A new study from Purdue University uses detailed local traffic accident reports to suggest that Pokemon Go caused a marked increase in vehicle damages, injuries, and even deaths due to people playing the game while driving. Ars Technica reports: In the provocatively titled "Death by Pokemon Go" (which has been shared online but has yet to be peer-reviewed), Purdue professors Mara Faccio and John J. McConnell studied nearly 12,000 accident reports in Tippecanoe County, Indiana, in the months before and after Pokemon Go's July 6, 2016 launch. The authors then cross-referenced those reports with the locations of Pokestops in the county (where players visit frequently to obtain necessary in-game items) to determine whether the introduction of a Pokestop correlated with an increase in accident frequency, relative to intersections that didn't have them. While the incidence of traffic accidents increased across the county after Pokemon Go's introduction, that increase was a statistically significant 26.5 percent greater at intersections within 100 meters of a Pokestop, compared to those farther away. All told, across the county, the authors estimate 134 extra accidents occurred near Pokestops in the 148-day period immediately after the game came out, compared to the baseline where those Pokestops didn't exist. That adds up to nearly $500,000 in vehicle damage, 31 additional injuries, and two additional deaths across the county, based on extrapolation from the accident reports. The study uses a regression model to account for potential confounding variables like school breaks and inclement weather, which could cause variation separate from Pokemon Go. The model also compares Pokestops to Pokegyms (where it was nearly impossible to play while driving) to account for the possibility that generally increased traffic to Pokemon Go locations was leading to more accidents, even among drivers who stopped and parked before playing. In all cases, though, being able to compare to intersections without a Pokestop and to the same dates the year before, helped provide natural control variables for the study.
Read more of this story at Slashdot.