There’s been a bit of publicity about a recent study of the effects of the Australian gun buyback. The central finding of the authors was that, while gun homicides declined after the buyback this was merely a continuation of a pre-existing trend.
I’m dubious about the whole approach. In the absence of a well-founded explanation for the trend, there’s no reason to treat maintenance of the trend, rather than the level, as the null hypothesis. The rate of gun homicides has clearly fallen (the authors find the same for suicides), so the data supports the policy, contrary to the claims.
And eyeballing the data, I’m doubtful that it’s even sufficient to establish the existence of a declining trend for the period up to and including 1996. It might be argued that the Port Arthur massacre in 1996 should be excluded and that a downward trend would then emerge, but, given that this was the even that precipitated the buyback, this seems like begging the question to me.
In any case, Andrew Leigh has the ultimate knockdown objection. If you look at the confidence intervals, the only way the gun buyback could have been shown to work, on the authors’ tests is if gun homicides fell below zero by 2004. Clearly, even if you buy the declining trend story, a linear trend is just wrong.
Mark Bahnisch has more, though quite a few commenters don’t seem to appreciate how conclusive Leigh’s refutation has been.
John Quiggin 
Limiting homicides to guns does not show the full picture, there is strong evidence that post regulation guns have been replaced as a weapon of choice in favour of other weapons including knives.
The proliferation of armed holdups has shifted from banks, with their increased security eg pop up screens to all night petrol stations, often the weapon of choice being a syringe or knife.
There is also a decrease in heroin use which has impacted on armed holdups.
Better to look at homicides and crime as whole.
Not exactly the “ultimate knockdown objection”. True, they should have used a one-sided error distribution, truncated at zero, but people draw graphs with error bars extending into infeasible regions all the time. One of my pet peeves. They get away with it because in most cases incorrect error modeling does not affect the conclusions.
Remember, “95%” was not handed down by God as the magical cutoff beyond which a hypothesis is proven.
To me it looks like there was a clear trend that has continued since the buyback. And think about this: why would the buyback initiate a trend? One would expect more of a one-off reduction. It’s not like the guns have been steadily taken out of circulation over the past decade.
However, what has obviously changed since 1996 is the variance: no more big peaks. Is that because the buyback stopped mass-murders? If so, that is a better argument in favour of the buyback than any argument about trends.
proust,
Does that that, if there is a big incident in the future, we discount that theory? A lot more work needs to be done than this. This is a valuable start that raises a number of questions and, over the next few years we can expect a number of papers looking at this in a critical fashion.
This is the start, not the end, of the debate.
Well, if the variance goes back up in the absence of any reintroduction of (semi-) automatic weapons – yeah, my theory is toast. But as a good Popperian, I always claim refutability as the least requirement of any theory, so I don’t get your objection.
Excuse my ignorance, but what are the pointy bits regularly occurring in graph? Is there are regular outbreak of gun deaths occurring, like sunspots?
Well, for example, the last big spike was 1996, the Port Arthur massacre (35 deaths) which prompted the gun buyback.
“I’m dubious about the whole approach. In the absence of a well-founded explanation for the trend…”
Aging of the population.
Any idea how the trend compares to other western countries?
“I’m dubious about the whole approach. In the absence of a well-founded explanation for the trend…�
“Aging of the population.”
Decreasing unemployment.
Increasing wealth.
I agree John, that a declining trend is not the only null hypothesis that could have been considered. To me, the prior data could better be seen as a series of plateaus – levels – with the only true downward trend appearing after the gun buyback. That is, pretty much the opposite model from that of the authors.
“is the variance: no more big peaks. Is that because the buyback stopped mass-murders?”
The buy back only bought back semi automatics, the kind of guns that will be used in mass murders, because large numbers of shots can be fired off very quickly. Shot guns and other guns that have to be reloaded with every shot, which are used in domestic murders, were not bought back. So we’d expect fewer massacres, but no change in domestic gun murders, resulting from the buy back.
Which from the data seems to be pretty much what we got.
I wonder if there will be an observable upwards spike amongst the last remaining well-armed segment of society, the farmers, from additional suicides during the current extended low rainfall period.
Unlikely, given the phenomenal level of largesse from the federal govt.
Give up a welfare entitlement like that through suicide? No way. At this rate, the farmers will be praying for more drought.
Anyone else here care to underwrite the success of my business with their taxes?
Depends, do you have a well organised and extremely over-represented conservative socialist political party behind you? Are you a ‘national icon’?
I think you meant national party icon…
Given the loss of liberty then surely the onus of proof should be on those that advocated this law rather than the other way around. If gun homocide dropped to zero but homocide overall stayed the same then I would view the laws as a complete failure. In murder motive is more significant than means and if murderers merely changed weapon than nothing has been gained.
The following NHMP report shows no real change in the overall homocide trend from 1989/90 to 2003/04. It is near constant in the range 300-350 incidents per year. The relevant graph is on page 34 of the electronic version (page 26 of the printed version). If there is any trend after 1996 it is a slight upward trend until 2001.
If access to guns becomes more restrictive and then homocide by firearm declines but homocide overall is steady then what should we conclude? I think it is reasonable to conclude that access to firearms is mostly irrelevant. Ristricting one means of murder has little impact on the rate of murder. The report essentially says the same thing in the following passage:-
Opps. Here is the report:-
Click to access rpp66.pdf
For those interested the issue is also getting discussed at the Australian Libertarian Blog site:-
http://alsblog.wordpress.com/2006/10/25/am-i-wrong-to-have-changed-my-mind/#comment-858
Terje Says:
Number of incidents isn’t corrected for population growth.
The graph on page 33, the one that’s actually labelled “Trends in Homicide Victimisation 1989-2004”, shows a drop after the buyback and a further decline since.
Nice try, Terje, but no cigar.
I’ll grant you that the numbers should be adjusted for population. And I’ll grant you that if you squint you can make out a downward trend in the national homocide incident from about 2 to 1.9 between 1989 to 2004. Maybe if I wasn’t reviewing it on my PDA at the moment the trend might even appear to be more than that.
However I don’t note any change of trend after the 1996 midpoint. Certainly nothing that would compel me to think the law reforms actually achieved anything. I am inclined to believe they were a monumental waste of money.
“drop after the buyback and a further decline since.”
Hardly: down, up, up, down, up, down down. Given that after Port Arthur it was guaranteed to be down whatever happened, that’s equal up years to down years.
The most interesting thing about that graph is the murder rate in NT is about 6 times the rate in the rest of Australia. Let me guess why that would be: Aboriginals beating their wives to death?
Another example of spectacularly successful lefty welfare policies. Oh sorry, I forgot: your intentions were good.
Maybe we just need to ban the possessions of knives in this country. Dangerous things. Feet and hands? Ban them too. There goes most of our homicides.
proust Says:
Jesus. I’m arguing statistics with someone who doesn’t know what they mean, and can’t even count.
Terje Says:
Of course nothing would compel you. That’d be kinda against your religion, wouldn’t it?
Andrew,
Yes except it still wouldn’t remove the primary factor which is motive. You can kill somebody with a chair or a snow dome or swimming pool acid or a broom or a hammer or a lamp or a truck or any number of readily available items. The best solution is a simple blanket ban on murder.
Regards,
Terje.
SJ,
Please stick to playing the ball. If you want to trade insults then please try somebody else. I don’t wish to waste my time on such a passtime.
Regards,
Terje.
Terje
Go preach to the converted then.
You’re making claims about trends you can’t see because your PDA screen is too small? Please.
You can’t see any change in trend after the trend began in 1996? The trend which for your argument to be viable shouldn’t exist at all? Get real.
Terje Says:
Perhaps a bit of perspective is in order, not just NRA talking points.
Can you honestly claim that the buyback of semi-automatic weapons was an inappropriate response to Port Arthur, on the basis that Martin Bryant could have killed 35 people with a chair instead?
SJ,
He could have made a fertilizer bomb or driven a car through a crowd or poisoned the town water supply or set a building on fire or any number of such things to satisfy his premeditated intent to kill lots of people.
I have now finished my commute and have seen the graph you refer to in a larger mode. I don’t see any “significant” trend change. Certainly not anything that stands out from the natural variability or that I would bank on.
Call me old fashioned but I take the view that when the government is going to deprive people of liberty then they should build their case on evidence that is beyond reasonable doubt. Or at a minimum they should have evidence that makes their case on the balance of probability. The decline you cite provides no such evidence and if you think it does then you are clutching at straws.
Regards,
Terje.
Terje Says:
Yes, he could have done those other things that are capable of mass murder. What he actually did was to use a semi-automatic weapon. The government took steps to reduce the liklihood of the episode being repeated.
You should be aware enough of world events to realise that the government has also taken steps to make it harder to purchase the fertilizer necessary to make the fertilizer bomb, or the chemicals necessary to poison the town water supply, etc.
Note that cars haven’t been banned, because there’s a trade-off between preventing mass murder and causing dis-utility to the 100% of the population who rely on road transport.
Pure ideological libertarianism is useless for the real world.
This is just silly. Semi-automatic weapons were proved to be a cause of mass murder by Martin Bryant. Beyond reasonable doubt. Maybe you wouldn’t be a mass murderer if you had one. But the rest of us don’t trust you enough with something that would seriously infringe on our right to life if you used it in a mass-murder.
Our right to life trumps your right to be a w****er with a machine gun.
I would have thought everybody elses right to a civilised discussion trumps your right to be a w****er with your keyboard.
I suspect that you’re using terms you don’t know the meaning of, i.e. “significant”, “trend” and “trend change”.
Terje Says:
“Natural variability”? Ditto.
Sanctimonious:
Yeah, sure, to a libertarian, civilised defense of mass murder is a worthy goal. To the rest of us, not so much.
What more insults? Obviously you’re a very clever person. Would you like to say anything that contributes to the discussion?
~~
For the interested reader it is hard to see the time scale on the image that John Quiggin has included. However if you link over to the article he references you can see that the chart starts in 1980.
However the trend is actually longer than that. On page 20 of the following chart you can see the history since 1915 and it is clear that the trend turned downward at the beginning of the 1970s.
Click to access facts_and_figures_2005.pdf
Terje
It’s obvious that you take any criticism of an argument that you present as a personal insult. It’s quite a natural reaction, but it’s a juvenile reaction. Grow up and get over it.
SJ,
You call me a w****er and then say that I am a defender of mass murder and then suggest that I shouldn’t consider it a possibility that you are trying to insult me? Hmmm let me think about that.
Okay I thought about it. I think you were being insulting. I think you do it pathologically. It’s a pattern of behaviour that you don’t even attempt to change because you’re proud of it.
You say it is juvenile to point out that somebody is being insulting and asking them to try and stick with discussing the issues. I find that quite amazing. I think that insulting people in the first place is the behaviour that is juvenile.
Kind Regards,
Terje.
Whatever.
Your lack of perspective is pretty amazing.
As is your lack of good manners.
Yeah, yeah. In libertarian fantasy world, manners are more important than murder.
Terje – ignore him. He’s not worth the keyboard wear-and-tear.
This seems to have got out of control. Everybody take a deep breath and stop, please.
This Dilbert cartoon (from the 28.10.06 calendar page) somehow reminded me of this badinage and the whole Godwin’s Law thing.
Curiously, despite people dying in massacre style numbers, I can’t find any mention of what type of gun was used in the Whiskey AuGoGo murders…..
Gun controllers are w*nk*rs
Most males (and a high proportion of females) are w##kers. While the phalic overtones associated with grabbing guns might make for “cheap shots” there is a serious debate that can do without this slanderous approach to discussion.
Those that want to regulate guns generally start from a set of presumptions that are well meaning. I was myself quite in favour of gun regulation at the time that Howard implemented the reforms in 1996. However logic and knowledge won me over.
Australia had its lowest incidence of homcide during WWII when the rate dropped to near 1 in 100,000 by population per annum. My best guess is that this correlates with a time when lots of young men were absent from the country. The rate then rose to a peak in the early 70s presumably because the baby boom ballooned the number of young men in the population. In spite of unemployment and other social unrest in the 1970s things generally improved during that period and have continued to improve. Presumably because boys grew up and settled down.
Whilst it is common enough to normalise the number of homocides against the total population when analysing such things it would be interesting to see the number of homocide over the last century normalised against the population of young male adults (who are the predominant offenders). My guess based on reviewing charts and studies done elsewher is that such a chart would be quite flat. I certainly see no evidence that it’s shape would be disrupted by changes in the law and plenty to suggest that the law is mostly irrelevant.
Does anybody have access to the raw demographic and homocide data for the last century and the courage to analyse it accordingly? That way this discussion could yield more light and less heat.
Hand, foot, fertiliser, knife, car, semi-automatic weapon.
Spot the odd one out?
Because we can’t ban EVERYTHING that COULD be used for murder, we shouldn’t ban something that can pretty much ONLY be used for murder.
Have I got this right, Terje? Please don’t insult me with a list of all the reasons farmers need to get off 100 rounds a minute. The crows have got to be ‘coming right at you’ pretty thick and fast.
Gun ownership is fetishism dressed up as either sport or ‘rights’. The fantasy of one day using your gun to enact justice or ‘protect your family’ simply DOES NOT OCCUR.
Quick cost-benefit analysis, people. I don’t have the stats, but it would be interesting to see them. Number of violent crimes sucessfully averted with guns (I’m feeling generous – throw in successfully avenged as well) versus number of violent crimes perpetrated with guns.
GO!
Are you asserting that the 600,000 guns that the government bought back in 1996/97 at a cost of around $350 million dollars were only ever used for murder? The murder figures from prior to 1996/97 suggest that there was no where near that amount of murder happening. There were around 350 murders per annum and firearms were not the dominant means of murder(and not even the second most dominant). And nearly 95% of gun related homocide is with an unlicenced firearm anyway.
I’m not a shooter so I can’t articulate a personal motive to own a semi-automatic weapon. However to suggest that 600,000 firearms were owned just in case the owners needed to murder, or were actually being used for murder is very cynical. Here is what an actually shooters had to say about the ban:-
http://www.theage.com.au/news/national/anger-lingers-among-those-who-lost-their-firearms/2006/04/27/1145861489398.html
As for your cost benefit analysis. Doing such an analysis might make sence if the relevant legislation actually changed the outcome. However if banning guns merely shifts the means of undertaking a violent crime, with the same net cost/benefit situation, then the law adds nothing to society but merely puts a lot of people out.
However if somebody could demonstrate that gun regulation achieved a statistically significant increase in safety for the community, allowing of course for other factors such as demographic shifts (ie the number of young males in the population) then I would be happy to support such a regulation. Assuming of course that the law was not broader than necessary to achived the result.
In essence I would like to see these laws justified empirically rather than hysterically.
FDB Says:
Terje Says:
Obviously, Terje, FDB is asserting no such thing. It does you no credit at all to so egregiously misrepresent the argument that has been put to you.
My question is simple enough. FDB has implied that these 600,000 weapons had only one purpose which was for murder. That implies that either all these people bought these guns with the intention of murdering somebody or else FDB is wrong and these guns actually have more than one possible purpose. It’s a sincere point of logic that merely needs a rational response. If in fact it is the case that these guns had no purpose other than murder then the question asked by FDB stands and I will concede that they should be banned. However I think that it is reasonable that FDB justify this particular assumption.
If on the other hand I have misunderstood FDB and FDB was not refering to the laws of 1996 but was in fact merely refering to the principle of hypothetical banning of a hypothetical weapon that has absolutely no purpose other than murder then I would hypothetically have no reason to disagree with such a ban.
You are telling straight out lies. And not for the first time in this thread. What’s the point of this crap?