Frame the driving question and inquiry arc

Where the evidence comes from

• PBLWorks, Gold Standard project design
• A Framework for High Quality Project Based Learning

Honest about the limits: the evidence for careful driving-question design is strong. Whether adding AI to that design improves the result has not been tested. Treat this as a faster route to a first draft, not a better question by default.

1. Fix your intent first

   Write the learning intent and the authentic context in one sentence each, before you open anything. The AI cannot supply these, and everything downstream depends on them.

2. Ask for contrast, not an answer

   Several genuinely different formulations. One suggestion gives you nothing to judge against.

   Prompt 1 · Generate Copy

   CONTEXT
   I teach [subject] to [year group] at [school type] in [country].
   The project runs [number] weeks.
   Curriculum goals it must cover: [paste them].
   The authentic context is: [the real situation, audience or problem].
   What my students already know: [briefly].
   Real constraints: [time, resources, trips, budget].
   
   REQUEST
   Write five contrasting driving questions for this project.
   Make them genuinely different in angle, not five rewordings of one idea.
   
   QUALITY BAR
   Each question must be open to more than one defensible answer,
   impossible to answer with a quick search,
   and it must require the curriculum goals above to answer well.
   
   FORMAT
   For each: the question, the angle it takes, and one sentence
   on what students would have to investigate to answer it.

3. Steer the closest one

   Pick the nearest miss and say precisely what's wrong. "Too broad" is not useful feedback. "A student could answer this from one Wikipedia page" is.

   Prompt 2 · Steer Copy

   Take question [number].
   
   What works: [say what you want to keep].
   What doesn't: [be specific, e.g. a student could answer this
   from one search; the audience is invented; it doesn't need
   the curriculum content].
   
   Rewrite it three times, each fixing that problem a different way.
   Keep what works. Don't soften the challenge to make it easier.

4. Make it critique itself

   Hand it your criteria and ask for the weaknesses. Read this as a prompt to look harder, not as a verdict.

   Prompt 3 · Critique Copy

   Here is my chosen driving question: [paste it].
   
   Critique it against these criteria, one at a time:
   1. Open to more than one defensible answer
   2. Not answerable by a quick search
   3. Requires these curriculum goals: [paste them]
   4. Authentic to this context: [paste it]
   5. Feasible in [number] weeks with [constraints]
   
   For each criterion: pass or fail, the evidence for your judgement,
   and the single change that would most improve it.
   Do not rewrite the question. Do not be encouraging.

5. Build the arc backwards

   With the question settled, work out what students must build to answer it.

   Prompt 4 · Inquiry arc Copy

   My driving question is: [paste it].
   The project runs [number] weeks and covers: [curriculum goals].
   
   Build the inquiry arc:
   - The sub-questions students must answer along the way, in order
   - The knowledge or skill they need to build for each
   - The decision points where a team commits to a direction
   - The revision triggers: what evidence should make them change course
   
   Keep it to one page. Mark anything that assumes resources
   I did not mention.

Check before you use it

• Could a capable student answer this in ten minutes online? If yes, it's not a driving question.
• Can you name two defensible answers that genuinely disagree?
• Does answering it well require the curriculum content, or does it sit alongside it?
• Is the authenticity real, or did we just claim it?
• Can your actual students, with your actual weeks, get somewhere with this?