Education research
How to report attrition and missing follow-up scores in education studies
Start with the original assignment list. By arm, report how many students have a valid follow-up, how many lack one, and how many are analyzed. Separate absences, timing violations, and invalid observations.
Use assignment records, planned and actual assessment dates, validity rules, missingness reasons, and the final analysis table. For delayed tests, record makeup timing so reviewers can judge comparability.
A result can have several denominators. Label whether each claim refers to all assigned students, students with valid outcomes, or the analysis sample.
How to do it
Draw a flow for each arm
Count participants assigned, exposed, assessed in-window, valid, and analyzed. Record exclusion reasons. If a student has multiple records, state the selection rule.
Calculate overall and differential missingness
Use the assigned count in each arm as its denominator. Report percentage-point differences across arms. Compare observed baseline characteristics of students with and without results where appropriate.
Check the original plan
Document the prespecified window and missing-data handling. Date decisions that changed after planning. Describe makeup sessions, imputation, or sensitivity analyses and their assumptions.
Reconcile the abstract and results
Put the actual denominator in the table and attach counts to proportions in the abstract. Explain which sample each result describes and where missingness limits interpretation.
Worked example
In this synthetic example, the prespecified primary outcome is measured on days 7–9. Missingness uses the original assigned counts.
| Record | Arm A | Arm B |
|---|---|---|
| Assigned | 60 | 60 |
| Valid in-window outcome | 54 | 45 |
| Missing in-window outcome | 6/60 = 10% | 15/60 = 25% |
| Reason | 4 absent, 2 out of window | 9 absent, 6 out of window |
Missingness differs by 15 percentage points. The table audits denominators. Inference also depends on reasons, clustering, baseline data, and the planned analysis.
Common questions
Can I record absent students as scoring zero?
Record an unobserved assessment separately unless zero is itself a prespecified, observable outcome. Follow the study plan for missing observations.
Can a makeup test replace a missing result?
Include it under the planned rule if timing and measurement conditions qualify. Keep the actual date and handle out-of-window scores as specified.
Verified sources
The official sources below were checked on 2026-11-13. Access conditions and update dates are rechecked before publication.
RigorMed workspace
Audit my follow-up counts
Bring authorized assignment and assessment summaries. Prepare an arm-by-arm flow table, abstract consistency checks, and questions about missing-data handling.
