Six NSW sites
Each speed-zone change is compared with control roads using weighted difference-in-differences, with a rolling coefficient of variation to track speed spread, not just the mean.
Path-level vs aggregate
Aggregated segment speeds mix vehicles that drive different paths. Comparing the aggregate estimate with path-level trajectories exposes a bias (aggregate minus trajectory) that can make a change look larger or smaller than it is.
Instrumented cars
Connected-vehicle speeds are compared with two instrumented SafeFuture research cars on the same roads.
What we will not claim
This note presents the method. Site-level effect sizes are not presented here as findings.