Field guide
Report missing follow-up without hiding the denominator
A result from the people who responded is not automatically a result for everyone who enrolled. Keep the response count, the original group and the wording of the question together.
Keep the groups visible
Suppose a fictional training pilot enrolled 20 people. Twelve answered a follow-up question, and nine of those twelve reported using a skill from the course. The defensible description names all three counts: 20 enrolled; 12 responded; nine respondents reported using the skill. Saying only that 75% benefited removes the missing eight people and changes a specific self-report into a broader outcome claim.
The numbers in this guide are invented to explain reporting practice; they are not research findings or a recommended sample size. The nine reports represent 75% of the 12 respondents and 45% of the 20 enrolled participants. Neither percentage tells us what happened to the eight people who did not answer. Choosing the larger percentage for a headline does not resolve that gap.
Define the denominator for each measure
Enrolment, attendance, completion and follow-up response answer different questions. Write a definition beside each field. An enrolled person who attended one session is not automatically a course completer. A person who completed the course but did not answer follow-up remains in the completion count and outside the response count. Do not quietly remove them from the original group.
Use a small table with the measure, count, denominator, collection date and source. For the fictional pilot, the skill-use item has nine positive responses out of 12 respondents, collected at the stated follow-up point. Record that the answer was self-reported. If a second question has only ten answers, use ten as that question's response denominator rather than copying twelve across the report.
Explain missing information without assigning an outcome
Record a known operational reason for missing data when it is relevant and appropriate to retain. A failed delivery address is an observation about contact, not proof that the service did not help. Equally, silence is not a positive outcome. Keep unknown results unknown. Do not substitute a team member's impression for a participant's answer just to make the table complete.
Review whether the follow-up process reached the intended groups. A method that relies on a particular device or language can leave some participants out. Document the recruitment and contact process alongside the limitations. Any additional contact should follow the permissions agreed with participants. A reporting deadline does not justify collecting extra personal information without a defined purpose.
Write the finding and its limit in the same paragraph
A clear fictional report reads: Of 20 enrolled participants, 12 answered the follow-up question; nine respondents said they had used a course skill. Outcomes for the eight non-respondents were not established. The result is self-reported and does not show that the course caused the change. This wording gives a reader enough context to interpret the observation without making them reconstruct the missing group.
Keep the question wording and follow-up timing available with the report. Separate this descriptive finding from decisions about improving the service. The team can use a partial response to identify a question for its next trial while remaining explicit about what it has not measured. If a funding agreement requires a different indicator, verify its definition rather than renaming this result to fit.