The Scribe’s Advantage

Data

The data behind the book

Every dataset the book rests on or takes a number from, as the Companion uses it: 31 datasets, 68 tables. Each one downloads as CSV, one file per table, and as JSON. The files are the ones the pages draw, so a number on a page and a number in its download are the same number.

How to read these files

CSV files are UTF-8 and open in any spreadsheet. Nested fields are flattened into dotted column names (for example axes.stability); a list of values sits in one cell, separated by |. The JSON file is the dataset exactly as the site reads it.

Scores are judgments made against written rubrics, not measurements. The rubrics, the calibration and the argument for each scale are on Methodology and Method and Sources. Where a cell was researched, its justification and source travel with it in the file.