How to write terms of reference for a final evaluation
What to put in evaluation terms of reference: purpose, scope, questions, method, team, deliverables and timeline, with a section-by-section checklist.
Short answer
Terms of reference (ToR) for an evaluation tell an evaluator why the evaluation is happening, what it must answer, how it will be done, who is responsible, what they must deliver and by when. Include background, purpose and users, scope, key questions, method, team qualifications, deliverables, timeline and budget. Keep it to a few pages and make the questions specific.
Key points
- The ToR is the basis of the contract and the yardstick for the evaluation's success.
- Say who will use the findings and for what decision.
- Write five to eight clear evaluation questions, not twenty.
- List deliverables with dates, including an inception report and a draft for comment.
- Plan three to six months for the whole process.
What are terms of reference for an evaluation?
Terms of reference, or ToR, describe an evaluation job. BetterEvaluation explains that a ToR gives an overview of what is expected, and for an external evaluation it provides the basis for a contract between the commissioner and the evaluator. It also sets the standard against which the evaluation can be judged.
A weak ToR leads to a weak evaluation. If the questions are vague, the report will be vague. If the users are not named, nobody uses the findings. Spend time on the ToR before you look for an evaluator.
What goes into the ToR?
Common elements, as listed by BetterEvaluation and the IFRC evaluation framework, fit into five groups. Use them as headings.
| Section | What to write |
|---|---|
| 1. Background | The project: goal, location, dates, partners, budget, main activities and progress to date |
| 2. Purpose and users | Why you are evaluating, who will use the findings and which decisions they will inform |
| 3. Scope | What is in and out: time period, locations, target groups, which parts of the project |
| 4. Evaluation questions | The questions to answer, often grouped by criteria such as relevance or effectiveness |
| 5. Method | Expected approach: data sources, sampling, how communities take part, ethics |
| 6. Team and roles | Skills needed, who manages the evaluation, who approves the report |
| 7. Deliverables and timeline | Outputs with deadlines |
| 8. Resources | Budget, existing data and documents |
How do you write each part? Eight steps
- Write the background in one page. Include the results framework or logframe and the monitoring data summary, or list where to find them.
- State the purpose and the users. Say why now, and what will change as a result. "To learn what worked so our next project in Rakhine is designed better" is clearer than "to assess impact".
- Set the scope. IFRC guidance suggests thinking about the unit of analysis, the time period, the geographic coverage and the target groups. Say what you will not cover.
- Choose five to eight questions. Link them to criteria. The OECD-DAC criteria are a common frame: relevance, coherence, effectiveness, efficiency, impact and sustainability. Do not use all of them mechanically; pick those that fit your purpose.
- Describe the method. IFRC guidance suggests covering what evidence is needed, what data collection methods will be used, who collects the data, how the sample is chosen, how participants take part, and how data will be managed and analysed. Add ethical needs: consent, confidentiality, and safeguarding.
- Define the team. List the skills needed, such as evaluation experience, sector knowledge, language skills and gender expertise. Name the person who manages the evaluation and who has to approve the deliverables.
- List deliverables and dates. Typical items: an inception report, a data collection plan, a debriefing workshop, a draft report, a final report, and any extra products such as a short summary in Burmese.
- Add the timeline and budget. Show key dates in a table. State the budget or the consultant days if your policy allows.
What deliverables should you ask for?
IFRC guidance describes common deliverables:
- Inception report. Shows the evaluator understands the ToR and has a realistic plan. For lighter reviews, a short work plan is enough.
- Debriefing or findings workshop. The evaluator presents early findings so stakeholders can check accuracy and add context.
- Draft report. Allow time for comments before the final version.
- Final report. At minimum: an executive summary, project background, methods and limits, findings, conclusions, lessons learned, recommendations and annexes with the ToR and data collection tools.
State the language, length and format. If the report must be read by partners who read Burmese, ask for a translated summary.
How long does an evaluation take?
IFRC guidance says a good evaluation can take three to six months, from designing the ToR, to hiring a consultant, to finalising the report. Plan backwards from the date you need the findings. If your grant ends in December, start the ToR in the middle of the year. Our guide on a midterm review covers a lighter, earlier check.
Common mistakes to avoid
- Too many questions. Twenty questions produce shallow answers.
- No users named. If nobody owns the findings, nobody acts.
- Copying a donor template without adapting it. The ToR should fit your project.
- Leaving out ethics. State consent, confidentiality, safeguarding and data protection needs.
- No management response. Plan how you will respond to recommendations after the report.
- Conflict of interest. The person managing the evaluation should not have a stake in the result. IFRC guidance expects evaluation managers not to have a vested interest in the operation being reviewed.
How EMPO can help
EMPO is a trusted training and consultancy provider based in Yangon and one of Myanmar's leading names in practical NGO training, with 1,200+ people trained and 15 partner organisations. We offer MEAL training and intensive support, in English and Burmese, online and in person across Myanmar. See our services or contact us to discuss your evaluation planning.
Sources
- Terms of reference — BetterEvaluation
- IFRC Framework for Evaluations 2024 — IFRC, 2024
Questions people ask
How long should terms of reference be?
Usually three to six pages plus annexes. Make it long enough that a stranger knows what to do, and short enough that people read it. Put long material, such as a report outline, in an annex.
Should the ToR state the budget?
It depends on your organisation's policy and your donor. Many ToR state the available budget or the number of consultant days, because it helps bidders propose a realistic method. Check your own rules.
Do we need to choose the method in the ToR?
You can give the main method or leave it to the evaluator to propose in an inception report. Either way, state any non-negotiable needs, such as consulting women and people with disabilities, or collecting data in Burmese and local languages.