Quiz & Assessment Creator

Quizzes have three jobs in an online course and most creators only think about one. Verification — did the student actually learn the material — is the obvious one. But the other two are where the real value is. Reinforcement: the act of retrieving information strengthens memory more than re-reading or re-watching ever does. It's called the testing effect and it's one of the best-established findings in cognitive science. A quiz isn't a measurement tool. It's a learning tool, and design choices either lean into that or waste it. Motivation: passing tells students they're progressing, failing tells them exactly where to focus, and both increase completion rates compared to courses with no assessment at all. Most quizzes are written like trivia games because the creator was only thinking about the verification job.

Question design is anchored in Bloom's Taxonomy and most course quizzes accidentally sit at Levels 1-2 — Remember and Understand. The skill targets a mix weighted toward Levels 3-4 (Apply, Analyse) because those are the levels that actually predict whether the student can do anything with the material. Pure recall is capped at 20% of questions and only used for essential vocabulary that's genuinely prerequisite. The rest is scenarios — given this client situation, which approach? Given these two strategies, what's the trade-off? Given this output, what's wrong with it? Those are the questions that test capability.

The rules for writing questions are strict and the rules for not writing them are stricter. Wrong answers must be plausible — if the wrong options are obviously wrong, the question tests elimination, not knowledge. Test one concept per question; compound questions confuse rather than assess. No trick questions, no "all of the above," no "none of the above," no negative framing ("which of these is NOT…") because errors from misreading aren't errors from misunderstanding. And no testing trivial details unless the date or the name is itself meaningfully important. The "what year was X invented" question is the marker of a course that thinks quizzes are decorative.

Answer explanations get written for every option, correct and incorrect. The explanation for the right answer confirms the reasoning; the explanations for the wrong ones diagnose the specific misconception that led there. That's where the testing effect actually pays off — the student who picked B when the answer was C reads why C was right and why their B-reasoning was flawed. The explanation is often more valuable than the question itself, and the skill treats it that way.

Output is the full question set with question text, options, correct answer marked, explanations for every option, and a Bloom's level tag per question; a separate answer key; assessment metadata (type, count, time estimate, pass threshold, level mix); a randomisation strategy that names which questions can safely shuffle (order-independent) and which can't (scaffolded scenarios); a retake policy recommendation; accessibility notes (no colour-only or image-only questions, reading level checked); and, for courses with retakes, a question bank sized 1.5-2x the quiz length with even concept coverage. There's a diagnostic pre-assessment option for courses that benefit from showing students where to focus. If a question doesn't map to a stated learning outcome, it gets cut — orphan questions testing untaught content are the second-fastest way to destroy trust in an assessment.