A map like this can be the source of pride (congrats northern New Englanders) or shame (bury your heads southerners). Of course, like all estimates, this one has its shortcomings. For example, having grown up in upstate NY in that area of dark blue where 24+ inches is allegedly needed to close school, I can tell you that schools closed regularly with less snow than that.
Another shortcoming is that driving conditions and safety are not necessarily directly related to how much snow falls. Winter storms are incredibly complicated. They can include freezing drizzle, freezing rain, sleet, and snow of differing water contents and crystal types. Road characteristics, such as temperatures, also vary from storm to storm and during storms. Last year's freezing rain event here in Salt Lake City shows how a very small amount of precipitation can produce extremely hazardous driving conditions. Conversely, I have driven on rural roads covered by several inches of low density snow with little or no problem. The timing of the precipitation (e.g., relative to rush hour) and the quality and preparedness of winter road maintenance infrastructure and staffs further contribute to winter storm impacts. It also doesn't help when our elected officials opt to bury their heads in the sand and ignore the forecast (you know who you are Georgia Governor Nathan Deal).
Forecasting and communicating all these complexities is a major challenge for the weather prediction enterprise. In addition, even a good forecast can be ineffective if the end user doesn't understand it, doesn't believe it, doesn't know what to do with it, or doesn't take effective action (see this excellent post on Chuck's Chatter). Ultimately, good decisions regarding school closures are based on a lot more than just how much snow falls during a storm.