On this page
Record the population before selecting
Enter the population definition, unit identifier and frame date. Describe excluded or missing units. The example row uses a fictional quarterly case study; its empty fields invite a specific design rather than silently recommending one.
Make the randomization reproducible
Record the selection rule, stratum or cluster definitions and any within-group stages. If a computer generated the draw, retain a reference to the procedure and its random seed in the private project record. The CSV does not perform a selection.
Track losses between stages
| Fields | Purpose |
|---|---|
| Target, unit and frame date | Define eligible units |
| Frame exclusions and selection rule | Identify coverage and mechanism |
| Groups and probabilities | Describe design structure |
| Selected, responding and usable | Track post-selection losses |
| Estimator and limitations | Retain weights and conclusion scope |
Update the record when the plan changes
A new filter or response treatment may change the estimate’s target. Retain the original and revised definitions rather than overwriting the reason for the change. The planning sheet documents the study; it does not certify randomization or representative coverage.