Pattern
A spreadsheet analysis can be a companion repository
Many published analyses were done in a spreadsheet, not in code. That's fine: a spreadsheet with its formulas intact, its data in open formats and a short README that says the order of the steps is a companion repository too.
Formulas are code, so keep them in the file
Claim: recommendation
A spreadsheet's formulas are the analysis. Pasting values over them, or sharing only a PDF of the final table, throws the method away. Share the workbook itself, in its native format and in an open one such as OpenDocument Spreadsheet, with every formula in place.
Put the data in open, plain formats a reader can open without the original software: CSV or TSV (values separated by commas or tabs), JSON or JSON Lines, XML, or Parquet for large tables. Then write a short README that says which sheet does what, in what order, and which cell or range produces each number in the paper. That README is the spreadsheet's entry point and declared outputs.
The same six parts apply. A spreadsheet just holds the code and the environment in one file.
What would show this claim is wrong
- Reviewers given a spreadsheet with a README can't reproduce its reported numbers any more often than reviewers given the paper alone.
- Spreadsheet results depend on software versions, such as how a function rounds, often enough that the file alone can't reproduce them.