Tag Archives: accidental deaths

When liars figure

Possibly you’ve seen reports about the new study that showed a sharp increase in firearms-related deaths after the Sandy Hook shooting, attributed to the sharp spike in gun sales. If not, here’s a fine example:

Gun sales spiked after Sandy Hook. So did people being shot and killed.
Gun sales in the United States spiked dramatically after a shooter walked into Connecticut’s Sandy Hook Elementary School on Dec. 14, 2012 and killed 26 people, 20 of whom were children. Now researchers say that as a result of that rush in gun purchases, 60 additional Americans — 20 of whom were children — were killed by accidental gunshots in the five months following the tragedy.

That’s… interesting, if really kinda vague. So off to the study itself. You can read it, or I’ll give you the TL;DR.

They used Google Trends to define the period of gun sales that they attribute to post-SH panic buying (you’ll giggle at their search terms), which they determined to be 5 months. They looked at firearms-related accidental deaths — concentrating on children 0-14yo –for that period and compared them to a 2008-2013 period and magically came up with this:

That’s… even more interesting. I happen to recall another gun buying spree in 2008 and 2009. If these folks really found a sales/accidental deaths correlation, I’d expect a bigger spike in those years.

Well, charts are misleading. What’s graphed there isn’t the number of children dead, but the anomaly in deaths; that is, they essentially figured the statistical average for a five month period for the years 2008-2013, and decide the period in question had 20 more than average.

Not having access to their undisclosed “restricted” data, I decided to use the CDC WISQARS online tool, which only breaks data down by year. For the researchers’ 2008-2013 time frame, I got this:

Right off, you — not being a grant-funded academic — probably notice, despite the apparent increasing trend in deaths, that 2013 wasn’t much of a spike. In fact, it’s five deaths lower than the 2011 “spike.” And the average annual deaths for the 2008-2013 time frame is 62.17. So the annual total for 2013 is only 7 higher than average.* Where did they find 20, 13 more in just a five month period? And haven’t we heard that accidental firearms deaths are decreasing?

That’s an oddly limited time frame, given that more data is readily available. WISQARS currently has data for 1999-2015.

And the rate per 100,000 for that age group:

The entire year of 2013 is only the sixth highest, both in numbers and per capita. And again –not being a grant-funded academic –you may have seen the trend: constantly down.

The researchers’ “spike” in their study appears to be largely imaginary, and what there was is more easily attributed to random variance. The lack of sales/deaths correlations with other known periods of increased buying (2008-2009, 2012) invalidates their Sandy Hook-related claim. The overall consistent decrease in child accidental deaths tells me that, rather than untrained newbies going out and slaughtering kids, more people are getting trained and demonstrating safety awareness. Remember: the number of guns are cumulative; if more guns really correlated with accidents then there be a constantly increasing number of accidents instead of the real decrease.

Researchers Levine and McKnight say, “No external funding was used to support this research.” It might be interesting to see what monetary resources are consider internal.

And just for info regarding that additional people of all ages killed?

In 2012, there were 548 people of all ages killed in accidental firearms-related incidents.

In 2013, it dropped to 505.

* The 17 year average for 1999-2015 is 64, meaning 2013 was only 5 higher than average.

Carl is an unpaid TZP volunteer. If you found this post useful, please consider dropping something in his tip jar.