Misuse of Australian and UK statistics in the US gun control debate

Introduction

I wasn't going to write any more concerning gun control, but a friend posted a link on my Facebook page to a blog featuring the dreadful video below:



The blogger asks: "What happened to gun control successes in Australia?" My Facebook friend wanted me to respond and, so to avoid having to repeat it elsewhere, I thought I'd deal with here in a comprehensive way.

And at least this time I can't be accused of "meddling in US issues" - because the statistics being misused here are not US ones but rather those of my own country and of the UK!

So let's examine what this video actually shows:

Some US network (Fox, I imagine?) reporter asks a few very disgruntled gun owners in Australia about the 1996 gun restrictions, then quotes some Australian crime "statistics" to establish a "link" between these restrictions and the general increase in crime rate in Australia.

What's wrong with that?

Well I'll say this: The video is titled "Watch and Weep" and that is precisely what I felt like doing - not because Australia has "lost the plot" with its gun laws, but because a "news report" that is so:

  • poorly researched; or
  • deliberately dishonest,
(take your pick) can be passed off in the US today as "journalism". It is nothing of the sort as I will now detail. In fact, it has all the journalistic credibility of a rant by Alex Jones.

Correlation? Causation?

It is true that in the past 2-3 decades Australia has had some increase in rates of "violent crime" (more on what that means – and how that meaning differs in various countries – in a moment). But the suggestion that the 1996 restrictions have any connection with that increase is, frankly, laughable.

For a start, as you can see from the adjacent graph, the rate of violent crime was already rising before 1996 – at pretty much the same rate it rose after that year. So attributing the rate of increase in violent crime to something in 1996 is totally without foundation even in the sense of a mere correlation.

But this observation misses a much bigger point: even if there were a correlation between the 1996 Australian gun restrictions and a rise in violent crime, we would need much more than this to establish causation. Put simply, correlation ≠ causation. You need something more than mere correlation to suggest causation.

Okay, I know that some people (like the makers of the above video) probably can't grasp such a "sophisticated" concept. So I'd ask them imagine this:

What if, for example, 1996 had also marked the start of a new trend in society – say a new fashion in haircuts or clothes, or the use of file sharing or social networks on the internet? Would they say that a rise in violent crime from 1996 might be attributable to such a trend?

"Of course not," would be their answer. "We're not talking about something so unrelated. We're talking about guns! You've taken them away, and now people can't protect themselves. Surely the connection is rather obvious?"

I'm afraid it isn't. You see, scientific analysis doesn't rely upon "feelings" about things "being obvious". Rather, it relies on observation and data. And it is important to note that those who study criminal statistics in a rigorously scientific way – in particular the researchers at our Australian Institute of Criminology – haven't even suggested a link between the Australian gun restrictions in 1996 and the increase in violent crime statistics from that time (and before!) to the present.

Why not?

Because, apart from the general observation that some crime rates (including those for violent crime – but notably not homicide) have been rising from the time of (and before) the 1996 restrictions, there is nothing – I repeat nothing – to suggest that we should be looking at those restrictions as a potential "cause".

Now, researchers at the Australian Institute of Criminology have some ideas about the factors contributing to the increase in our rates of violent crime – but gun restriction isn't one of them.1 (I'll let you read through just one of their reports here. You'll note that they don't mention guns except to note that gun-related homicides have declined.)

Factors explaining the rise in "violent crime" rates in Australia

So what sorts of factors might underlie the increase in rates of violent crime in Australia? Well, for starters, things like changes in the rates of reporting. For example, you'll note that rates of reporting of sexual assault against young children have increased – probably as a consequence of both more effective policing and greater awareness of such crimes in the community. Greater reporting = higher statistics for sexual assault – even if the actual incidence of such crime remains largely unchanged.

As a legislative drafter, former prosecutor and former student of criminology, I am also aware of the changes in definition for various crimes over the past 2 decades – both in legislative and statistical terms. So comparing data from one decade to another can be quite problematic and potentially misleading unless one probes a bit deeper.

Other general factors not dealt with in the above report, but commonly considered by criminologists in attempting to understand trends in criminal statistics, include things like abrupt increases in population density, changes in societal homogeneity, economic factors, cultural aspirations, political upheaval and other social changes (including statistical "spikes" due to things like riots).

Why criminologists don't even think of attributing causation to the 1996 restrictions

I know what the people behind the opening video might say in response: "But why aren't the criminologists investigating whether there is a link between the gun restrictions and the rise in violent crime?"

Well, I suppose they might. But there would first need to be some good reason to do so – at least some logical impetus for such a hypothesis.

For example, if it were the case that:

  1. Australia had (by international standards) high rates of private gun ownership; and
  2. data existed to support the proposition that those guns were being used to deter crime; and
  3. a high percentage of those guns had been confiscated,
this would provide a good reason to examine the link between that confiscation and any rise in crime statistics.

But this is very far from the case.

First, Australia has (before and after 1996) always had a very low rate of gun ownership by international standards (in 2007 we had a rate of 15 per capita, where the US had 88.8). This makes the role of guns in relation to crime deterrence a very small one at best.

And we can be even more certain that our 1996 restrictions had absolutely no statistical impact on Australian crime rates since the guns subject of the restriction (ie. high capacity, semi-automatic rifles) comprised only a very small precentage of the already low number of guns privately owned in Australia.

Not only that, but it is worth noting that these restricted firearms were never likely to be used in self-defence anyway, nor was there ever any sense of fear in criminals that these guns would be used in self-defence by home owners or store-keepers. This is because these firearms (as rare as they used to be in private hands!) were also required to be stored unloaded in locked cabinets, with ammunition stored in other locked cabinets.

Furthermore, no one has ever been allowed to "carry" such firearms here, nor has any store-keeper ever been allowed to keep one at work etc. for "self-defence".

So the notion that restricting them has led to a "crime wave" is borne out neither by the statistics nor logic.

Problems with comparing international "violent crime" statistics

That should really be the end of the matter in terms of showing that there is no reason to suppose that the 1996 restrictions played a role in any "rise in violent crime" from the 90s to the present. But I know that many will persist in arguing that they still comprise evidence that "more guns = a safer society". I've heard this argument from a number of people, notably Alex Jones in his "debate" with Piers Morgan.



I know a lot of folks find Alex Jones' arguments persuasive. Emotively, I suppose can see why. But logically they hold as much water as a colander. Let me explain:

Essentially Jones is making the same sort of mistake as the first video makes: he's citing "statistics" for violent crime from another country (the UK), noting a "correlation" with firearm restrictions in the mid-90s, and imagining a causal link between them. All without any basis; when no criminologist has even postulated such a link.2

But in order to satisfy some people who still think Jones has a point, I know that I will have to dig a bit deeper to establish exactly why his argument is a nonsense.

Comparing "apples with apples": international comparisons

Most significantly, when analysing crime statistics, you need to compare "apples with apples" – and preferably apples of the same sort. The closer the compared data, the more meaningful the comparison. The greater the difference in your compared data and the more your conclusions will become strained, if not completely untenable. And the latter increases exponentially the greater the difference in data.

So you really should not be comparing Granny Smith apples with the Pink Lady variety – never mind with oranges. And certainly not with potatoes, giraffes or air conditioners. To get any sort of meaningful comparison, you need your data to match very closely indeed.

I say this because we're all familiar with the common refrain from the gun lobby that the US doesn't have very bad rates of violent crime in international terms. And sure, it doesn't: not when compared with, say, South Africa, Colombia, Estonia, Brazil, Mexico, Philippines or even Taiwan. As this site notes, the non-gun murder rate of each of the latter three is in excess of the total US murder rate (ie. gun and non-gun!).

But here we're starting to compare apples with something else entirely. Because, as any criminologist will tell you, culture plays a huge role in criminal statistics – a role so huge that it can render any attempted comparison meaningless. There is simply too much data – social, economic and political – muddying the water.

If you doubt me, take a trip through China, Japan and South East Asia and try to find examples of graffiti (I thought I saw one in Taiwan but it turned out to be a builder's mark). Then take a trip through any of our Western nations and note the difference.

Why don't young people in the Far East do graffiti? The answer has nothing to do with higher penalties or better policing. The people there just don't tend to do it. Their culture is different.

Okay, so when it comes to violent crime, to what countries should the US be compared? Clearly other Anglo-Celtic nations would provide the closest "match": Canada, Australia, New Zealand the UK.3

Comparing "apples with apples": comparison of "violent crime" statistics

"Aha!" I hear some people cry. "That's why Alex Jones has compared the US to the UK. And the video at the start compares the US to Australia. Gotcha!"

They might then cite the above graph (from this report of the Australian Institute of Criminology) as "proof" of their claims.

I have to admit, that graph certainly looks alarming. But once again, you have to start looking a bit deeper.

First, what are we comparing here? Apples with apples? Or is it apples with oranges? Or potatoes etc.?

"Don't be stupid!" is the inevitable response. "You've asked for a comparison of similar countries didn't you? And violent crime is violent crime!"

Except it isn't. As Wikipedia will tell you: "The comparison of violent crime statistics between countries is problematic. Valid comparisons require that similar offences between jurisdictions be compared. Often this is not possible because crime statistics aggregate equivalent offences in such different ways that make it difficult or impossible to obtain a valid comparison."4

You see, in the US the only "assault" covered by "violent crime" is "aggravated assault". In Australia and the UK, the term also covers "common assault" which includes such things as a fight between two men/women in a bar or outside nightclub (whether a full blown punch-up or even just bit of "push and shove"), a domestic assault, sport hooliganism and even an unauthorized medical procedure.4

Similarly, in terms of sexual assault the US definition of "violent crime" only covers "forcible rape". The Australian and UK examples cover many other forms of sexual assault where consent is absent but no physical force is actually used (at least not of the kind people commonly associate with that term), eg. raping someone who is unconscious due to drink or drugs.4

For that matter, the Australian and UK definition of "violent crime" would cover "groping" (ie. "indecent assault" – what the Japanese call "chikan") where the US definition would not.

If you pause to consider how the inclusion of "non-aggravated" and "non-forcible" assaults dramatically increases the figures, you'll appreciate just how ignorant and simplistic – indeed laughable – arguments such as Alex Jones' actually are.

But it only gets worse for people like Jones. If you read the Australian Institute of Criminology report from which the above graph was taken you'll note that the apparent spike in "violent crime" in the UK is actually an artefact of the data.

The report states that the "increase is largely the direct result of major changes to the way crime data are recorded in England and Wales. First in 1998 and then again in 2002, amendments were introduced to include a broader range of offences, to promote greater consistency, and to take a more victim-led approach where alleged offences were recorded as well as evidence-based ones."5

Indeed, as Chris Smith points out in his blog, an examination of data from the British Crime Survey actually reveals that violent crime has been decreasing in the UK from 1996 – and quite significantly so (see the above graph).6

Meaningful comparisons

Okay, so we can't use "violent crime" for the purposes of a meaningful international comparison. What can we compare? Well how about "apples with apples" – comparing, say, murder rates with murder rates. That would be a start.

When we do that, the picture changes drastically from that suggested by the opening video, Alex Jones and various others in the gun lobby. In comparison to the other Anglo-Celtic nations, the US goes from having the lowest rates to having the highest.

Consider the above graph. A quick glance will confirm that the US murder rates are at least 4 times higher than those in the UK and almost 5 times higher than those in Australia.

Does this show that higher rates of gun ownership in the US are responsible for this difference? Far from it! Unlike the gun lobby, I'm not going to note a mere correlation and suggest necessary causation (even if the correlation in this case is real, unlike the illusory correlations upon which Alex Jones and others try to rely!).

No, in order to examine the effect of guns (and gun control) on crime, the only statistics worth examining are those that actually relate to guns! In every other case the "firearm data" simply cannot be separated from other data.

Consider for a moment that if you compare rape statistics to get a picture of the effect of gun ownership or gun control on rape, somehow you have to filter out all those rapes in which guns did not play (and/or could not have played) any role – either in offence or defence. Date rapes would, for example, be one example. Rapes of minors would be another.

In fact, as a prosecutor I noted that very few rapes actually involved the use of a firearm by the perpetrator. And very few involved cases where the victim was realistically likely to be have or use a firearm – even assuming a "gun culture". The "criminal breaks down your front door while the woman waits with her glock" scenario might reflect how we think (or want to think) violence plays out in society. But in fact, violence mostly follows a far less "scripted" pattern that this.

Gun-related homicide statistics: the only ones capable of meaningful comparison in the gun control debate

Okay, so what "apple with apple" gun-related crime statistics do we in fact have? Really, for international comparison there is only one:

Gun-related homicide.

And this, ladies and gentlemen, is why Piers Morgan was trying to bring attention to the fact that in 2012 the number of gun deaths:

  • in the US, was than 11,000;
  • in the UK, was a mere 39.
(See this report.)

He wasn't trying to force some little "factoid" onto Jones. He was citing the only criminal statistic one can relevantly and reliably cite in relation to the gun control debate.

It might not prove that "more guns = more crime". But it does go some way to establishing that "more guns = more gun-related homicide"!

In this regard, the 2012 figures are particularly interesting since they reflect trends that are even more damning of Jones' argument: the US figure was up from 2010 (when the US had a total of "only" 8,775 gun homicides) while the UK figure was down from 58 (see this report). In other words, in terms of gun-related homicides, the picture in the US appears to be worsening. The picture in the UK seems to be improving.

Whether or not an increase in gun-related homicide is off-set in the US by an decrease in the total number of homicides is another matter. I very much doubt that is the case from a comparison of the overall murder rates mentioned previously. However I also acknowledge that as we move away from gun-related to overall statistics, the conclusions we can reach relating to guns start to become a bit more tenuous (albeit still a darn sight more solid than any conclusions reached by Alex Jones!).

Conclusion

Analysis of crime statistics is not something that can be done superficially and unscientifically. And mere correlations are not indicative of causation.

Instead, to get any meaningful comparison, one must evaluate the data carefully to see first whether there is even a correlation in the data.

Once you have established a correlation, you need to make sure you're comparing apples with apples – not apples with something else. This involves removing from the equation such issues as varying rates of reporting, changes in definition, societal upheaval etc.

For international comparisons, this involves picking countries that are substantially similar in cultural, social, economic and political terms. Then you need to examine only that data that is directly relevant to your study.

When the gun lobby and its supporters try to appropriate Australian and UK statistics, they do none of this:

Their attempted comparisons of Australian and UK vs. US "violent crime" statistics ignore very large (indeed, insurmountable) definitional, logistical, social, economic and political factors that might be at play.

Accordingly if you want to examine the effect of gun ownership or gun control on crime levels by reference to international statistics, you need to look at those statistics that are comparable and to which guns are directly and unambiguously related. In this case, criminologists the world over have only ever been able to agree on one set: those pertaining to gun-related homicide.

A comparison of these figures does not flatter US gun policy at all.

Footnotes:

1. In fact, it is hard for criminologists to point to any effect at all from the 1996 restrictions – good or bad – on broader criminal statistics. That is because it is very hard to divorce social, economic, political data from that pertaining to the restriction of a discreet class of firearms.

2. Alex Jones would probably argue that the criminologists, like most of the media, are part of some sort of "conspiracy" to cover up the truth. I won't bother addressing such blatant nonsense, except to say that scientists (including social scientists like criminologists) aren't being dictated to by anyone – never mind some sort of leftist "conspiracy".

3. It is important to note that the UK has a far older European culture than, say, the "frontier" cultures of the former colonies. And it is worth noting again that New Zealand's culture is closer in some respects to the UK's than it is to Australia's. These sorts of things have to be kept in the back of the mind when making any comparison between these countries and the US.

4. See the Wikipedia entry on Violent crime.

5. See this report of the Australian Institute of Criminology.

6. See also this report.

Copyright © 2013 Dejan Djurdjevic