>focusing on 246 deceased drivers who were tested for THC following a fatal crash”. That means there could be selection bias at play.

that wording definitely sets off warning alarms for selection bias. but it looks like there were approximately 350 traffic deaths in montgomery county during that period [1]. that probably about lines up with 246 drivers dying during that period, so it seems likely they tested all or almost all deceased drivers.

[1] https://dam.assets.ohio.gov/image/upload/statepatrol.ohio.go...