How the record gets made.

The inclusion boundary is the single most consequential decision in an event dataset: it determines what the record can and cannot be used to argue.

The research team’s definition of a codable event goes here, with the worked examples and edge cases that show where the boundary falls — what is in, what is deliberately out, and why.

Every event is placed in exactly one of 33 categories, which roll up to 9 themes. The taxonomy is what makes the record comparable across states and across years.

The decision rules coders apply, how ties between neighbouring categories are broken, and what a coder does when an event fits none of them.

• The nine themes
• The category guides

A hand-coded dataset is only as good as its agreement between coders, and that number belongs in public alongside the data.

The double-coding procedure, the inter-coder reliability statistics, and how disagreements were adjudicated and fed back into the codebook.

The full codebook publishes with the dataset, so anyone can check a coding decision or extend the scheme to their own sources.

A link to the released codebook, its version history, and the note on how to cite it.