The data is from police reports where there was an accident. Do you think that it's underreported, and that people would not volunteer to an officer they were u

The data is from police reports where there was an accident. Do you think that it's underreported, and that people would not volunteer to an officer they were using their cell phones? Your first link talks about this in the Data Limitation section. Also, and they don't note this, but people can be intoxicated and using a mobile device. I suspect their model substantially unreports DUIs where distraction was a factor. It is very easy to do a field test and chalk it up to a DUI and not investigate further. "Honestly, I think the real number of fatalities tied to cell phones is at least three times the federal figure " - https://www.insurancejournal.com/news/national/2017/10/18/467836.htm https://www.insurancejournal.com/news/national/2017/10/18/46...