Tuesday, 15 January 2013

Guns in the US and the UK: Piers Morgan vs. Larry Pratt

One of the big topics in the business world is how to deal with big data. I've been doing some reading about multimillion dollar software programs that larger firms are now using to handle the massive amount of information that they have, but can't digest. Entire branches of technology and professional services firms are dedicated to wading through big data in the pursuit of gleaning relevant information from an overwhelming heap of noise.

Yet we are expected to sort out big data on a daily basis. This is most clearly evinced by the current debate about gun control, which is raging in the United States. Flick on any news program and you will be besieged with data about guns. Both sides of the debate cherry-pick statistics that help advance their positions. Curiously, some draw wildly different inferences from the same data. 

In the past month, I've heard a lot of different numbers thrown around about guns. For instance, I've heard no fewer than three different numbers tossed about regarding gun related homicides in the US. On my quest to find the right number, I realized that—well, there is no right number. Gathering this data isn't always easy, which might be why the FBI statistics vary from the CDC statistics, which vary from the United Nations statistics. Nevertheless, you're usually in the ballpark. While the numbers do show some variability, they almost never vary more than 5%. 

But in an interview with Larry Pratt, President of the Gun Owners Association of America, Piers Morgan contended that there were 39 firearm related homicides in England and Wales in 2011. Pratt's number: 970. This number would represent a 2387% increase in the incidence of firearm related homicides when compared to Morgan's 39. These drastic differences are endemic of the gun debate in America. Both guys cited data. And if you're citing data, you must be credible, right? Turns out it's important to cite the right data. The Home Office of the UK—the governmental department responsible for immigration, crime and policing—published this report listing crime statistics for 2010/2011. On Piers Morgan's blog you can find a link to the most recent data for 2011/2012, which hasn't been compiled into an official report yet.

I took screen shots of the relevant data:

These numbers corroborate Morgan's statement. So where on earth did Larry Pratt's 970 come from?  Pratt asserts that: "according to your investigator of your constabulary, the data that you’re using for the murder rate in England is a sham. There’s a monumental misreporting of what constitutes murder. If three people are murdered, it’s likely to be counted as one event."

Is this true!? No, not exactly. Pratt was referring to two articles from The Telegraph. The first is a seemingly extinct 1996 article in which UK police were accused of massaging crime statistics. While this may very well be true, these statistics mainly related to burglaries and vandalism, not homicide. 

The second Telegraph article, which the GOA had archived on its website, gets to the heart of Pratt's assertion. Here are some excerpts: 

Figures to be published by the Home Office this week will massively understate the scale of the problem.
Data provided to The Sunday Telegraph by nearly every police force in England and Wales, under freedom of information laws, show that the number of firearms incidents dealt with by officers annually is 60 per cent higher than figures stated by the Home Office.
The explanation for the gulf is that the Government figures only include cases where guns are fired, used to "pistol whip" victims, or brandished as a threat.
Thousands of offences including gun-smuggling and illegal possession of a firearm - which normally carries a minimum five-year jail sentence - are omitted from the Home Office's headline count, raising questions about the reliability of Government crime data.
"Firearms offences are comparatively rare in Britain, and the vast majority thankfully do not result in a serious or fatal injury. But if the police already collect this information it is difficult to understand why it should not be put routinely into the public domain."
The Home Office crime figures document states: "Firearms are taken to be involved in a crime if they are fired, used as a blunt instrument against a person, or used as a threat."

While we thank Mr. Pratt for enlightening us on the current state of police work in Britain, the manner in which the Home Office reports non-violent gun crimes overextends the framework in which the current argument ought to be couched. Yes, England and Wales are not without their faults, and perhaps the Home Office should be more meticulous when it comes to accounting for non-violent firearm related crimes. However, it is important to note that the statistics that are underreported have to do with firearm related incidents that do not include a weapon being fired, used as a blunt object, or brandished. The underreporting involves smuggling and illegal possession. Violent crimes involving firearms are, in fact, accurately reported. 

Let's stick to the question on the table. Where did Pratt get that 970? 

Well Pratt was citing an article from (you guessed it!) The Telegraph. This article reports that the UK (England, Scotland, Wales, Northern Ireland and some small islands) had 927 total murders in...2009. So, let's just throw all analytical structure out the window and arbitrarily cherry-pick data from whatever year, source and, hell, statistic, we feel like!

Let's compares apples with apples. 

According to the UK's Office for National Statistics there were 550 homicides in England and Wales in 2011/2012. (Looking back at Pratt's figure of 970, I find it very curious that there were more firearm related homicides in the UK than total homicides). Guns accounted for 7.1% of total homicides. There were 640 homicides in all of Great Britain, of which guns accounted for 6%. (These numbers are corroborated by data from the latest Home Office report, which extends to 2011).

Let's compare these figures to the United States.

According to the Center for Disease Control there were 16,259 homicides in the US in 2012. Guns accounted for 11,078 of those homicides, or 68%

The FBI's latest crime statistics show that there were a total of 12,664 homicides in 2011. Guns accounted for 8,583 of them, or 68%

The National Institute of Justice cites a Department of Justice report, showing that 68% of homicides in 2006 were committed with a gun. 

And the final piece of evidence I will submit is this table from the United States Census Bureau, illustrating that guns accounted for 66% of all homicides from 2000 to 2008. 

Larry Pratt may have been right to criticize the UK for certain practices that they employ. Heck, the UK might have some work to do itself. But we have a serious problem here, and the numbers attest to that fact. Whether or not conservatives want it, the gun debate is at the epicenter of public attention. Maybe the argument for stricter gun control won't be borne out by the facts. But let's have an honest debate. Let's get our numbers right. And let's show some intellectual integrity here. 

No comments:

Post a Comment