6/13/2012

MLive Media Group does an important investigation into data problems with justifiable homicide data

I have long been concerned about the problems with the FBI's justifiable homicide data, so it is nice to see John Barnes at MLive Media Group do a careful investigation of this data.  
Officially, there were 117 justifiable homicides involving civilians in Michigan from 2000 to 2010. Another 95 were killed by police, according to the statistics.
One reason those numbers are low is simple, MLive’s investigation found.
Police are reporting the cases as criminal homicides. When it’s later determined to be justifiable, they don’t change the easy-to-recode electronic records. It’s as simple as changing a “1” to a “4.”
That was the case in Kalamazoo County, where FBI statistics show only one justifiable homicide between 2000 and 2010. There were eight, three by civilians and five by police, MLive found.
“It’s all a data-input problem,” Kalamazoo Public Safety Chief Jeff Hadley said. “What happens is when the initial reports are generated, it’s coded as a homicide. It never gets recoded.” . . .
See also this piece from Monday. 

Labels: , ,

4/04/2012

Reasons not to rely on the FBI's justifiable homicides

Possibly the biggest problem with justifiable homicides is that in many years states don't report the number. For example in 2010, large states such as New York, Illinois, and North Carolina didn't report their numbers at all. Indeed, 14 of the 50 states didn't report a number and the 36 that do report the number are largely not doing so accurately, since just because a state reports the numbers doesn't mean that all or even most police jurisdictions within the state are reporting the data. The laws over what is classified as a justifiable homicide have also been changing a lot over time. Even when New York did report justifiable homicides up until 2006, Hawaii has a significantly higher justifiable homicide rate than New York. The rate is about a third higher. In 2006, Illinois has a rate that is about a sixth of the rate of Hawaii. It appears that what is counted as justifiable homicides is politically determined.
2000 164
2001 222
2002 233
2003 247
2004 222
2005 196

2006 238
2007 257
2008 265
2009 266
2010 278

I don't know how the WSJ gets the claim that justifiable homicides went from 176 in 2000 to 326 in 2010. I get a 70 percent increase versus their claimed 85 percent increase, but more importantly 2000 appears to be an unusually low year -- going from 2001 to 2010 gives only a 25 percent increase.
The WSJ piece did acknowledge that the increase could be due to the changing definition of what is classified as self-defense and not an increase in shootings.
UPDATE: Rob Barry at the WSJ was nice enough to respond to my question on this last point and he notes that they included Florida in their numbers. Florida reportedly had 12 cases in 2000, 12 cases in 2001, and 48 in 2010.
UPDATE 2: There was apparently a 25% increase in justifiable homicides by law enforcement from 2000 to 2010. The data by state is still more complete for justifiable homicides by law enforcement than for civilians. For example, during the years 2003 and 2005 20 states did not have the data for civilians while 7 did not have it for law enforcement.
2000 309
2001 378
2002 341
2003 373
2004 367
2005 347

2006 376
2007 398
2008 378
2009 414
2010 387


UPDATE: A further example of the problems with reporting the numbers even in the states that report the numbers is available here.

Labels: , ,

10/22/2008

Justifiable homicides rising but still dramatically undercounted

USA Today has this. My problem is that so few of these cases are reported that it is hard to know whether there is an increase in the number or a slight change in the rate that they are reported.

The 391 killings by police that were ruled justifiable in 2007 were the most since 1994, FBI statistics show. The 254 killings by private individuals found to be self-defense were the most since 1997. . . .

Florida State University criminologist Gary Kleck says the FBI underestimates self-defense killings by citizens because the ones that are not precipitated by felony crimes may not get counted. "Less than a third of (citizen killings) are reported," he says.

Labels: ,