Template and Outline
Section-by-Section Writing Guide
The template follows the order most commissioners expect. This guide says what each section needs and a rough length for a mid-sized evaluation (about 45 pages of main text). For what an evaluation report is and why it matters, see the evaluation report entry. The report should deliver what the inception report promised, so keep the two aligned.
| Section | Length | What it must do | Watch for |
|---|---|---|---|
| Executive Summary | 2 to 3 pages | Let a busy reader act: what, why, how, main findings, conclusions, recommendations | Numbers that differ from the body |
| Introduction | 2 pages | Purpose, users, decision, scope and what is left out | A purpose that sounds like "to evaluate the program" |
| Program Description and Context | 4 pages | Describe the program and only the context that helps judge the findings | Repeating the proposal |
| Evaluation Questions and Criteria | 2 pages | The final questions under their criteria, with any change from the Terms explained | Questions that quietly shifted |
| Methodology and Limitations | 5 pages | Design, sources, samples planned and achieved, analysis, ethics, limits | Reporting the plan, not what happened |
| Findings | 18 pages | By question: finding, evidence, strength of evidence | New judgments creeping in |
| Conclusions | 3 pages | What the findings add up to, against the criteria | New evidence appearing here |
| Recommendations | 3 pages | What to do, who, by when, in priority order | Advice with no owner |
| Lessons Learned | 2 pages | Lessons for other programs, separate from recommendations | A repeat of the recommendations |
The title page, table of contents and list of acronyms come before the summary. Annexes (20 to 60 pages) usually hold the Terms of Reference, the evaluation matrix, tools, lists of people and documents, data tables and team biographies.
How the Template Helps You Keep the Chain Intact
- The Findings section is a table with four columns: question, finding, evidence and strength of evidence. It stops you from stating a finding without saying what supports it.
- The Conclusions table has a column for the findings each conclusion rests on, and a column for the judgment against the criterion.
- The Recommendations table names the conclusion each one follows from, who it is addressed to, its priority and a timing.
For the difference between the three kinds of statement, see findings vs conclusions vs recommendations.
Filled Example
Filled Example: One Chain from Evidence to Recommendation
This extract is from a made-up final evaluation of a community water program. It follows one question through the three tables.
Finding (effectiveness). 82 percent of pumps were working when visited (167 of 204). Households using them spend less time collecting water: the median round trip fell from 70 minutes at baseline to 25 minutes now. For the poorest households the fall is smaller, from 85 to 45 minutes.
| Evidence | Strength |
|---|---|
| Inspection records for 204 pumps; spot check of 30 pumps in person (28 matched); household survey of 571 households | Strong for pump function: three sources agree. Moderate for time saved: no comparison group, but focus groups with women describe the same change |
Conclusion. The program has largely achieved its main purpose: most pumps work, and households use them and save time. The gain is smaller for the poorest households. Judgment: effectiveness is met, with a gap for the poorest households.
Recommendation. Review the locations of pumps in the five villages where collection time is still over 45 minutes, and consider added pumps. Addressed to the program team and the district water office. Priority: medium. Timing: the next planning cycle.
Now the sustainability link, which carries a stronger message:
| Finding | Conclusion | Recommendation |
|---|---|---|
| Only 31 percent of committees hold savings above the cost of a typical major repair. Spare parts are available only in the district town, about 3 hours away. | Gains are at risk because repair funding and spare parts are not secure. Sustainability is partly met and likely to weaken without action. | Agree with the district water office to hold a standard stock of common spare parts at sub-county level (high priority, within 6 months). Set up a shared repair fund per sub-county, with committee contributions agreed with committees (high priority, within 9 months). |
Check three things in the extract. Each finding carries numbers. Each conclusion is a judgment, in words the criteria use. Each recommendation names an owner and a time, and a reader can tell which finding made it necessary. If you cannot trace a recommendation back, either a finding is missing or the recommendation does not belong.
Check Before Sending
Check Before You Send
- The summary matches the body: same numbers, same names, same dates.
- Every evaluation question in the matrix has a finding, or the report says why not.
- Planned and achieved samples are both reported.
- Findings are separated from conclusions. Conclusions add no new evidence.
- Each recommendation names who should act and when.
- Limitations say what each one does to the findings.
- Quotes and case details follow the consent given and do not identify anyone who asked not to be named.
- Disaggregated findings (sex, age, location, disability) appear where the data allows, and gaps in the data are noted.
- The report follows the commissioner's format and quality checklist, such as the UNEG checklist for evaluation reports, where one applies.
When the report is final, the commissioner writes a management response and plans how to share the findings; see management response and dissemination.