We recently outlined the fact that Pokémon Go still has a thriving community, a year and a half after it first launched. But there might have been a dark side to the game’s unprecedented success.

According to a new study from professors at Purdue University, there was a 28 percent increase in traffic crashes near PokéStop locations in the approximately five months after the game first launched. The study, which focused on a single county in Indiana, also found that at least two of the crashes resulted in fatalities. When you extrapolate the study’s data, it’s possible that Pokémon Go may have resulted in upwards of 150,000 traffic crashes, 250 deaths and billions of dollars in damages nationwide. 

So is it safe to say that, for a time, drivers playing Pokémon Go behind the wheel was a national epidemic?

Not quite.  

While the findings of the study are eye opening, there are certain things to consider. First and foremost, as many as 40,000 people died in traffic accidents in the US last year. Two of those 40,000 were from a single county (Tippecanoe County to be precise) in Indiana near PokéStops. That correlates to 256 nationwide. However, if even just one of those deaths were coincidental and not actually related to the game, then that number gets cut in half.

Furthermore, a county in Indiana is not a county in Los Angeles, New York, Illinois or even Montana. While the data collected in the study might be rock solid, it’s dangerous to try to extrapolate it on such a grand scale as nationwide.

You can read the entire study here and come to your own conclusions based on the full set of data,  but keep in mind that it has yet to be peer-reviewed and should be considered accordingly. 

There’s no denying that the findings of the study are interesting. And the two professors behind it, Mara Faccio and John J. McConnell, did a good job at attempting to deal with the many variables that were out of their control.  Plus, there were obviously plenty of dumb driving around playing the game, as evidenced by countless YouTube videos and driver admissions. But keep in mind that making grand conclusions based on a relatively small data sampling from a yet to be peer-reviewed study is a dangerous game. Almost as dangerous as playing Pokémon Go while you’re driving.