Tuesday, October 16, 2012

Smoke and mirrors at the BBC

"Oh goody, another survey. It's got numbers in - no, not just numbers, statistics - so it must be true. Plus it supports what I want to believe, so I won't look at it too closely."
And so, with a flourish of a mouse, the BBC website finds itself carrying this nonsense:
Smoking in the car breaks toxic limit (BBC Health, 16 October 2012)
The article, by Health Editor Michelle Roberts, claims that "smoking in the car, even with the windows open or the air conditioning on, creates pollution that exceeds official "safe" limits. Roberts got the story from a journal called Tobacco Control, which is run by the BMJ (formerly the British Medical Journal). All fairly respectable, and I daresay I'd be tempted to take the evidence at face value if the BMJ claimed that homeopathy is bunk. But a glance at Tobacco Control's website shows that it's already nailed its colours to the mast. Tobacco Control isn't science, it's advocacy. Any journalist should treat it with scepticism.

For those who don't know, BMJ is wholly owned by the British Medical Association, which recently called for a ban on smoking in cars and later had to admit that it had relied on sensationalist, long-discredited data. One might also criticise them for authoritarian control-freakery, but that's a subjective judgement and a journalist should rise above it. But whatever one's views, a good journalist should immediately ask themselves what agenda, if any, is being pursued.

So, let's have a look at those claims:
  1. "A Scottish team who took measurements during 85 car journeys found readings broke World Health Organization limits"
  2. "The researchers … analysed air quality data during a number of journeys ranging from about 10 minutes to an hour in duration"
  3. "In 49 of the 85 journeys in total, the driver smoked up to four cigarettes"
Wait, wait, hang on a minute. None of those journeys lasted more than an hour and yet in 58% of them the driver got through FOUR cigarettes. There's no evidence that people smoke much in cars with kids in them, yet this survey had the drivers practically chain-smoking. That on its own is enough to dismiss the entire study. All we've learnt so far is that the University of Aberdeen, which carried out the research, is a bad place to study science.

Mind you, the article says "up to four". So it could be as low as zero. You can try to get something better out of the abstract if you like, but that's just as vague. Finally:
4. "During these 49 smoking journeys, levels of fine particulate matter averaged 85µg/m3, which is more than three times higher than the 25µg/m3 maximum safe indoor air limit recommended by the World Health Organization"
Pretty damning, huh? Since the headline cites the WHO's recommendations, we ought to check them. I've provided a link so you can check them yourself, but the relevant paragraph says this, with the number the researchers relied on highlighted:
Guideline values
PM2.5
10 μg/m3 annual mean
25 μg/m3 24-hour mean
PM10
20 μg/m3 annual mean
50 μg/m3 24-hour mean
Now, this is particulate matter, and I have to take it on trust that the harmful bits in tobacco smoke fall into that category. But WHO is quite clear: this is not occasional doses or a safe limit. The WHO's guidance is that average concentrations of small particulate matter should not exceed 25 microgrammes during the course of the whole day. Isolated or occasional readings above this limit are irrelevant. So the headline isn't just misleading: it's wrong.

What this study actually shows is that, if you lock yourself and your child in a car for 24 hours and chain-smoke constantly, it probably won't be very good for the child. It says nothing about the real world. Put it another way:
"We have created an entirely fake and unrealistic scenario that has produced results that bear no relationship whatsoever to the real world but which we think should be used as the basis for legislation."
Moral: You can always get the numbers you want if you skew the research, and you can always find a journalist who's gullible, lazy or as unprincipled as you.

1 comment:

  1. Thank you Pat for de-bunking bollocks (again)

    ReplyDelete