Tuesday, 24 October 2017

The cell phones hypothesis

Are cell phones to blame for the uptick in New Zealand's crash rates? I don't know, but I have a potential way of checking - an exercise I'll leave to the committed reader, or whoever MoT hires as their new data analyst.

Get maps of cell phone coverage going back as far as the data goes. Not all uses of cell phones while driving require signal, but phone use should be increasing in signal availability. Here's a map of towers, but you'll need a time series and you'll need dead zones.

Get maps of accidents.

See whether expansions of cell phone coverage predict increases in the number of accidents in the places getting more cell coverage.

Confound: endogeneity where cell companies will put up towers in response to demand. But that demand should lead the accident rate rather than lag it, unless the cell companies are real good at predicting where people are going to start wanting to drive more. If the towers are more likely to come after the increase in traffic, then it should be fine.

No comments:

Post a comment