Of Iain Dale, Boris and Gateway Drugs
Posted by rantingkraut on September 23, 2007
During his recent one to one with Boris Johnson, Iain Dale stated that Marijuana was a gateway drug. According to Dale, this can be inferred from the fact that one would be unlikely to find a single Heroin addict who had not consumed Marihuana previously. Boris duly conceded that taking drugs was a stupid idea and expressed his scepticism over taking a more relaxed attitude to Marijuana.
So, is Boris loosing his libertarian instincts? Are potheads doomed? Let’s look at the gateway drug issue first.
Iain Dale: Barking up the wrong tree
Iain Dale’s assertion amounts to saying that, once you are known to be a Heroin user, it is virtually certain that you have consumed Marijuana previously. He effectively guesses that the probability of current Heroin addicts having used Marijuana is one. As far as guesses go, this is not an unreasonable one, but is this the kind of information we should be interested in?
To answer this question, we should look at some statistics. The Department of Health knows a lot about drugs, and they tell us the following:
“Cannabis was the most frequently reported illicit drug used in the last year, used by 13%.
One per cent had used heroin in the last year and 1% had used cocaine. In total, 4% had used Class A drugs in the last year.” (Source)
So the probability of a randomly selected respondent having used Marijuana  is:
P(M) = 0.13
while the probability of such an individual having used Heroin is:
P(H) = 0.01
If, as seems likely, these data don’t change dramatically from year to year and if Marijuana is a gateway drug leading on to Heroin, then the probabilities of being a Marijuana user and that of being a Heroin user should be similar. They are not: a randomly selected respondent is 13 times more likely to be a pothead than a junkie.
Iain Dale’s guess, that the probability of heroin users also having used Marijuana is one, can be expressed formally as:
P(M|H) = 1
i.e. the conditional probability of an individual having used Marijuana given that he is identified as a Heroin user is one. This guess still seems plausible, it is just not the information we should be looking for.
Once we accept that P(M|H)=1, we also know that, which ever way you look at it, the probability of having used Marijuana and Heroin is: P(H^M) = 0.01. We know that because we know that 1% of respondents used Heroin and we accepted that all of them smoked Marijuana before. Formally:
P(H^M) = P(M|H)P(H) = (1)(0.01) = 0.01
which, incidentally, is just P(H). Of course, the concept of belonging to two categories simultaneously is inherently symmetric: P(H^M) = P(M^H) and we are able to argue by analogy that:
P(H^M) = P(H|M)P(M) = P(H|M)(0.13) = 0.01
I have no data, or guesses, for P(H|M), the probability of being (or becoming) a heroin user once you are known to have consumed Marijuana. We can, however infer from what we have got that :
P(H|M) = 0.01/0.13 = 0.077
In other words, of those known to have consumed Marijuana, 7.7% are predicted to move on to Heroin. This probability is clearly higher than the 1% incidence recorded for the population at large. It is also substantially below the 100% implied by Iain Dale.
Whether or not Marijuana is a gateway drug depends on more than crude probability calculations. How important this 7.7% probability is compared to the general incidence of Heroin use is a matter of subjective judgement. That Iain Dale looked at the wrong conditional probability is, however, a pretty clear cut case.
Boris: loosing his libertarian convictions?
Boris has so far been one of the more libertarian Tories, so hearing him say that he is against decriminalising drugs sounds like a change of direction. One reason he mentions for this is the supposed seediness of parts of Amsterdam where the approach to drugs is more tolerant.
A number of things need to be said here, because that point gets made again and again. The last time I looked, Amsterdam was a city in the Netherlands, not an independent state. National drug laws should therefore apply as much to Amsterdam as any other place in the Netherlands. I am no expert on regional crime statistics, but I would expect Amsterdam to have more of a drugs and crime problem than Zandvoort or Gouda, just like London will probably have more of a drugs and crime problem than Bridge of Allan or Arundel.
This brings us to the second question of just how bad Amsterdam or the Netherlands really are compared to the UK. I don’t have city by city crime statistics available, but Mercer, an HR consultancy publishes a ‘quality of living’ ranking for a number of cities in Europe. In this, Amsterdam is on place 13, London on place 39 while heroin infested Zürich comes top.
The website nationmaster.com, publishes comparative crime statistics by country. Looking at drug offences makes no sense in this case: if a drug is legal in one country but not the other, differences in crime statistics may show just that . However, looking at burglaries per capita, the UK makes place seven, leaving the Netherlands on place 20. The picture is similar for robberies, with the UK on place 8 and the Netherlands on place 16. Drug related crime is likely to be part of both categories, so it would conventionally be expected to increase with the incidence of drug addiction.
The Netherlands and Amsterdam then aren’t doing all that badly, Amsterdam does seem to attract drug tourism though. It has the reputation, the counter culture and the nightlife to attract it. This is something which London could well emulate.
Drug liberalisation is a social experiment. There is therefore a case for gradualism and regional experiments. The question now is whether London is the best place to pioneer such a regional experiment. The parallels to Amsterdam suggest that London might attract a similar amount of drug tourism, and if this is something Boris is keen to avoid, he would be tempted to leave the first move in liberalising drugs to a less conspicuous place.
So how authoritarian has Boris revealed himself to be? Power corrupts and, even if it didn’t, running a city requires more pragmatism than libertarian ranting. Considering this and the fact that he proceeded to a conciliatory message advocating low key enforcement of drug prohibition, one probably shouldn’t make too much of this point.
 These data refer to under 25 year olds in 2003.
 Of course we could note, in general terms, that: P(M|H)P(H) = P(H|M)P(M) so that P(H|M)= P(M|H)P(H)/P(M), which is just a form of Bayes’ rule and returns the same result.
 For what its worth, the UK (rank 12) would indeed record more drug offences than the Netherlands (rank 20).
Sorry, the comment form is closed at this time.