How to write an evaluation plan: purpose, questions, indicators and data collection

Include the background, purpose, users, program overview, evaluation questions, judgment criteria, indicators, data collection and analysis methods, arrangements, schedule, ethical considerations and planned use of results.

Link evaluation questions, judgment criteria, indicators and collection methods. One question may require multiple indicators, and one source may support multiple questions. Traceability of the grounds for judgment matters more than a one-to-one relationship.

• Evaluation purpose and intended users
• Program overview and logic model
• Evaluation questions and judgment criteria
• Indicators and data collection methods
• Analysis and reporting methods
• Schedule, arrangements and ethics

Too many questions for the evaluation purpose can make it impossible to collect the evidence needed. Effects, reach and improvement can be addressed together, but estimate the design and burden for each question and agree priorities.

Confirm who will use the results for which decisions and prioritize questions needed for those decisions. For example, if deciding whether to continue, effectiveness, reach and value for cost become important.

1. Identify users of evaluation results
2. Confirm decisions to be made after evaluation
3. Identify questions needed for those decisions
4. Select main evaluation questions according to intended use and collection burden
5. Set judgment criteria for each question

Indicator names alone are insufficient in an evaluation plan. Specify from whom, when, how and approximately how many records will be collected, and whether existing data or new research will be used.

Decide ethical arrangements at planning, including personal information, consent, audio and video recording, anonymization and retention periods. Evaluation credibility depends on data management as well as methods.

Use the evaluation plan as a document of stakeholder agreement rather than treating completion as the end. Check that program staff and managers share an understanding of purpose, questions, data provision burden and reporting timing.

When external evaluators are involved, design must protect both evaluation independence and frontline learning. Agree participant contact arrangements, data access rights, interim reporting and the scope of final-report publication beforehand to avoid later friction.

• Document who will use results and for what
• Check data provision burden and deadlines
• Set procedures for changes during implementation
• Agree publication scope and anonymization policy