A house that demands evidence needs to declare its own limits. This is our list, valid for all four proving grounds.
We measure
- What the person did when facing the case: decision, argument, leadership under pushback.
- What stayed after the AI left the picture — retention, transfer and autonomy, by the LUMEN standard.
- Progress against one's own baseline.
We don't measure
- Who the person is: no profile, no type, no diagnosis.
- Future performance: no prediction.
- A verdict by comparison with peers.
Why declare it
Because a declared limit is what makes feedback usable. A strength tied to what the practice demanded can be carried into the next meeting; a reflection presented as a hypothesis can be tested in real life. A label cannot. And because that is how we hope to be read: with the source beside the number — and the caveat beside the conclusion.
Some of the data we use is self-reported by those who produced it; we say so when it is. Some calculations are our own; the method is available on request. The LUMEN standard is under validation; we say that too.
Feel free to quote — with the source.