Asking Good Questions (ChatGPT)
Learn the structure of effective prompts—then iterate: Ask → Review → Revise.
🌟 Why asking good questions matters
- Clarity saves time and reduces back-and-forth.
- Context improves accuracy and relevance.
- Constraints (format, length, audience) shape usable output.
- Iteration (Ask → Review → Revise) quickly converges on quality.
🧩 Quick Prompt Template (copy & adapt)
Role: (optionally set a persona)
Task: (what you want)
Context: (course/project/audience, constraints, sources)
Format: (bullets, table, outline, steps; length)
Tone: (concise, formal, friendly, neutral)
Criteria: (rubric, must-include items, cite style if needed)
Verification: (ask to show sources, uncertainties, or next steps)
Example prompt:
"Act as a research tutor. Task: explain Porter’s Five Forces for Nike.
Context: undergrad business class; cite 2 reputable sources.
Format: numbered bullets, ≤200 words.
Tone: clear, neutral. Verification: note any assumptions."
Tip: If unsure, start simple—then add one missing element (context, format, length, or audience) each revision.
📝 Copy-ready examples
🎒 Students
"Explain the difference between primary and secondary sources for a marketing paper.
Audience: first-year undergrads. Format: 5 bullets + 2 examples. ≤150 words."
🎓 Faculty
"Create a rubric (table) to assess literature review quality: criteria, 3 performance levels,
and quick feedback phrases. Context: MBA research course."
📚 Librarians
"Draft a 10-minute mini-lesson: 'Using Business Source Complete for SWOT.'
Include learning goals, steps, demo queries, and a 3-question exit ticket."
✅ Do / ❌ Don’t
✅ Do
- State the task and the desired output format.
- Add context (audience, level, constraints).
- Specify length and citation style when needed.
- Ask for assumptions and uncertainties to be flagged.
- Iterate with Ask → Review → Revise.
❌ Don’t
- Assume the model knows your exact context.
- Leave the format ambiguous if you need a table or bullets.
- Request “everything” without prioritizing what matters.
- Skip source checking for claims or data points.
🧠 Helpful Prompt Patterns
- SCQA (Situation–Complication–Question–Answer): “Given X and Y, what’s the best Z?”
- 5W1H: Who, What, When, Where, Why, How—ensure coverage and clarity.
- Chain-of-Thought (concise): “Show brief steps (3–6) before the answer.”
- Critique → Revise: “Assess this draft with 3 fixes; then rewrite.”
- Persona: “Act as a librarian/instructor/TA for this audience: …”
🎯 Mini-Exercises (Try It Now)
- Basic → Better: Start with “Explain SWOT.” Revise by adding audience, format, and length.
- Format Switch: Ask for a table; then a 5-bullet summary; then a 120-word abstract.
- Verification: Request 2 citations and a list of uncertainties/assumptions.
Loop it: After each answer, revise one variable—context, format, or constraints.
Pro Tip: If you’ll reuse a task, save a prompt scaffold with placeholders for context, format, and length.
Authored by ChatGPT. Paste-ready template for the “ChatGPT Getting Started Guide.”
↑ Back to top