Science in the Service of Politics

Australian Criminology’s Integrity Test

 21 January 2003


In 1996-97 Australia spent almost $500 million confiscating 600,000 firearms from a  low-risk population of ordinary, law-abiding Australians.   This buyback was implemented in a climate of moral outrage,  following the massacre of 35 innocent people at Port Arthur in Tasmania. 

The criminologists who made and promoted this policy include at least three past and present Directors and a number of staff and Associates of the Australian Institute of Criminology (AIC).  It originated in the report of the National Committee on Violence (1989) which put forward  a gun control wish-list unsupported by research or even a named source . 

Following the buyback, the Prime Minister appointed the AIC to run a National Firearms Monitoring Project to report the results. 

The AIC is a Government body, formed in the early 1980's to provide policy advice without the prejudice of public politicking over law and order.  Though a publicly-funded body, like the ABC it has a notional degree of independence through supervision by a Board of Governors.  In practice it works closely with the Attorney-General's Department, and lives with the annual tension of earning its funding from the Howard Government.

It is five years since the gun buyback was completed (written in late 2002).  The data are in.  The effects in future years will become less and less measurable as they are diluted by further interventions and social change.  In a string of publications, Mouzos and other AIC authors more or less honestly reported data as it accumulated.  However, culminating in Carcach and Mouzos (2002) and Reuter and Mouzos (2002) the results appear to have been deceptively interpreted by unsupported speculation.  In October 2002 the AIC staged a media event to misrepresent this speculation as proof of benefits from the buyback, and were silent when Australia’s Prime Minister promoted this falsehood in the media for political gain

Australia has paid up over $500 million in two buybacks, on the recommendation of leading lights of Australian criminology.  Where is the cost-benefit research? 

The AIC has wavered in its duty of honesty.  And here is a challenge to the rest of Australia's criminologists: will you speak the truth?  Will you challenge a lie made in the name of your profession? 


Talking Religion at the Dinner Table

In Australia's  new class system, 'Cosmopolitan' values are required for admission to public discourse.  Anti-racist, anti-sexist and other values are almost mandatory, and status is displayed by waving opinions as weapons.  The issues which this applies to are relatively few, and are known as 'marker issues' (Betts 1999). 

Rational discussion on marker issues like gun control is not expected.  Moral value is assigned to the speaker’s opinions by the use of emotionally-charged language.  Win-win solutions are not possible because the terms of debate, set by activists of the cosmopolitan class, are defined in contempt for those who disagree (Kates 1992).  In Australia, the cosmopolitan value system almost exclusively qualifies entry to public discourse.  In the United States the marker issues are not as closed, at least in the public forum.    

If moral status is assigned to the individual belief on the marker issues, academic peer review can be part of a ‘status auction’, and reinforce an activist approach rather than build scientific detachment.   Even passive academic language may carry a coded moral significance. Activist attitudes coercively frame the interpretation of evidence for marker issues, including gun control.   

‘Data Torturing’ (Mills 1993) may go unchallenged in peer review of research on marker issues.  Data trawling in the absence of experimental design catches any supportive data but throws back non-supportive observations. This cognitive trap is called confirmatory bias.  It builds a sense of ‘justification’ as a mechanism for social discrimination, whether against minority races or minority sports.   The dominant ideology in this way reinforces itself with ‘moral’ self-justification on marker issues.  Data trimming to fit the ruling hypothesis is made acceptable. 

As in other kinds of discrimination, the dominant social group projects or ‘sees’ unacceptable values onto the out-group.  They represent them as morally contaminated and allow their opinions into debate mostly as straw-man arguments.  They pass legislative sanctions to validate the dominant group’s moral position.  In this kind of politics, symbolic action is approved in contemptuous disregard for evidence.

 The obvious historic examples are America’s Prohibition, Western Australia’s mandatory sentencing law, and the Australian anti-Chinese immigration laws that evolved into the White Australia Policy.  These well-known examples of legislative sanction are ultimately founded in moral status display.


Research activism disguised as science

Research advocacy is the conduct and use of partisan research for the purpose of directing public opinion.  “Research advocacy” may include:

(1)    Conducting “research” that is beset by error or breaches ethical and scientific principles;

(2)    Using such “research” to reinforce pre-formed opinions, beliefs or ideologies;

(3)    Misrepresenting such “research” as competent, impartial or objective.

 Interest groups may do accurate, ethical research, in accordance with scientific principles, and publish their results without objection. Advocacy research - which is not objective in design, observation or interpretation - should be clearly distinguished from science.  

 Publicly funded, independent scientific bodies have a powerful implied authority, which to activists is a very desirable tool for changing public opinion.  Researchers there may be rewarded in prestige, funding and publicity when they produce ‘acceptable’ outcomes on the marker issues. 

Unabashed activist research papers have been published as science and used to influence public policy.  Wright et al (1983) found that the literature on gun control was almost entirely of this kind.   The scientific literature on gun control is still heavily contaminated.  In some cases, major works have been acclaimed because they support the dominant ideology but later exposed as fraudulent (most obviously Bellesiles 1998, but others exist such as Kellerman and others 1993).  The data may be quite blatantly adjusted to produce ‘socially beneficial’ results (eg Alpers and Morgan 1995, Kellerman 1993).   

Work which is non-supportive, even if credibly scientific, is severely attacked.  The form of attack is principally ad hominem and gets wide media support.  Examples include the attack on Lott (1997) on the grounds he worked for a university body which had received funds from the Olin Foundation and Alber and Messner’s 1995 attack on  Kleck (1991).  

Pro-gun activists respond with their own ad hominems (eg the ‘gun-aversive dyslexia’ diagnosed by Kates et al. 1994) but  do not get any significant media support.  Mainstream pro-gun literature propagates a few falsehoods and urban legends to its own uncritical audience, but does not dress it as science.  Most of them are obvious propaganda anyway.  As examples we have the blatantly false Sarah Brady ‘socialist agenda’ quote and the ‘1922 Communist International action agenda’, a piece of propaganda read into the US Congressional Record by conservatives on two or three occasions.  

Within the criminological profession it appears that activist papers are not publicly distinguished from science.  Neutral researchers will recognise contamination and avoid citing those works, but not have them ejected from the canon as Talent (1989) did for one person’s fraudulent geological publications.  Nor do criminologists appear to alert public policy stakeholders to the issue of contamination.  

Activism and Australian Criminology

Criminologists in Australia have strongly supported increased gun control.  From the (almost fair) seminal work of Harding (1980) to Graycar (2000), the published Australian literature has a clear subtext:

1.    Social costs of guns are ‘bad’ at any magnitude.   

2.    Benefits will not be intellectually recognised; for instance self-defence value is denied or excluded on moral moral grounds.  Recreational benefits are ignored 

3.    Control measures are a social ‘good’ per se.  Utility or benefit-cost are not of interest, except when future and hypothetical.   

4.    Those who disagree in any way are at best misguided, at worst are promoting social evil.  Their arguments suffer moral contamination and will not be tested except by straw-man techniques.     

5.    The symbolic meaning of ‘guns’ and the debate are to be interpreted morally, in terms of the ruling values.  Alternative symbolic views or values disqualify the holder from participation  in the discourse.    

Activist contamination in Australian criminology is clear in its signs:       

1. The author does not report contamination of the literature; questionable works continue to be cited, but balanced or opposing works are cited less or not at all.

The AIC’s Trends and Issues (T&I) papers do not acknowledge contamination through the entire series.  Questionable authors are cited without comment (Zimring, Frank and Hawkins 1987 in T&I 10, Kellerman et al. 1993 in  Mouzos 2000; Alpers in T&I 70, T&I 151).  

Mouzos (T&I 151, 1999) in a footnote acknowledges data selection by Alpers explains his contradictory findings.  A scientific auditor would read this as a sign of deeper trouble but the AIC leaves this buried in the footnotes.  Australian policymakers are not alerted to contamination.

2. Discussion is framed to contain negatively coded implication, and echoes activist falsehoods.

To Mukherjee and Grabosky in T&I 70, shooters ‘become enlightened’ when they leave the sport. 

‘Gun Deaths’ as a combined statistic (eg RPP4, T&I 70, T&I 269) is an activist tool, capitalising on high suicides and implying accidents and murders using guns are also at a high level.  To test this idea, consider the scientific value of linking ‘ligature deaths’ such as suicide hangings, industrial and car accidents involving strangulation with strangulation murders.

Murders are examined to see what % are committed by licensed shooters.  The implication is that a high statistic discredits shooters, but even if it were 100% there would be injustice in implying that the huge population of non-murderers were in some way guilty of or even more likely to commit murder.  This is framed around the activist belief that anyone can ‘snap’ and murder if a gun is handy.  It is not considered realistic to expect ‘snap’ murders by ordinary people in the presence of cutlery or automobiles.  It is not polite to point out a genuine 900% higher likelihood of indigenous people to commit murder, yet this statistic implies shooters are in some measure dangerous.

‘Firearm-related’ deaths or crime - in linguistic sense - implies that the presence of a firearm in part caused the crime.  Similar terms such as ‘ligature-related’ or ‘knife-related’ are not widely employed.  This implies the writer has a view of causation by an object rather than a scientifically neutral view.   It is widely claimed (eg Graycar 2000) that there is a proven link between the general availability of guns and violence, but the evidence on both sides is generally selected with confirmatory bias and ignoring cultural variables.  Substance abuse, entrenched social disadvantage and criminal history have a far stronger relationship. 

Articles are even published with titles like 'Deaths Caused by Firearms'!

Mouzos and Reuter (2002) use a ‘note on data sources’ to say that ‘until recently there was no neutral material on the Sporting Shooters Association website’.   This note seems to imply that the data published there, which is based almost entirely on ABS statistics and AIC research, was not to be trusted.   No comment is made about pro-control websites such as the Gun Control Australia website.  Check the websites yourself, and decide if this footnote is fair. 

Note also that no such accompanying note is made when Mouzos and Reuter use the Coalition for Gun Control opinion that “In New South Wales and Queensland, where the “shooters” lobby had been more politically prominent, the share may only have been 50 percent of the prohibited guns.”

When armed robberies using syringes and knives (but not guns) rose sharply, the AIC reported that ‘% of armed robberies using guns has fallen sharply’, and presented a % graph without any base numeric data (T&I 116). 

 ‘Small arms and light weapons’ control issues are used to appropriate war-zone social problems to validate Australian policy decisions on sporting firearms (T&I 104).

Graycar 2000 discussing ‘people who have guns’, ‘gun supporters’ – distinguishes and ‘others’ those who may disagree with his activist agenda. 


3. Social costs are emphasised but benefits are treated as negatives

This is legitimate as a viewpoint, but over a body of scientific work should be balanced.

Graycar (2000 and in Introductions to T&I papers) repeatedly says ‘guns are important as tools of crime’ but by implication not important as sporting or self-defence tools or possessions with symbolic meaning.  AIC research design focuses on trafficking, hospitalisations and deaths, but excludes the legitimate instances of self-defence, economic evaluation of sporting benefits, relative risk levels or the effectiveness of police procedures in approving individuals for firearm licensing.


4. Press releases and public activities are conducted with clear bias.

The undertone of disapproval for gun owners as people is unique in AIC publications and press releases.  Compare this to serial rapists, money launderers and domestic violence perpetrators who are be referred to dispassionately.

One exception: an AIC press release by Senator Vanstone against gay-hate murderers, clearly not a criminology publication.

Selective data was used in AIC press releases to support the 1996 buyback, despite suspiciously poor control in scientific terms (eg Lax Laws mean More Deaths, AIC press release 4 Nov 1996).  On the other hand, research news which exonerates sporting shooters is not reported in press releases. For example the strongly positive findings of Mouzos (T&I 151) were excluded from the press release.  False impressions are actively conveyed in media – eg Graycar (2002) implied that the Buyback is responsible for falling gun use in armed crime, when the evidence is clear that a pre-existing trend has continued without change. 

Criminologists’ quotes are reported in media which confirm or support false perceptions of the people affected by gun laws: for example,  Lincoln (1998) implying that the privilege of lawful gun ownership had until then been relevant to armed robbers and murderers, and by extension joining them with all Australians who exercised the privilege.


6. Interventions proposed are one-sided in intent

·      Interventions are typically not about utility.  More typically they are referred to in abstract  eg ‘policy options for reducing availability’(T&I 10) .  In contrast, options which might reduce misuse while not changing availability are not considered.   

·      No component of any gun law is examined for usefulness except with the assumption it may need tightening (eg Warner 1997).

·      Despite characterising the 1996 intervention as being against a ‘gun culture’, no cultural interventions are proposed; only mechanistic measures against ‘access’ targeted at people who obey the law .

·       Sunset clauses are never proposed to encourage examination of benefit.  

·      None of the many  pointless inconveniences in the law are pointed out.  Activism encourages such inconveniences, to raise costs to shooters and discourage participation.   

·      The cumulative effect of interventions is one-way; a ratchet of increasing barriers against lawful ownership.   

Monitoring Effects of the 1996-1997 Buyback

The effects of the Buyback have been carefully monitored by the AIC’s National Firearms Monitoring Program (NFMP).  Despite spin in the media (eg Graycar 2002), the actual observations reported by Mouzos and Reuter (2002) are:

1) No change in a downward trend in murders since 1989

2) No change in a downward trend in firearms use in murder since 1980

3) No change in a downward trend in firearms use in armed robbery since 1993

4) No change in a downward trend in firearms use in suicide since 1979 (but suicide is rising anyway)

5) No change in the rate of firearms use in attempted murder.

6) 40% to 60% of the banned guns are estimated to remain in the community.

7) Handgun murders rose significantly - related to drug gangs arming themselves in a few districts of the biggest cities. 

 Actual benefits are reported to be:

1) No firearm-related mass killings until the Victorian Police gave licences for seven handguns to a foreign Monash student.  Mass killings (defined as more than four victims) have occurred but by other means - the Childers hostel arson, the Snowtown murders and two family murder-suicides using exhaust gas. 

2) A moderate decline in firearms being stolen.  

It is worth noting that any good effects of the NAF may have arisen from low-cost components.  For instance, it is a small cost to issue shooters a licence after a simple police check for a violent record, and perhaps extremely effective.  In general, it would be appropriate to require a safety course and safe storage facilities at the owners cost. 

It did not cost money to bring state laws into alignment, but it has made it a lot harder for prohibited persons to obtain firearms. This is a clear increase in control, and may account for rising armed robberies being mostly committed with other weapons. 

The evidence does not support many other measures.  It is possible that the ban on military-style semi-automatic centrefire rifles acted in Australia's case to discourage copycat violence, as US media interest in the weapons is known to stimulate sales of the publicised types (Cramer 1993).  

However, the bulk of the buyback guns were not of this kind; they were mostly .22 rifles and shotguns.  The pattern of gun use in murder now suggests that the reason these are often used in crime is that they are so very common, not because they pose any particular danger to the community.  This accords with the routine activities model.  The high social negatives attached to buying back these arms were clearly not worth while.  

Registration, waiting periods and denying self-defence to be a ‘genuine reason’ do not appear to have contributed significantly to violence reduction. 

Actions against replica firearms and ‘soft air’ imitation arms have helped increase demand for real handguns on the black market.   

Contagion effects have been neglected in public discussion of the causes of public shootings, in an apparent rush to head off accusations against media violence However, Phillips (1980) has demonstrated the contagion effect so conclusively for suicides that media organisations generally follow restrictive guidelines on reporting them. 

I speculate that in Australia the media campaign and general public opprobrium against shooters may have created a climate in which these shootings became unfashionable even to the individuals who might otherwise identify with perpetrators – a negative implementation of the contagion effect.  

The evidence is that these killings are subject to contagion effects (eg Cramer 1993, Cialdini 2001), and that contagion effects are actually global; the worst massacres in Britain, New Zealand and Australia are often within a short time of others.  The Monash killings were at the height of publicity for the Washington sniper.  Unfortunately, Australia may have reached a break in a very desirable drought. 

It is also notable that suicides increased significantly in 1996-1998.  It has been pointed out (Tiffin 2002) that a number of suicides were known to be gun owners, in many cases using other means than guns.  Could the community hostility against shooters have actually contributed to suicide? 


Reporting of the Buyback Results

 The AIC was commissioned the ‘official umpire’ to report the results of the Buyback.  This has created a conflict of interest for the AIC.  It also meant that research by other criminology groups has not been initiated.   

Criminology Research Council grants have been very sparse; $30,000 to investigate the effects in Tasmania, by Professor Kate Warner (1999) whose close ties to advocacy groups are reflected in the quality of her science.              

Within months of the close of the Buyback it was asserted in public by the Government that ‘we could already see results’ after a release from an AIC staff member.  The Director then announced that many years data would be needed. 

The AIC trickled out results, and each time they were misrepresented in the media despite the basically objective work done by the authors. 

Then, on 2nd October 2002 the news broke: there was evidence the Buyback had saved lives!   

Editorials were launched, but there were no press releases by the AIC as we would normally expect.  The paper was not released, and the journalists did not attend the conference presentation.  The Director of the AIC seems to have invited the media with the promise of evidence lives were saved by the Buyback (ABC pers. com.).  He announced there had been falls in a number of violent crimes and no mass murders in that period. 

He did not advise the press that these were continuations of pre-existing trends, despite knowing it full well. 

He did not advise the press of the evidence for contagion effects in spree killings, despite the work done by the AIC showing he should have known (Carcach and Mouzos 2002).  

The AIC output is scraping to find ANY benefit, to the extent that with five years of post-buyback data they augment non-supportive data with blatant speculation.  In Mouzos and Reuter 2002: "Given only five years of postban data, one could not reject the hypothesis that it had reduced homicides by 10 percent."  Not gun homicides, but homicides - 35 lives a year or halving the  gun murder rate.  Far worse in Carcach,  Mouzos, and Grabosky 2002: "This sudden increase can be interpreted as evidence of a "contagion effect", with the significant response by the Australian Government contributing to the incidence of homicide resuming its long-term trend, rather than continuing to increase."   

Where is the evidence for that claim?   Contagion effects typically act for a few days only.   

We need to ask what measures are being used to evaluate the buyback.  If lives have been saved, will there be a benefit-cost analysis?  What are the indicators of success?  No targets or measures have been published, no experimental design offered. 

This author was advised by an AIC researcher (Mouzos pers. com.) that the media had broadcast what they liked after failing to attend the presentation; and by the ABC that they had fairly represented interviews and written materials provided by the Director Adam Graycar and the authors Mouzos and Reuter.

The AIC are now beyond opportunistic fishing expeditions, or implying that correlation is causation.  They appear to have moved to basic misrepresentation.   


Conclusion: Integrity Test For Australian Criminologists 

The evidence is that the Australian Government and the Australian Institute of Criminology will not evaluate the buyback by appropriate measures such as benefit-cost analysis.  

If the profession of criminology has any responsibility to the Australian people and to its own code of ethics, we need to ask:

·        What research could fairly measure the effects of the National Agreement on Firearms? Who will  do it?

·        What could we have done with over $1 billion to reduce violence in Australia?  

·       If the Government-appointed researchers are not going to report honestly, what should the  profession do? 

 At the 1999 ANZSOC conference, Don Weatherburn (1999) questioned the impact of criminology on public policy and public perception of crime, and attributed the poor showing to the profession itself. 

Australia has now paid up over $500 million to test Australian criminology's integrity. 


Published with the emailed approval of the Director of the Australian Institute of Criminology, who was unwilling to point out anything needing correction.

Corrections and improvements are welcome.  Please send them to thinkfocus@c-l-a-s-s.net .



 Alba, R, and Messner, S 1995.  Point Blank against itself. Journal of Quantitative Criminology 11:391-410.

Alpers, P and Morgan, B 1995.  Firearm homicide in New Zealand: victims, perpetrators and their weapons 1992-1994.  National Conference of the Public Health Association of New Zealand, Knox College, Dunedin.

Bellesiles 1998. Arming America: The Origins of a National Gun Culture. Vintage Books.

Betts, Katharine, 1999.  The Great Divide: The Politics of Immigration. Duffy and Snellgrove, Sydney.

 Carcach and Mouzos 2002.  The Mass Murder as Quasi-Experiment. The impact of the 1996 Port Arthur Massacre. Homicide Studies, Vol 6 No 2 109-127.

Cialdini, Robert 2001.  Influence: Science and Practice 4th Ed. Allyn and Bacon, pp121-130.

Cramer, C 1993.  Ethical problems of mass murder coverage in the mass media.  Journal of Mass Media Ethics 9.

Graycar 2000 – Crime, Safety and Firearms.  Prevention 2000 conference, Canberra

Graycar 2002 – public statements re buyback effects, 2 October , http://www.abc.net.au/pm/s692800.htm

Harding, R 1980 Firearms and violence in Australian Life

Kates, D. 1992. Bigotry Symbolism and Ideology in the Battle over Gun Control. PUBLIC INTEREST LAW REVIEW. http://teapot.usask.ca/cdn-firearms/Kates/bigotry.html

Kates, D., Schaffer, H.E., Latimer, J.K., Murray, G.B, & Casem,E.H. 1994 Guns & Public Health: Epidemic of Violence or Pandemic of Propaganda?  Tennessee Law Review 61, 513-596. http://www.constitution.org/2ll/2ndschol/58tenn.pdf

 Kellerman, A.I. et al. 1993. Gun Ownership as a risk factor for homicide in the home. 329 New England Journal of Medicine, 1084.

Kleck, G 1991. Point Blank: Guns and Violence in America. Aldine, New York.

Lincoln, R 1998. “Bond University criminologist Robyn Lincoln said the increased use of guns in crimes last year...was "because (criminals) knew the weapon was going to be withdrawn, they may have wanted to use their firearms before they lost the 'privilege'," she said. Armed Crime on rise - The Sunday Mail - Brisbane 18/10/98

Lott, J 1997.  More Guns Less Crime.

 Luna, E 2002.  The .22 calibre Rorschach test.  Houston Law Review 39, 59-131. http://www.law.uh.edu/Journals/hlr/downloads/39-1%20pdf%20files/HLR39P53.pdf

Mills, J 1993.  Data torturing.  New England Journal of Medicine 329, 16. http://content.nejm.org/cgi/content/extract/329/16/1196

 Mouzos, J  2000. Homicidal Encounters: A study of Homicide in Australia 1989-1999,  Australian Institute of Criminology Research and Public Policy Series No 28. http://www.aic.gov.au/publications/rpp/28/index.html

 Mouzos, J and Reuter, P 2002. “Australia: a massive buyback of low-risk guns” to be published in Ludwig J and Cook PJ (eds) “Evaluating Gun Policy: Effects on Crime and Violence”,  The Brookings Institution, Washington 2002

Talent, J 1989. "The Case of the Peripatetic Fossils," NATURE 338 (20 April, 1989): 613-15.

T&I – Trends and Issues papers of the Australian Institute of Criminology are available online at www.aic.gov.au .

Tiffin, B 2002. Pers. Com. listing suicides of acquaintances. (and look at this!)

Warner, K 1997.  Audit of implementation of APMC resolutions. 

 Warner, K 1999.  Firearm Deaths and Firearm Crime after Gun Licencing in Tasmania

Weatherburn, D 1999.  Criminology, Public Policy and the Fourth Estate.  ANZSOC Conference paper

Wright J.D.(et alia)  1983, Under the Gun: Weapons, Crime and Violence in America.