Not only that, but the parameters of the study and the collection protocol leave a lot of room for error. All that aside, I do have to weigh my skepticism against the fact that the study does account for every type of accident in its definition of dangerous, so it counts not only fender benders, head-on collisions or getting rammed by a maniac who ran a stoplight, but also accidents due to low visibility, slippery roads, whiteout levels of snow, even getting hit by a moose. Those last few would at least account for how two cities in Maine made the list.