‘Licence to blog’ agent Maxine Vaccine focuses her telescopic sights on unreliable statistics and the dangers of passing them on.
Mark Twain, the patron saint of the anti-bullshit movement, said there were three kinds of deception: “There are lies. There are damned lies. And there are statistics.”
It’s not literally true that statistical data can mean whatever you want them to. After all, numbers are only numbers. But a number – regressing for a moment to my misspent student days, which I wasted in studying – is only a signifier, easily detached from its referent in the real world. Between the raw data and the qualitative conclusion lies a world of potential distortion.
For example, take this statement: 50% of NHS trusts are performing below the median standard. Shut them down, right! Except that in any set of values of a single variable, 50% are below the median value.
Try another one: 20% of company CEOs in a survey of the UK pharma industry said they wished they’d taken up kitten breeding instead. That’s a disturbing picture of the industry... but faintly less so if the 20% means one CEO out of five who answered a leading question on Facebook at 3 am, when he was alone in a hotel room and the minibar was empty.
Another problem with statistics is that they induce meaning-blindness in the audience. They sound so unquestionable, so ‘hard data’, that the possibility of their being poorly sourced, fictitious or just plain wrong doesn’t occur. So when David Cameron announced in 2010 that 50% of girls on housing estates were becoming mothers while in their teens, the power of the statistic stopped most of his listeners thinking about the plausibility of what they were hearing.
The next day, in some embarrassment, Cameron declared he’d been incorrectly briefed: the correct figure was 5%. But how did he not question it at the time? Perhaps he thought something along the lines of: Housing estates... single mothers... promiscuity... feral children... what do you expect? But more probably, he didn’t even process it that far. The weight of the number stifled its own meaning.
The pharmaceutical industry may be experiencing a faint sense of relief that the first medical shock story of 2012 concerns breast implants rather than drugs. But it’s the relief of a soldier who sees the next platoon drawing sniper fire rather than his own. There isn’t much comfort in it. Just remember that a couple of years ago, a certain global pharma corporation was charged with publishing a fake clinical journal in Australia in which its product was praised by non-existent clinicians. If true, that’s chutzpah on a cosmic scale.
Both cases come down to statistics. In the fake articles (if they existed), the holy writ of spreadsheets and pie charts made the lies invisible. And in the case of the PiP breast implants, where industrial-grade silicon intended for use as mattress filler was substituted for medical-grade silicon intended to make women look even better than their natural selves, the company’s marketing director told sales executives: “Sales are more important than what the shells are made of.” Why? Because the lies could be hidden in perfectly formed pyramids of data, safe, protected – until they ruptured and the whole structure collapsed.
Moody statistics (including misinterpreted or misrepresented data) are like counterfeit banknotes: bright and new, easy to fold, easy to spend, but they will come back to haunt you. So keep your facts straight... and keep your figures even straighter.
Note: the views of Maxine Vaccine are not necessarily those of Pharmaceutical Field.