On this page
Record the plan before analysis
Write the population, sampling unit, null claim, alternative direction and data window. Include planned filters and the number of hypotheses in the family. A log completed only for successful tests conceals the search that produced them.
Keep estimate and uncertainty together
| Fields | Purpose |
|---|---|
| Test ID and hypothesis | Identify the question |
| Alternative and data window | Retain planned direction and sample |
| Family size and adjustment | Describe multiple-testing treatment |
| Estimate, SE, statistic and p-value | Keep magnitude and uncertainty visible |
| Assumptions and deviations | Record conditions and changes |
Use the fictional row as a format example
The CSV includes a demonstration of the known-sigma mean exercise and an empty row for an independent record. Its example assumptions are synthetic. Replace them with your actual study definitions instead of treating the row as an approved investment procedure.
Use a log alongside the underlying evidence
The CSV documents a calculation; it does not validate the sampling design, assumptions or software. Preserve the underlying data and analysis instructions in an appropriate private research location. A test log alone is insufficient to recreate the study.