Method

How this was built — and what wasn't done

The short version: the construction was careful, and the validation hasn't happened. Both halves matter.

What was done

• Constructs defined first. Each dimension has a written definition with explicit out-of-scope boundaries and pre-decided edge cases, so items couldn't drift toward whatever sounded good.
• A blueprint before writing. Each scale was allocated a fixed number of items per facet and per severity level — mild, moderate and extreme — so the test discriminates across the range rather than only in the middle.
• 194 items written, 65 kept. Roughly three times what was needed, so the selection could cut anything carrying a known confound.
• Recorded confounds. Items that might track culture, religion, income or circumstance rather than the construct are flagged in the source data with the reasoning attached.
• Cost and upside for every dimension. Enforced by a build check: a scale missing either half fails the build.

What wasn't done

• No independent review. Nobody but the author checked whether the items measure what the definitions say.
• No cognitive interviews. No item has been read aloud to a stranger to find out what they think it's asking.
• No pilot study. There are no reliability figures, no factor analysis, and no evidence the nine dimensions separate as intended.
• No norms. Which is why there are no percentiles anywhere in the results.
• No differential item functioning testing. Items that may behave differently for religious respondents, or across cultures, haven't been checked. The most affected dimension is deliberately withheld from the report entirely.

What would change that

Everything above is recoverable from ordinary use: reliability figures, factor structure and item performance all fall out of enough responses. If a contribution option is added later, it will be explicit, optional, identifier-free, and reversible with a deletion code.