# Randomness First: Protocol for a Baby-Name Null Model

## Question
Does observed name concentration exceed what should be expected from the number of births and a realistically unequal distribution of name preferences?

## Sequence
1. Establish annual births and sex-specific denominators.
2. Measure concentration, turnover, rank persistence and tail size.
3. Fit a fixed unequal-probability null model.
4. Test time-varying but independent probabilities.
5. Add imitation or shared shocks only when simpler models fail out of sample.
6. Compare many names, not only Olivia.
7. Report which mechanisms remain non-identifiable.

## Stopping rule
If a simpler model accounts for the feature under investigation within pre-specified tolerance, do not infer a more elaborate cultural mechanism from that feature.