Create diagnostic questions that reveal reasoning and inform the next teaching move
7 min setupTested with:ChatGPTReviewed: 2026-08-28
1
Add your context
Your text stays in this browser. AILesson Prompts does not send it to a model or server.
2
Your prompt
Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt
Generate formative assessment questions grounded only in the supplied content.
Material, objectives, prerequisites, and boundaries:
[content]
Learner evidence, misconceptions, language, and access:
[evidence]
Count, time, response mode, grouping, and feedback timing:
[delivery]
Map every question to one objective or prerequisite and state the reasoning it should reveal. Avoid trivia, ambiguity, accidental clues, and keyword-only answers. Make selected-response distractors correspond to named misconceptions. For each item provide the acceptable answer, source-grounded rationale, likely response patterns, what each pattern does and does not prove, and a concrete teaching response. Sequence items so early answers guide later probes. Preserve cognitive demand while making wording accessible. End with coverage and timing audits; flag objectives that cannot be validly checked under these constraints.
Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.
From input to outcome
A worked example
See how concrete context turns this recipe into a usable result
Actual input
Content and objectives
Objective: distinguish correlation from causation and identify a plausible confounder. Correlation means variables vary together; causation needs evidence that changing one produces change in the other; a confounder relates to both. Observation alone does not establish causation.
Learner evidence
Adult introductory class. Several learners say strong correlation proves causation. Plain language; no calculation.
Delivery constraints
Four questions in 6 minutes using individual cards, then immediate feedback. Include one multiple choice and one transfer item.
Example output
Multiple choice: “People carrying lighters have higher lung-cancer rates. What is justified?” Correct: they are associated, but this alone does not show causation. Distractors claim direct causation, reverse causation, or that strength proves cause. If causal options dominate, reteach the observation-versus-intervention distinction.
Short response: name one factor related to both lighter carrying and cancer risk and explain both links. Accept smoking only when both links are stated; a name alone is incomplete.
Error analysis: a headline says neighborhoods with more firefighters have more damage, so firefighters increase damage. Identify the unsupported causal jump; fire size is a plausible confounder. Naming fire size without the reasoning error is partial evidence.
Transfer: employees using a wellbeing app take fewer sick days. Write one supported observation, one unsupported causal claim, and one confounder. Accept varied answers matching the supplied definitions.
Allow about 75 seconds per item plus one minute to scan cards. Both objective parts are sampled twice. Study-design knowledge is outside the stated objective and is not assessed.
Why this works
1
Response-linked actions turn questions into teaching decisions rather than mini-tests.
ChatGPT