
Pair the concept with a small practical task. Ask the candidate to implement the fallback behavior and show an example where a superficially related document still does not support an answer. That creates a clear bridge from selecting the right principle to applying it. Use MCQs to sample relevant foundations, then spend practical assessment time on the capabilities the role depends on most.
Choose concepts that matter to the role
Start with decisions the candidate may need to make at work. A backend engineer may need to understand request behavior and data consistency. A data engineer may need to reason about joins and aggregation. An AI engineer may need to recognize a retrieval or evaluation problem.
Avoid filling the assessment with facts that are easy to look up and unrelated to the role's responsibilities. The useful question tests understanding that helps someone act correctly when the situation changes.
Write questions that reveal a meaningful distinction
A good question has a clear context, one defensible best answer for that context, and alternatives that reflect plausible misunderstandings. The wording should not require candidates to guess what the author intended.
For example, a question about API retries becomes more useful when it specifies whether the operation creates a new record and whether the same request might be submitted twice. That context lets the reviewer assess the intended concept rather than test-taking intuition.
Combine breadth with practical depth
Use a small set of knowledge questions to cover important foundations. Then choose a practical exercise that lets the candidate demonstrate the most consequential skill for the role.
For an AI application role, that might mean a few questions about retrieval and evaluation followed by a small RAG task. For a software role, conceptual questions can precede a coding or debugging exercise.
Keep the total assessment proportionate. A candidate should be able to understand why each section is part of the process.
Interpret MCQ results in context
A high MCQ score shows success on those questions. It does not establish that the candidate can implement, debug, or explain a production system. A missed answer may identify a useful follow-up rather than a reason to reject someone automatically.
Review the relevance and difficulty of the questions, the rest of the assessment evidence, and the role's actual requirements. Where question variation is used, check the quality and balance of the underlying pool.
A review framework your team can use
The following is an editorial checklist for this workflow. Adapt it to the role, the candidate journey, and your configured environment.
| Review area | What to establish |
|---|---|
| Context | A clear scenario and its constraints |
| Best answer | One defensible choice in that context |
| Distractors | Plausible, distinct misunderstandings |
| Follow-up | A practical demonstration of the concept |
Avoid trick wording and “all of the above” shortcuts that obscure the concept. Review the explanation for each option before using the question.
The documents do not contain the answer. What next?
Every factual answer must be supported by the supplied documents. Select the best response under that rule.
Next, ask the candidate to implement this behavior. Recognition and practical application provide different evidence.
Put the guide to work
Answer the example below, then write one role-specific MCQ. Explain why every option is right or wrong under the stated assumptions.
Download the editable worksheet to record the decisions and open questions with your team. The examples in this guide are illustrative, not customer results or validated selection rules. Product availability and setup depend on the workflow agreed for your account.
A focused checklist, a worked scenario, and questions to resolve together.
Frequently asked questions
Are MCQs enough for a technical hiring decision?
They can establish useful foundational evidence, but practical engineering roles also need evidence of application. Combine them with appropriate tasks and interviews.
Can MCQs be included with other assessment formats?
Yes. LunaPrompts' assessment offering combines knowledge questions with practical formats such as coding, AI tasks, and system design.
How many questions should we include?
Include enough to cover the necessary concepts without adding repetitive work. The right number depends on the role, question complexity, and the rest of the assessment.
Keep building your hiring workflow
- JD to assessment: The job description is ready. The assessment still needs judgment.
- Coding assessments: The code passes the example. What happens on the second event?
- Question randomization: Different questions. The same skills to demonstrate.
Explore mcq assessments with LunaPrompts or browse all 19 hiring field guides.
