How to Write Assessment Questions That Actually Predict Performance
A bad assessment question measures test-taking. A good assessment question measures job-relevant judgment. If you want better hiring outcomes, start by designing questions around the work candidates will actually need to do.
Start with the work, not the question.
Most weak assessments begin with the wrong prompt: what should we ask? Strong assessments begin with a better one: what does success in this role actually require?
Before writing questions, define the situations the candidate will face on the job. What decisions will they make? What tradeoffs will they manage? What does strong communication look like? What errors would be costly? The answers become the foundation for your evaluation.
A predictive question has a clear signal.
A question is only useful if the answer tells you something specific. Generic questions produce generic answers. Role-grounded questions reveal how a candidate reasons under realistic constraints.
For example, instead of asking whether someone is good with customers, give them a customer scenario. Instead of asking whether they are analytical, ask them to explain how they would diagnose a messy business problem. The candidate should have to demonstrate the skill, not simply claim it.
- Tie each question to one or two explicit competencies.
- Use realistic scenarios instead of trivia or brainteasers.
- Define what strong, acceptable, and weak answers look like before reviewing candidates.
- Avoid questions that reward familiarity with your company rather than ability to do the job.
Rubrics matter as much as questions.
Even a strong question can fail if the scoring is vague. Without a rubric, reviewers bring their own assumptions into the evaluation. One person rewards confidence. Another rewards depth. Another rewards speed. Suddenly the question is structured, but the decision is not.
A good rubric names the criteria, describes observable evidence, and gives reviewers a shared standard. It turns evaluation from personal preference into a repeatable process.
The best assessments create decision-ready evidence.
The purpose of an assessment is not to generate a score for the sake of having a score. The purpose is to help the team make a better decision. That means every question, answer, rubric, and recommendation should connect back to the role.
When assessment design is done well, hiring conversations change. Teams stop asking whether they liked the candidate and start asking what the evidence shows. That is where better hiring outcomes begin.