Did the buyback lower gun suicides?

Discussing the Baker & McPhedran paper, commenter “Ungrateful Troublemaker” asks:

Suicides: Shootings, hangings (whether vertical or assisted horizontal), leaps from high bridges and medication overdoses are rather obvious but what about death-by-motor-vehicle? An area that, for a lot of reasons, is notoriously underreported. Wonder if there has been a marked increase in this form of suicide since the gun buyback?

Excellent question. The graphs above show trends in gun suicide (top panel) and non-gun suicide (bottom panel). If you’re like me, you see big declines in both – no evidence of method substitution there. But if you’re Baker & McPhedran, you might see something else:

Examination of the long-term trends indicated that the only category of sudden death that may have been influenced by the introduction of the NFA was firearm suicide. However, this effect must be considered in light of the findings for suicide (non-firearm).

If the fall in gun suicides had been accompanied by a rise in non-gun suicides, a natural conclusion would have been method substitution. But because the fall in gun suicides was accompanied by a fall in non-gun suicides, we are asked to ignore the result on the basis that there must have been some other societal intervention going on that affected both.

A case of shooting the messenger, perhaps?

This entry was posted in Australian Politics. Bookmark the permalink.

7 Responses to Did the buyback lower gun suicides?

  1. Pingback: More on the Guns Study « Criticality

  2. Robert says:

    The graphs are a bit fuzzy, but I think I can make them out. Of course, I’m no statistician so I’m still confused: Isn’t it odd that in the first graph, the observed result is below both the upper and lower confidence intervals for many years?

  3. Christine says:

    Robert: An ‘open link in new window’ trick might work.

    To pick up on Andrew’s point a little less politely:

    One of the authors was quoted in the Age article as directly saying that we can’t be sure that there was an effect on gun suicides, even though deaths went below the set-up-to-fail 95% CI, because non-gun suicides also fell, suggesting a general drop in suicides. (This is of course possible, and one of the difficulties with the method used. It would be nice to at least have included some socio-economic variables that might explain suicides in order to try to get at this problem, though probably wouldn’t be enough.) Anyway, what this means is that suicides could have fallen to zero immediately after 1997, and the authors could claim that wasn’t proof of an effect of the gun laws because other suicides were falling.

    But the paper also suggests that if the gun murder rate had gone down but the non-gun murder rate had gone up, then this would be evidence of a substitution from gun to non-gun murders, and therefore would not allow us to say that the laws had an effect. So, given that the murder rate was increasing slightly over this period, gun murders could have fallen to zero immediately after 1997 and the authors could have discounted this effect because other murders were rising.

    In other words, whatever the outcome of the empirical analysis, the authors can claim there was no effect. From the SMH article on the matter: “The findings were clear, she [Dr Baker] said: “The policy has made no difference. There was a trend of declining deaths that has continued.”” But the findings were not clear, the paper did not identify the trend, and performed no test for a break. These public statements are misleading at best.

    Also from SMH: “The significance of the article was not who had written it but the fact it had been published in a respected journal after the regular rigorous process of being peer reviewed, she [Dr Baker] said.” I truly question the rigour of the peer review process, given that it took Andrew a quick look at a graph and elementary logic (alright, possibly backed up by a bit of experience analysing stats :) ) to completely demolish the paper.

    I will also say that I appreciate that Baker and McPhedran revealed their personal interest in the topic, and I don’t think that means we shouldn’t take their work seriously. If we were to refuse to consider analysis done by people with a personal interest, we’d be ignoring a lot of good research. But taking the work seriously means actually trying to figure out whether this is good work or not.

  4. ChrisPer says:

    I am the author of a couple of papers of amateur quality at the above site, most especially ‘Science in the Service of Politics” and “Ideas Kill: Science Shines a Light on Port Arthur Deaths”.

    I don’t pretend to econometric expertise, I only have an Honours in Science and spatial stats experience in industry. Nevertheless, I take your point about the impossibility of negative murder data.

    But tell me, why was Baker and McPhedran’s paper necessary? The fact is that we have had very poor sources of information and analysis:
    1) Journalism. Highly biased, treats activists as reliable sources when they support new-class biases such as toward gun control.
    2) Public Health Approach papers. Written by people such as Dr Ozanne-Smith. No pretense at objectivity but cheerleading for acivism.
    3) Australian Institute of Criminology papers. The AIC is supposedly Howards ‘official umpire’ but has published the most saccharine papers. The lead researcher, Jenny Mouzos, is honest in data and footnotes but the former AIC Director created a negative climate for honest evaluation.

    Benefit-cost evaluation is missing in action at the ‘official umpire’. It is interesting that in AIC general surveys of homicide, murder is calledan ‘extremely rare event’ in Australia. This phrase recurs in general studies but never appears when studying gun homicides.

    And now it appears that application on the contagion effect is also missing in action. More massacres are happening, and media stories are probably to blame.

    To sum up, Dr Baker and I and others are frustrated that the research results that vindicate honest, non-violent people are de-emphasised by the news process and the selectivity of academics, especially activist academics. Only by doing our own work can we get anywhere. I was thrilled last year when Andrew Leigh started taking an interest, though I saw no reason to expect him to be biased to our side, I expected a QUALITY analysis. Sadly he did not find what he was after and dropped the project.

    But the basic model of human behaviour in gun control thinking has not been questioned. How does raising barriers to non-violent people affect the black swans that carry out massacres? In fact, the first assumption is that all people are average, that decent people suffer some form of derangement when angry in the presence of guns that they do not in the presence of cars, swimming pools or kitchen utensils.

  5. Peter W. says:

    A few quick comments: The ABS a couple of years ago DID highlight the fact that although suicide by firearm had dropped, the drop had been exactly matched by an increase in suicide by hanging. Overall, the efforts of such help and support groups as “Beyond Blue” and increased funding for mental health has helped reduce suicide in some sections of the population. If only such support and funding had been given in 1996, rather than crushing the family heirlooms, fox guns and rabbit rifles of people who had never committed any crime, we would all have been a lot better off.
    In a Country like Australia, with the lowest unemployment in 30 years, we SHOULD have had reducing crime, but AIC noted that such crimes as assault had been increasing by 6% per year, since 1996. That’s five times the pop. increase. AIC stated in one of their reports that improvements in medical services, (plus access to mobile phones) had meant that the murder rate had not increased by the same rate. Now, what if the one BILLION dollars wasted on gun crushing had been invested in our medical system???
    Our local shopping centres and bigger pubs are now protected by security guards, armed with revolvers. Perhaps John Howard really does think he is safer, since his security guards carry Glock 9mm semi-automatic handguns!

  6. ChrisPer says:

    The ABS earlier published data showing a slight rise in total suicides, with a substantial fall in firearm suicides.

    Question: would a statistical category for ‘rope deaths’ that combines suicides, stragling murders, climbing accidents and broken-windlass mining deaths offer meaningful insights?

    IMHO, the firearm-related deaths category is a work of art. It allows the huge number of completed suicides to be appropriated for misleading people about the risks of murder. Same now is happening in IANSA propaganda, but with ‘war deaths’.

  7. Pingback: Andrew Leigh » Blog Archive » Did ‘96 make a difference?

Comments are closed.