Vaccine policy changes and autism outcomes Results tables

The PDF and HTML provide the finished interpretive summary. CSV files retain
complete records for reuse. All104 event-study pools contain both raw percent
changes and uptake-scaled ratios, explicitly labeled; those units must not be
mixed. The44 native reference families and446 alternatives reuse contrasts and
studies. Archived local reference families have their own memberships and
denominators; they do not replace the14-state manuscript estimate. Tables of
country deletions, holdouts, dependence scenarios and service-use models are
overlapping sensitivity checks, not new independent observations. Ratios, hazard
ratios, odds ratios, rate ratios, native service-use units and risk differences
remain distinct. Source bindings preserve the exact input-file hashes.
