Linked to
Miaow miaow, croak
March 24, 2010 25 comments
In the latest drug scare to hit the UK, the news media is excitedly reporting that Cake mephedrone has been linked to several deaths of young people. Have coroners and toxicology reports determined that mephedrone was actually the cause of death in these cases? Er, no. All that is known is that the unfortunate users had taken some of it, along with other drugs (in what is usually known as a cocktail). So far, then, mephedrone is thus linked to these deaths in the same way that what you had for breakfast is linked to your being run over by a bus three hours later. Your breakfast could have been to blame (perhaps it was spiked with LSD, causing you to wander out abruptly into the road), but that remains to be empirically demonstrated.
As things stand, then, to report that mephedrone is linked to instadeath is not to report a causal link or a statistically robust correlation; it is simply to report the existence of, and thus to perpetuate, a media fiction that, as cast in the passive voice, also dissuades us from wondering about the identity and medical authority of the original hypothesizer (who is doing this linking, and on what evidence?). On the basis of this fiction, our guardians of health and sanity are demanding that users of mephedrone be criminalized forthwith. This kind of evidence-agnostic hysteria, of course, is standard procedure for official drugs policy, as David Nutt and his colleagues know all too well.
What have you linked to what on the basis of no more evidence than a post hoc ergo propter hoc fallacy lately, readers?
This kind of myth-making reminds me of Bagley’s derisive cynicism in How To Get Ahead In Advertising:
Priest: The bag was full of drugs, it [the newspaper] says so.
Bagley: The bag could’ve been full of anything. Pork pies, turnips, oven parts… it’s the oldest trick in the book.
Priest: What book?
Bagley: The distortion-of-truth-by-association book.
Idiot:
90% of heroin users started on cannabis.
Rational person:
100% of heroin users started on milk.
Unfortunately Nutt himself has used the “linked to” formulation in his otherwise reasonably sensible Guardian piece.
*spits mouthful of milk onto monitor*
I like Charlie Brooker’s article about this.
And in other news today:
Probably two separate issues (although they often go together): 1) the use of phrases such as “has been linked to” in a way that some linguists would describe as “lost performative”, “deletion”, etc; and 2) the science of the effects of substances on humans, etc.
http://www.dailystar.co.uk/new.....rugs-ring/
Also:
I didn’t quite get that, as I associate hysteria with strong belief, but “agnostic” implies noncommittal. Being agnostic about the evidence would seem a calm/rational thing to me.
Yeah, but you tell that to the Prague schoolboy who got hit by a bus because he thought he had the whole week to cross the road, Steven.
Of course it’s been linked to something – because I just linked it!
“Facebook ‘linked to rise in syphilis’.” (via @tankthetories.)
Yes, sure, skeptism about “linked” is important: who is doing the linking, on what grounds, and, crucially, what is the nature of the purported link?
But could you just explain how we get to “a causal link” and “robust statistical significance” without first, um, wondering about and isolating links like this one? Because isn’t this where epidemiological investigation begins?
And of course, I can understand your loathing of the “guardians of health and sanity” – they’re complete wankers, they are – and I can understand why one might be concerned about a rush to legislation; but if, left unreported by scientifically-illiterate journalists, there is what might be called a “signal” or some other concerning statistical blip (which may not be of “robust statistical significance”), suggesting that mephedrone alone or in combination with other drugs may be lethal, how long do you think mephedrone should be sold legally and freely to teenagers? Or anybody else for that matter?
Er, by investigating whether the ‘links like this one’ actually exist or are just made up, of course, eg and inter alia by, say, waiting for the Advisory Council on the “Misuse” of Drugs to report any evidence of possible harm before jumping on the ‘link’-reporting bandwagon and/or joining in calls for criminalization etc.
@sw
There are a number of unstated assumptions here:
1) Where there exists good (i.e. scientifically and statistically sound) evidence that a thing is “bad for you”, it is automatically a good idea to ban that that thing. The previously-linked Nutt article lists several practical reasons why banning mephedrone might do more harm than good. Remember also that “banning” is in no manner the same thing as “preventing the use of”.
2) Bad journalism has no side-effect other than on the precise subject matter being reported on. Or, that a journalist has no responsibility to consider the “bigger picture”. What of the responsibility to consider the cumulative effect of repeatedly misrepresenting the concept of evidence-based science?
3) Certain types of “linking to” where there is no good evidence are more valid than other types. It is precisely because there is no good evidence, that “linking to” mephedrone is just as valid as linking to (say) what one had for breakfast.
The broader point of correlation pointing the way towards an investigating into possible causation is valid – but that pre-supposes a correlation actually exists. That absolutely has not been demonstrated.
Ian @13:
Good point. Perhaps, in addition to sampling some substance, a recently-deceased person was recently put on a 60-hr shift in a low-paid soul-crippling job, and was threatened with legal action for non-payment of council tax, etc. Who’s going to investigate all the “links”. Epidemiologists? Everything is linked, man
Steven @12, er, yes, but that begins with observations that are not fully supported such as this one; I don’t dispute your concerns about the word “link” as I point out in my first paragraph, but as a matter of epidemiology, such unsupported “links” may be the first clue that merit further study and the development of good evidence. “Links” can famously be wrong; but spotting connections without fully understanding the causality and without having good evidence has produced a lot of “links” that have kept us alive. So you think that all “links” should be suppressed until they meet some objective criteria?
Of course, Ian, there are unstated assumptions – I live by them.
1) I disagree with your first sentence in point number 1 (which you seem to be ascribing to me?). For example, I do not think it is automatically a good assumption to ban fast food restaurants, although there “there exists good (i.e. scientifically and statistically sound) evidence that [they] are “bad for you””. “Bad for you” in a lot of ways. Bad for all of us, really. But there are times when the stakes change, when risk/benefit ratios change. I noticed in the news today that a company voluntarily withdrew a million baby-slings, because 3 babies were smothered in them: I don’t know if there is “good evidence” that the slings were properly used and that the babies did not die some other way (pre-existing cardiac or respiratory illness?), etc.: the “link” is strong enough (“smothered in them”) and the stakes are high enough to act decisively. This, of course, is a private company, and not the government. Instead of ascribing a false assumption to me, I would ask of you the question Steven also didn’t answer:
And, this reminds, me: nobody here has argued that there is any evidence that Mephedrone is is safe? Is it? Have you all just left out the evidence for its safety because it is so obvious?
2) I disagree with your first sentence in point number 2 (also somehow ascribed to me?). Bad journalism has a lot of bad side effects. Where do I purport to say anything else? (See, for example,
I have long supported and enjoyed Steven’s thoughtful critiques of the use and misuse of science in journalism.
3) Yes, I do agree with this first sentence and it would be an assumption of mine: “Certain types of “linking to” where there is no good evidence are more valid than other types.” And you’ve nailed something important here, where science becomes imprecise, uncertain. There are types of “linking” that are not supported by “good evidence” but that have such things as plausible mechanisms of action and testable hypotheses about the link that are generated from what we know. These guide us in how to look at the gazillion potential links that might be made in any one event; there is ample room here for discussion, but simply saying that a link without its very own body of “good evidence” is as shabby as any other link would end the scientific project here and now. (And maybe that wouldn’t be a bad thing?)
sw@15 — quite right, I didn’t answer your question, because (as usual on unspeak.net) this post is not about what I would do if I were in charge of policy, but about a sloppy and misleading use of language in the media that tends to degrade the quality of debate about policy, a point with which you appeared at #11 to agree, albeit grudgingly, so I am not quite sure what exactly you are disagreeing with in what I wrote.
Do feel free, though, to continue disagreeing with what I didn’t write. I probably don’t agree with it either?
sw @ 15:
Yes, but prior to having “good evidence” there are a lot of connections to spot. Scientists don’t have infinite budgets – I suspect that often (prior to having “good evidence”) their choices about where to focus attention may be influenced by the kind of thinking which is always “linking” tragedies to substance-ingestion of a certain kind (rather than other factors, which may, in the absence of good evidence, be just as worthy of consideration). Prior to good evidence.
sw@15
On the first point (my assumption of your hypothetical assumption): it seemed (and still does seem) to me that the whole thrust of your post was to argue that there might be a case for “banning” mephedrone – or at least no longer “allowing it to be sold legally and freely”, which I take to mean the same thing. It also seems to me that, should it be the case that mephedrone is in fact dangerous, you would consider a “ban” to be a valid response. Does that sound reasonable?
As to your repeated question, I’ll give the same answer I gave the first time – there is no “statistical blip”, therefore no action is required. Further, if there was such a blip, my sympathy would be with Nutt – further investigation would be required as to the benefit (or otherwise) of regulatory action.
On the second, I think I went too far. It was my intention to point out that there are good reasons why a journalist might choose not to print something. You were supporting the supposed benefit of a journalist publicising something for which there was no evidence (yet), my comment gave a reason why he shouldn’t.
On plausibility, I agree – it’s a valid approach. I have to say though that I haven’t come across much discussion in the mainstream media along those lines – other than the usual mindless “OMG IT’S A DRUG INNIT!!1!” blather.
some lady from the east yorkshire police was on the radio last week.
“There are two reasons” she said, “for this campaign. One, to gather any new information regarding this case and.. one, to raise awareness amongst the young people who are thinking of taking this drug, which may have been responsible for these tragic deaths”
I think what has been quite an open campaign by the police who are dealing with the deaths has really been aimed at dissuasion. At first I bucked at the knee jerk commentary in the national press but then when I heard the police on the radio and telly who had obviously had little time to plan this campaign I felt a genuine pang of sympathy for them. They don’t want people to die, so they’re muddying the lines and definitions just to get the message out there.
If anyone can be arsed to check the first day’s satements out of east yorks constabulary- the ones i heard on various radio stations that day after the two lads died, you’ll find them to be far less misleading and far more sincere.
Ian @18:
Before I get to point 1 – I agree with you fairly strongly about point 2. There are times when a journalist may responsibly refrain from reporting something. There are also ways of responsibly reporting on the Rumsfeldian world we live in, in which there are known knowns, known unknowns, and unknown unknowns. Avoiding unspeak, identifying (most) sources, ensuring that the journalist is literate in the field he/she is covering and can thus provide adequate contextualisation – these are the obligations of the responsible journalist, I think. Choosing when to report or not to report should therefore be a matter of journalistic judgement and responsibility. I’m glad, for example, that there were journalists willing to talk about the links between human activity and global warming before the evidence became as strong as it is today.* And, in this specific case, I am not convinced that journalists should have refused to report on this matter. Alerts coming in from casualty departments of morbidity associated with the use of this drug, as well as possible lethalities recorded elsewhere, may not constitute “evidence” – and we should be careful not to get too confident or unspeaky in our use of “evidence” – but they may constitute a – shall we say – reportable pattern.
And I don’t have a disagreement with point 3. The teeth-gnashing, wailing, brow-beating, nipple-plucking, spittle-splashing rage accompanying issues like this are rarely edifying.
So let me just try to respond to point 1.
My main thrust was not intended to be about whether mephedrone should be banned or not – although, yes, to answer your question, I do consider banning to be a “valid response” (depending on the circumstances). I’m a fan of product safety: I am grateful for many of the public health interventions that have, through the force of law, imposed limits on what people can buy and sell, although I accept that there are moral, political, social, cultural, religious, and even medical repercussions that might be, shall we say, problematic. And I would add the following: the decision to wait is not without consequence. Calls for patience until all the “good” evidence is in so often comes from companies who want to continue producing and selling products that are increasingly understood to be dangerous, even deadly – cf history of lead-based paint, fossil fuels – and by the people who want to keep using these products. Happens all the time. Why should I think it isn’t happening here? So what is the consequence of waiting until all the evidence is in? Waiting, by the way, is not innocent (though it’s nice to feel that way); it’s not a disinterested position in which you have no responsibility for the consequences of waiting; waiting is not a suspension of policy, but a policy. And, in any case, why should the demand not be turned on its head: “Okay, show us this shit is safe”?
*Not that it’s done a lot of good.
sw @ 20:
I think you’re setting up a false dichotomy (also with your global warming example). It’s not a question of “choosing when to report or not to report”. It is reported (eg by the wires), but if it’s not much of a report (eg: “untested hypothesis about plausible mechanism allegedly linking drug to woman’s death”) then it doesn’t make the headlines. The responsibility is to be accurate, not to be right about when to suppress information.
“Product safety”? “Public health interventions”? Interesting ways to frame the issue. I’m reminded of the billions spent by drug companies to derive, from cannabis, a substance which reduces pain without making you feel good.
Bruce @21:
I continue to agree with Ian’s sentence (assuming my riff on it @20 demonstrates an understanding of what he is saying):
“Product safety” and “Public Health Interventions” are indeed ways of framing an issue; I too find them interesting.
sw @ 22:
I don’t think anyone was arguing the case for suppressing such a story outright (one side of your dichotomy) but for not printing something misleading about it.
The Wonkette blog in the US today did an excellent job of linking the attempt of Sarah Palin to become vice-president to the collapse of the world’s economy. We can blame her for everything:
sw, what do you mean by ‘safe’? As far as I’m aware, no inquest has confirmed that mephedrone was the cause of the seven deaths it has been ‘linked to’ so far. On the other hand, there are some 6,000 alcohol-related deaths and 100,000 tobacco-related deaths per year, not to mention deaths from other relatively risky activities that don’t make the news because they are commonplace and therefore accepted as a matter of course.
Well, I like John Stuart Mill’s principle that “the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others”. But even if you think, unlike Mill, that “his own good, either physical or moral” is “sufficient warrant” to exercise power over someone, broadly speaking in the UK I think we (ought to) have a principle that the person who wants to make an activity illegal* must show that (1) doing so is necessary and proportionate and (2) that the consequences of making it illegal do not outweigh the consequences of doing nothing (or doing something else about it).
* or a substance a ‘controlled substance’ (there’s a misnomer), etc.