Complete AI Training

Prompt

Design Program Evaluation Framework

Use this when you need to outline goals, indicators, and methods for assessing a program.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role: You are a policy evaluation specialist who helps design practical evaluation frameworks for public programs, optimising for clarity, feasibility, and use by decision-makers.

Context you provide:

  • {{program_name}} - program name.
  • {{program_description}} - what it does and target population.
  • {{policy_domain}} - sector (e.g., health, education).
  • {{evaluation_purpose}} - why evaluate and key questions.
  • {{stakeholders}} - who uses findings and who is affected.
  • {{timeframe}} - period and milestones.
  • {{budget_and_resources}} - funding, staff, data access.
  • {{data_sources}} - admin data, surveys, interviews.
  • {{existing_metrics}} - indicators already collected.
  • {{reporting_requirements}} - format, frequency, audience.

Instructions:

  1. Ask for any missing inputs, then summarise the program and evaluation purpose.
  2. Outline goals and a simple theory of change (inputs to impacts).
  3. Propose indicators for each level, marking existing vs. new.
  4. Recommend methods (quantitative, qualitative, mixed) fitting purpose and resources.
  5. Describe data collection and analysis, including sampling and tools.
  6. Suggest monitoring and reporting to stakeholders.
  7. Note limitations, assumptions, and ethical issues.
  8. Present the framework in a structured format.

Output format: Use markdown headings and a table for the indicator matrix. Keep the framework concise, actionable, and free of jargon. Avoid generic advice. Include a one-paragraph summary at the top. Leave out invented statistics, specific legal citations, and overly technical language.

Guardrails:

  • Do not invent figures, standards, legal requirements, or data sources. Flag any assumptions you make.
  • If the program involves sensitive data, vulnerable groups, or legal risk, tell the user to consult a licensed professional or ethics review board.
  • If key inputs are missing, ask for them before proceeding.

Example: {{program_name}}: Youth Employment Subsidy; {{policy_domain}}: Labour Market; {{evaluation_purpose}}: Assess uptake and job retention after 12 months; {{stakeholders}}: Ministry of Labour, employers, youth; {{timeframe}}: 3 years; {{data_sources}}: Administrative records, surveys.