When reporting numerical data, what should be documented?

When reporting numerical data, what should be documented?

• A. Only the final numbers without any source notes.
• B. Rely on memory and not document methodology.
• C. Document the methodology and round numbers consistently per policy.
• D. Publish the primary data without any explanation.

Answer: (C)

Accurate numerical reporting hinges on documenting how numbers were produced and how they’re presented. This means recording the data sources, the steps used to derive the figures, any data-cleaning or transformations, the assumptions made, and the units and time frame involved. It also means stating how measurements were obtained and how any adjustments or exclusions were handled. Importantly, rounding rules should be applied consistently and clearly described so readers know the precision level and can compare figures reliably. This combination—methodology documentation plus consistent rounding per policy—ensures transparency, reproducibility, and trust. Without documenting the process, readers can’t verify results or understand potential limitations. Relying on memory is prone to error, and publishing raw data without context leaves essential metadata and interpretation unstated.

Accurate numerical reporting hinges on documenting how numbers were produced and how they’re presented. This means recording the data sources, the steps used to derive the figures, any data-cleaning or transformations, the assumptions made, and the units and time frame involved. It also means stating how measurements were obtained and how any adjustments or exclusions were handled. Importantly, rounding rules should be applied consistently and clearly described so readers know the precision level and can compare figures reliably.

This combination—methodology documentation plus consistent rounding per policy—ensures transparency, reproducibility, and trust. Without documenting the process, readers can’t verify results or understand potential limitations. Relying on memory is prone to error, and publishing raw data without context leaves essential metadata and interpretation unstated.