Are kids really safer in schools?

A couple of comments on my last post challenge my assertion that "Kids are much safer in school than any other place."

I admit, I was lazy, I didn't do any research to back up that statement; I'm parroting Lenore at Free Range Kids, who I trust because I've followed her blog for years and she is always rational and data-driven. And maybe "much safer" and "any other place" are exaggerations; probably kids are just as safe or maybe even safer when they're at home, sleeping in their beds at night (or maybe not, because I'm pretty sure domestic violence is a lot more common than school violence).

But it got me to thinking: how would I go about proving or disproving that kids are safer in school than anyplace else?

I'm a skeptic, and I'm very aware of all the ways we can fool ourselves into "proving" something that we already believe. So if I had the time to research safety in schools, here's how I'd go about it:

First, I'd try to get more specific about what I mean by "safe" -- really, I mean "safe from physical harm." I'm not going to wade into a debate about whether our kids are being developmentally stunted or mentally harmed in school.

Then, I'd find a data source (or, ideally, two or three) that I trust that tracks statistics on deaths or injuries. I'm sure there are databases that keep track of emergency room admissions by age, time, etc.

Before I looked at any of the data, I'd decide what to look for. Any large dataset can be sliced and diced a million ways, and 5% of those ways will give you statistically significant (but meaningless) results.

In the "are schools the safest place for kids to be" case, I'd look at the injury/death rate for school-aged kids during school hours, and compare it against the injury/death rate at other times: weekdays during non-school hours, and weekends.

But that's not quite enough-- maybe the middle-of-the-day school hours are just safer in general. So I'd also look at injury/death rates for kids who are too young for school, to try to isolate the "in school" versus "not in school" factor.

To really do it right, I'd hire a statistician to run a regression over the data (or would take that statistics class I'm always thinking I should get around to taking...). But before doing all that, I'd run a Google Scholar search to see if anybody's already done the hard work.

A couple of minutes of searching "school safety" turned up a bunch of studies that focus JUST on violence in schools, but also "The Inherent Limits of Predicting School Violence"which says:

... violent deaths are a rare event, with less than 1% of the homicides and suicides among school-age children occurring in or around school grounds (Kachur et al., 1996). Moreover, the rate of violent crimes committed by juveniles remains low during the school day, but it spikes at the close of the school day and declines throughout the evening hours (Snyder and Sickmund, 1999), indicating that school hours are probably the safest time of the day for adolescents.

I didn't find any contradictory evidence, so short of crunching the numbers myself (which I'm not going to take the time to do), I'm going to stick with "kids are safer in school."