How to Lie With Jaywalking Stats

Ever since video of a Seattle cop punching a 17-year-old girl in the face surfaced last Tuesday, the city has been debating the relative merits of a relatively victimless crime: jaywalking. Thanks to reporting by Laura Onstot, we now know that the Rainer Avenue-Martin Luther King Way intersection where the incident took place hasn't had a pedestrian collision in five years. The SDOT study Onstot referenced showed that 74 people were struck on an eight-mile stretch of Rainier between 2002 and 2004. The agency doesn't say how many of these unfortunates were crossing against the light. But even if it did, there's good reason to distrust the agency's claims.

Any fan of HBO's "The Wire" is familiar with the term "juke the stats," a reference to the way institutions of all stripes shape statistics until they're molded into a form that confirms what they want the public to believe. In his invaluable piece for Slate, traffic guru Tom Vanderbilt shows that jaywalking stats aren't immune to creative sculpting.

For one thing, most jaywalkers who get hurt are, in fact, drunk off their ass -- "as the National Highway Traffic Safety Administration notes, 'about 25 percent of fatally injured pedestrians have a BAC greater than .20'" -- a condition for which there is already another law on the books.

For another, sometimes what is marked down in a police blotter as jaywalking is in fact something else entirely:

Consider, for example, this case of a driver who killed a pedestrian said to have been crossing outside the crosswalk. The driver was drunk and traveling at least 60 mph on a street whose limit was 30 mph. Statistically (and more-or-less legally), this enters the book as a "jaywalking" fatality, but it was predicated not merely on an illegal crossing but the active contribution of a driver whose reaction time was compromised--and who was traveling at a speed that made the pedestrian's death much more likely.
comments powered by Disqus

Friends to Follow