A method someone else can follow

A finished chart can make an investigation look straightforward. The path to it may have involved several files, a unit conversion, a decision about missing values, and a calculation that changed during the work. A useful method makes that path visible.

Write for someone who was not in the room

Imagine returning to a spreadsheet six months after making it. A column labelled “adjusted” may have been obvious at the time. Now it raises questions: adjusted for what, using which rule, and before or after the rows were filtered?

The reader who needs those answers might be a colleague, a reviewer, or your future self. Start by identifying the question, the inputs, and the output. Then describe the steps connecting them in an order that can be followed without private knowledge of the project.

A file name is useful only when the file it identifies is unambiguous. A note such as “use the latest table” becomes less helpful each time the table changes. Recording the specific version gives the instruction a stable meaning.

Preserve the choices between input and result

USGS guidance treats documentation as something that accompanies data through collection, processing, and analysis. Its guidance on processing also emphasizes recording the changes made to data. That distinction is practical: describing the final calculation does not explain how the starting material became ready for it.

For the imagined spreadsheet, a compact record could include:

The source and version of the input files, with the original files kept separately from working copies.
The meaning and units of each column, including codes for missing or unavailable values.
The rules for correcting, excluding, or combining observations, and the reasons for those rules.
The calculations and processing order, with software versions or settings that affect the outcome.
The expected output and a simple check that would reveal a failed step.

These details do not require a complicated system. A small project may need a clearly named folder and a short methods file. The structure should make the important decisions easy to find, review, and update.

Distinguish running again from testing again

Terms vary across fields. In its 2019 report, the National Academies uses reproducibility for obtaining consistent computational results from the same data and analytical steps. It uses replicability for obtaining consistent results in studies that collect new data to address the same scientific question.

Those are different checks. Being able to regenerate a chart helps establish how it was produced. It does not, by itself, show that the measurements were suitable or that the explanation will hold under new conditions. A faithful calculation can preserve a mistaken assumption.

Equally, a difference in a later investigation deserves examination. It may point to a changed condition, an overlooked detail, or a limitation in the original explanation. A clear method gives that examination somewhere to begin.

Try the handoff before it matters

Ask someone to follow the record using an appropriate copy of the materials. Notice where they need to ask a question or make an undocumented choice. Each pause identifies a place where the method can become clearer.

Where information cannot be shared, describe the restriction and the access conditions. A useful record can explain what is needed without exposing private material. The goal is a traceable account of the work and an honest statement of what another person can check.