Pokémon GO may have waned in popularity since its explosive release in July of 2016, but the groundbreaking augmented reality (AR) game that’s now been downloaded more than 500 million times still stands as a great opportunity to do some interesting incidental research. One recent study from researchers at Purdue University, for example, unambiguously titled “Death by Pokémon GO,” claims that the wildly successful mobile game may have been responsible for up to 145,632 vehicular accidents, and up to $7.3 billion in damages across the U.S. over the first 148 days after its release. Who knew “Gotta catch ’em all” could also be applied to insurance claims?

The yet-to-be-peer-reviewed study, which comes via Gizmodo, was conducted by Mara Faccio and John J. McConnell of Purdue University, and focused on accident reports filed with the police in Tippecanoe County, Indiana. The authors frame the research with the fact that a steady 25-year decline in vehicular fatalities reversed course in the U.S. in 2011, which was almost certainly due to smart phone app usage. Hence the desire to study one of the most popular apps and its effects on driving.

Faccio and McConnell examined 12,000 police accident reports from the county from March 1, 2015, through November 30, 2016, and found that there was an increase of 286 crashes across the county in the 148 days after the game’s release on July 6, with 134 of those occurring near PokéStops. (PokéStops being places, usually landmarks or other popular sites, where players go to collect in-game items like Poké Balls.) Using these figures, it’s estimated that 47% of the increase in crashes in Tippecanoe County in those 148 days was due to playing Pokémon GO near a PokéStop while driving.

That claim raises two questions. One: How can we be certain that those 134 accidents that happened near PokéStops were due to drivers playing Pokémon GO while driving? And two, how can countrywide figures be extrapolated from conclusions from a study in a single county in Indiana?

As for the first question, most of the paper works hard to control for any variables that may have caused the increase in accidents other than drivers playing Pokémon GO. For example, to invalidate the idea that perhaps PokéStops simply drew more traffic, and hence had more accidents, the researchers compared the accident rate at PokéStops with the accident rate at Pokémon GO Gyms, which are other popular game-related destinations that drew lots of traffic. It was found that accidents were less likely to happen in the vicinity of Gyms, where it was “basically impossible for a player to play the game while driving,” and that there was “a significantly greater increase in the number of crashes in the vicinity of PokéStops…”

Other factors, such as Tippecanoe’s population fluctuating significantly due to Purdue’s student body leaving and coming back for breaks from school, were also controlled for, and deemed to be non-contributors to the increase in accident rates near PokéStops.

As for how the study justifies its speculation of Pokémon GO’s effect on a countrywide scale, the authors make clear that the claim they’re making in this regard is “speculative.” Essentially the authors say that if Pokémon GO was responsible for 47% of the increase in traffic accidents in Tippecanoe, it may have been responsible for 47% of the increase in traffic accidents across the U.S. in the same time period, thusly arriving at that 145,632 number for accidents. The $2 billion to $7.3 billion in damages on a countrywide scale due to Pokémon GO is also scaled up from the numbers in Tippecanoe, which range from $5.2 million to $25.5 million (that number varies so significantly because it’s uncertain whether Pokémon GO was responsible for two driving-related deaths, which were given monetary values).

As for the study’s validity in Tippecanoe, there seems to have been due diligence in controlling for variables, and honing in on driving and playing Pokémon GO as a real cause of the increase in traffic accidents. Although it’s not exactly clear why the police reports that were analyzed didn’t say the exact cause of accident (i.e. “driver was playing Pokémon GO“) rather than simply “Driver Distracted” or “Cell Phone Usage,” which were apparently the extent of the descriptions provided in the police reports. Plus, there’s been plenty of anecdotal evidence to back up the claim that driving while playing Pokémon GO is (so obviously) a horrible idea.

The countrywide speculation should probably be taken with a Snorlax-sized grain of salt however, as it seems that even doing one other study like this in one other county could come to different conclusions and thusly throw a wrench in the idea of scaling up Tippecanoe’s results to stand in for the entire country.

Either way, you should never ever Pokémon GO and drive, or Harry Potter and drive for that matter. ‘Cause Niantic is coming out with a Harry Potter AR game next and as much as more science would be cool, the magic of not crashing into each other would be even cooler.

What do you think about these findings regarding Pokémon GO and driving? Do you think trends in Tippecanoe can stand in for the entire country? Give us your thoughts in the comments below!

Images: Paintimpact, Eduardo Woo, Health Gauge

More Pokémon news

• Here’s what Pokémon would look like in real life
• Watch the trailer for the newest Pokémon movie
• There’s an adorable Meowth plush on the market