"Won't come back from Dead Man's Curve"

Especially if you're listening to "Dead Man's Curve" at the time...

April 11 Update: For those of my readers who thought I was making this all up, I am vindicated: see the New York Times article of April 10: “Is a Big Album Dropping? You Might Want to Watch the Road.” Car Charts readers, you read it here first, in February!

(You may want to listen to the song itself while reading this. Just don’t drive while you do!)

With thanks to Marginal Revolution for flagging this, hot off the press we have Smartphone, Online Music Streaming, and Traffic Fatalities, by Patel, Worsham, et al., over at NBER. (As always, read the underlying paper to figure out how I managed to misrepresent the authors’ findings!)

Taking it from the top: we all know that distracted driving is a bad thing. If a driver is arguing with a passenger, or fiddling with the seat controls, or sipping coffee, less attention is paid to the road, and at 65 MPH during 1 second of inattention you’ve covered about 95 feet without noticing, and sooner or later this leads to collisions, injuries, and death.

We also know that people do listen to the radio, streaming services, etc., while in the car.

The authors of this work have put these two things together, very cleverly mapping vehicle fatalities onto the release dates of major pop albums in the US.

And the results are as you might expect not good:

Using event study analysis, we show that music streaming – an indicator for smartphone use, where streaming most often occurs – sharply increases, by nearly 40%, on dates of major music album releases, while U.S. traffic fatalities increase by nearly 15% on those same days.

Of course there is a chart:

The uplift between the gray and the red bars is the cited 15%.

Note: [ the figure ] shows the adjusted mean number of traffic fatalities on album release days and on non-album release days (the 10 days before and after each album release day). Estimates are from models which adjust for federal holidays, day of the week, week of the year, and year fixed effects. Error bars indicate 95% confidence intervals.

This works out to about 18 excess deaths per popular album release.

Here’s the list of the albums studied:

And as you can imagine, the authors put a lot of effort into adjusting for confounding variables (e.g. was drinking involved?) and they feel their answers are robust: excitement around a major new record drop alters driver attention enough to lead to a heightened number of crashes.

Because death and injury are involved here, I will refrain from any further wisecracks. But the lesson I take away from this is to keep your eyes and ears on the road, people: you are in control of heavy machinery moving very quickly, F = ma no matter what, and even a great new jam is not worth a lapse in focus.