Skip to content
LEAPDocs

The quiz builder

Authoring questions: types, Bloom tagging, misconception distractors and reuse across quiz groups.

Question types

MCQOne correct option plus distractors. Auto-graded.
True / falseAuto-graded. Cheap to write, weak diagnostically — use sparingly.
Short answerA brief free-text response, AI-graded against the expected answer with a confidence signal.
Open-endedLong-form. Graded against a rubric, AI-assisted, with your override final.

Bloom levels and concept tags

Every question carries a Bloom level and marks. The Bloom level is not decoration: it determines a student’s Bloom ceiling within the gate, which is what separates a student who is shallow from one who has surface gaps.

Concept tags map a question to the concepts it tests. They drive the concept map and misconception clustering in class analytics — an untagged question still marks correctly but contributes nothing diagnostic.

Aim for a spread across Bloom levels within a gate. A quiz group that is entirely remember will show every student at a ceiling of remember, which tells you nothing about any of them.

Misconception-based distractors

Generated MCQ distractors encode specific, known misconceptions rather than being plausibly wrong at random. That is what makes wrong-answer clustering work: when a third of the class picks the same distractor, analytics can name the misconception behind it instead of reporting that the question was hard.

Reuse and overrides

Questions live in a bank and are pulled into quiz groups. The same question can appear in a practice quiz and a final with different difficulty and Bloom overrides per use, so you are not maintaining two near-identical copies that drift apart.