"Data is neutral.
I am not."
Every number got there by a choice. We make those choices.
"Data distorts." "Data lies." "Data aggregates." "Data reveals." In fact, data doesn't do any of those things. We do these things with data. How could data lie? Numbers have no intentions of their own.
But neutral is not the same as innocent. Every number in a dataset got there by a choice — what to measure, how to categorize, whom to count, what to discard as noise. The spreadsheet didn't make them. You and I made them.
There is no neutral place to hide — not in the number, not in the method, not in the model.
A codebook is where those choices live. Written down, shareable, inheritable. Every decision legible — to you, to your reviewers, to whoever picks up the work next.
The work is ours, all the way down.