More on US Pedestrian Deaths

Last week I wrote about the unexplained rise in pedestrians killed by motor vehicles in the US. Since 2009, pedestrian deaths have increased by almost 80%, following multiple decades of falling rates.

I looked through data on pedestrian fatalities from the National Highway Transportation Safety Administration (NHTSA), but I couldn’t find a singular explanation that satisfied me. The piece of evidence that most compelled me was this: the rate of pedestrian fatalities has increased substantially across many different states, even as overall pedestrian collisions have stayed close to constant. This suggests that pedestrian accidents are getting more deadly, which in turn suggests that the increase in popularity of big SUVs might be causing the increase in deaths. However, the number of pedestrians killed by sedans seems to be rising as well. That can’t be caused by SUVs getting bigger and more popular.

This article got lots of attention, and readers had many theories about the drivers of the increase. In this follow-up post, I wanted to further investigate some of the most compelling theories and angles on the question.

Unfortunately, this exercise left me more confused than before. What previously seemed like the most promising explanation (the rise of large SUVs) now looks less compelling. But no other explanation has emerged to take its place.

Deaths and darkness

Some folks pointed me to a New York Times article on the problem of US pedestrian deaths from 2023. Contra my original claims (that the increase in deaths doesn’t seem specific to one time of day), the NYT states that nearly all the increase in deaths is at night.

The NYT is correct; This was an error on my part that stemmed from not looking closely enough at the data behind a bad summary statistic. I was correct that there hasn’t been much change in the proportion of deaths during the hours of 6pm to 6am. But 6am is night at some times of year and location, and day in other times and places.. Lighting conditions are not just a function of the clock. If you look at deaths by lighting condition, you find that the increase in deaths is mostly in times of darkness.

Deaths in daylight are only up 28% since 2009; deaths in darkness or lit darkness (streetlights, etc.) are up 103% and 87%, respectively. On its own, this data doesn’t tell us what’s causing the increase in deaths ...