Tag Archives: statistics

“Spike” in law-enforcement-officer deaths: cause for concern?

Today, the National Law Enforcement Officers Memorial Fund (NLEOMF) (probably not an unbiased source) decried a “spike” in “firearms-related” fatalities of law enforcement officers in the United States in 2014.  There were 50 “firearm-related” deaths in 2014, compared to 32 in 2013.  NLEOMF Chairman and CEO Craig W. Floyd said, in NLEOMF’s news release,  “Enough is enough.  We need to tone down the rhetoric and rally in support of law enforcement and against lawlessness.”  The “spike” in “firearms-related” deaths was part of an “increase” in deaths of law enforcement officers from any cause.

Should law enforcement officers be worried?  Is this, like Floyd claims, a result of the recent “rhetoric” against law enforcement officers?  (I don’t need to provide links to news about the events in Ferguson, Missouri, and New York City recently, do I?)

To be clear, I generally support law enforcement officers, although I do not support their illegal actions any more than I support the illegal actions of any person.  And, to be honest, given the trust that society places in law enforcement officers, I think that illegal acts by law enforcement officers, especially involving violence against citizens, should be punished much more severely than illegal acts by non-officers.  But I still think that law enforcement officers dying is a very bad thing.

But I’m not worried.  This “spike” in deaths, even “firearms-related” ones (which, as far as I can tell, includes suicides and firearms accidents, although the NLEOMF does not define its terms), is more likely a result of the random nature of events.  In other words, I think this is an example of Poisson clumping (related to the Poisson distribution), the phenomenon where rare or random events have the tendency to occur in clumps.  If, say, there are an average of fifty law enforcement officer deaths per year, there won’t be 50 every year; some years there may be many more, some years many less.  If you want more information, check this out for a start, or perform a search for “Poisson distribution.”

If we were to become concerned over this one-year rise in law enforcement officer deaths and attribute it to “rhetoric” against police officers, then we would have to wonder if there was some “rhetoric” by sharks against swimmers in Australia in 2009, given the rise in shark attacks in Australia that year:

Shark attacks in AUSTRALIA, 2004-2013
YEAR Total Attacks Fatal Non-fatal
2004 13 2 11
2005 10 2 8
2006 7 1 6
2007 13 0 13
2008 9 1 8
2009 22 0 22
2010 14 1 13
2011 13 4 9
2012 14 2 12
2013 10 2 8

(Source for table)  But, as you can see, the attacks declined again the following year and every year since then.

In fact, the NLEOMF’s own information shows that 50 firearm-related deaths per year is low.  Since 2004, there were the fewest firearm-related deaths in 2013.  The 2014 number, 50, is the next lowest number of firearm-related deaths among law enforcement officers in the last eleven years.  Fifty deaths is still below the eleven-year average of 54.45.  I think it’s a little premature to claim that there is some sort of new problem.

A look at the NLEOMF’s graph of law enforcement officer deaths since 1964 seems to show, if anything, a slight continuous decrease since a peak in 1974 (the true spike in 2001 was due to the 9/11 attacks):


Excuse me a sec while I get out my soap box and climb on it: I tried to see if I could compare these statistics to the numbers of people killed by law enforcement officers.  That is, however, impossible.  Although the FBI keeps detailed statistics on the number of police officers killed each year, and about crime generally, it does not keep statistics on the number of people killed by law enforcement each year.  As a Gawker blogger said, the absence of such a database is likely intentional.  Given that police agencies across the country regularly report detailed data to the FBI, it would not be that difficult to add a few lines to the submission by each agency to detail killings by police officers and for the FBI to compile those statistics.

OK, I’m off the soap box now.

But really, let’s not use fallacious interpretations of two data points to argue that there is a national crisis developing.  Let’s look at some real data, perhaps by examining each killing and any discernible reasons for them.

Leave a comment

Filed under Law Enforcement