Wednesday, April 24, 2013

Academic refutes 'soft on cannabis' media claims

On 5 April the Daily Mail published this news article: The price of going soft on cannabis: Labour's experiment 'pushed up hard drug use and crime'.

The public letter reproduced below and here (PDF) from Dr Nils Braakmann of Newcastle University emphatically refutes the way that the Mail and a number of newspapers reported his research.

Contrary to the news reports, his research (which were only provisional findings presented at a conference, not yet published in a peer reviewed journal) did not show that reclassifying cannabis from Class B to Class C led to an absolute increase in cannabis use or crime. He says that he never looked at this, and the research results: "should not be interpreted as evidence that the declassification was “bad”. "

He goes on to say:

"...our estimates do not contradict potential aggregate crime reducing effects of cannabis depenalisation. As stated earlier, it is quite possible that the aggregate or regional effects of cannabis depenalisation are positive."

The Daily Telegraph piece making similar inaccurate claims for the research has now been removed, but the other reports mentioned in Dr Braakmann's letter, including by the Daily Mail, remain online. The report was cited again in the Mail on Sunday in this article from 21 April.

It was robustly challenged by Ewan Hoyle of Lib Dems for Drug Policy Reform on in an article titled 'The sloppy journalism that misrepresents cannabis use'.

Prof Alex Stevens of the University of Kent also challenged another piece of research discussed in the coverage, carried out by the Institute for Fiscal Studies, which relates specifically to the experiment with tolerant cannabis policing in Lambeth (London), and was cited by the Mail and the Telegraph as further evidence of the negative impact of cannabis depenalisation.

This story provides yet another example of how the need to support a particular policy perspective can distort objective science reporting. Whilst a common theme in science reporting generally, drug policy has a particularly poor record.   

Dr Nils Braakmann's letter
11 April 2013 
To the interested public, 
Some further comments on the press coverage and contents of my research on cannabis consumption, consumption of other drugs and crime. 
I am the lead researcher on the cannabis research piece that received (somewhat distorted) coverage in the Daily Mail and the Telegraph on Friday, April 5, 2013, and the Daily Star on Saturday, April 6, 2013.  
Several members of the interested public have contacted me to ask questions about the research in question. The following is a brief reply to these questions. It is also an extension to our initial reaction to the press overage published on on Friday, April 5, 2013. I also recommend the excellent comment and summary of our findings (as well as those by Adda, McConnel and Rasul) by Ewan Hoyle. 
First and foremost, this research is in its early stages and was presented in front of a professional audience at the Royal Economic Society annual conference on Friday. It was never intended to reach an audience beyond professional scholars at this conference. The paper is not publicly available, we never made any press release and we never talked to any journalist.  
My personal opinion is that research should only influence public policy or public opinion after undergoing peer review, not necessarily because all peer-reviewed research is correct, but because (a) peer review ensures that the work has at least received some outside scrutiny and (b) only after peer review and the final publication of a piece of research can we be sure that the respective study will not change anymore (of course, results can still be overturned by later research – and often are as human knowledge progresses).  
As such, I am deeply unsympathetic towards premature press coverage of work in progress. Of course, I understand the freedom of the press to cover any story in the public domain, but I think it is vital that the press and the public are aware that academic conferences are not press conferences. Discussion of early-stage academic work at conferences is a necessary step in the development and maturing of academic papers, but results are often still preliminary and work at this stage will regularly undergo changes. As such even competent and best-case coverage of such work always runs the risk of commenting on results that might not be there in the next revision of an academic paper. 
In this case my work has also been misquoted and misrepresented by sections of the press. While I would still prefer not having to discuss the results in the open at this point in time for the reasons explained above, I feel it is important to be clear about the things we do, the things we find and in particular the things we do not find: 
1) Contrary to press reports, we do not find any absolute increase in cannabis consumption, (a) because we never looked at absolute increases in cannabis consumption and (b) because as far as I know there has never been any absolute increase in cannabis consumption. 
2) We also do not find any absolute increase in crime, essentially for the same reasons. 
3) We also do not evaluate the 2004 declassification. Our interest was whether cannabis consumption might lead to increased criminal behaviour among consumers. The obvious difficulty here is to rule out that criminal behaviour causes cannabis consumption or that things like lifestyle changes cause both cannabis consumption and criminal behaviour. The 2004 declassification provides a relatively clean experiment to answer this broader question as it should only influence cannabis consumption but not the other things. Our basic idea is that the declassification and the associated changes in punishments has different effects on different groups of people: There are some people who did not consume cannabis under the old punishment regime but start doing so after the declassification. For some of these people, deterrence through the earlier tougher punishments mattered. We compare the behaviour of the previous non-consumers with the behaviour of people who already smoked cannabis prior to the declassification. The 25% reported in the press is the relative difference in the change in annual consumption between those two groups (note that this is slightly simplified and the actual piece is more technical). It arises as previous non-consumers have increased their consumption post-2004, while previous consumers decreased theirs. 
4) As pointed out by Professor Alex Stevens from the University of Kent there is a risk that these changes just reflect that consumption for non-consumers can only increase, while it can change in both ways for the other group. We are aware of this possibility and are currently looking for ways to investigate and possibly get around this issue. One reason why we present research at conference is to have an informed conversation with other academics about such problems and look for solutions. 
5) Again somewhat simplified: We find similar changes for (low-level) crime and behavioural problems. We do not find anything for cannabis consumption and the consumption of other drugs. 
6) We make it very clear in the paper that our study does not say anything about the overall effect of the 2004 declassification and our results should not be interpreted as evidence that the declassification was “bad”. To quote from the conclusion: “Overall, the estimates indicate that cannabis consumption may induce people to adopt a riskier lifestyle that goes hand-in-hand with low-level criminal activities, such as criminal damage, anti-social behaviour, fighting and victimisation. One should keep in mind that our estimates do not say anything about whether individuals are turned towards a life of crime – in fact this seems somewhat unlikely given the choice of criminal activities and the overall picture that emerges from the estimates (after all hardened criminals do not necessarily spend their time spraying graffiti). […]  
Finally, it should be stressed that our estimates do not contradict potential aggregate crime reducing effects of cannabis depenalisation. As stated earlier, it is quite possible that the aggregate or regional effects of cannabis depenalisation are positive as found in Adda, McConnel and Rasul (2011).”
I hope this clarifies a few things. I am very happy to have a further discussion with the wider public on these results, but I would suggest that this should wait until a point in time when this research is finished and published. 
Kind regards, 
Dr Nils Braakmann

No comments: