One in 58 statistics are wrong.

The recent story in the Observer that headlined the one in 58 figure for autism prevalence is no longer available on the Guardian Unlimited website. When I enquired about it I received this reply.

As there is a legal issue with the article that appeared on page 1 of the Observer on 8 July, the article has been removed from the website and digital edition. Therefore, I regret that we are unable to provide you with this article.

Please accept our apologies for the inconvenience this causes, however in such circumstances we stop providing articles as soon as a legal complaint is received.

But the internet does not work like that. The article and the probable reasons for the legal complaint  have been widely blogged and numerous references to it remain online. Unfortunately, many of these internet references are uncritical endorsements of the the original assertions in the Observer that autism rates had doubled and that two members of the research team, “leading experts in their field,” believed that the MMR vaccine was partly responsible for this increase.

The Observer supported these assertions with quotes from members of the research team at Cambridge University. Unfortunately for the Observer, both the assertions and the quotes were fabricated. Hence the swift removal of the offending article from their website, once legal proceedings were invoked.

This does not matter to true believers in the vaccine induced epidemic. They will repeat the 1 in 58 figure as an article of faith. But they are preaching to the choir. What about all the journalists out there who are read by the general public? The Observer story was taken up by newspapers throughout the world. Are they going to take down their stories or issue corrections? When the issue is finally settled will the Observer run a front page story with banner headlines correcting the misinformation in their original story? Probably not. But even if they do, 1 in 58 is out there now like a virus infecting all subsequent discussions about autism prevalence.

About these ads

6 thoughts on “One in 58 statistics are wrong.

  1. And did you expect them to behave differently???
    I sure wouldn’t…

    Just like Generation Rescue leading people to believe in the mercury=autism issue…..no retraction needed. Both of them got exactly the reaction they wanted.

  2. I’m not so sure 1 in 58 is wrong. It might just be pushing the envelope a bit far. Either way, I thought this was going to happen sooner or later, as I wrote back in February.

  3. Right, wrong or not even wrong, the 1 in 58 number should serve as a wake-up call to the “autism epidemic” screamers.

    At some point (and 1 in 58 is probably well past that point) a “disorder” that becomes so common no longer qualifies as a “disorder” and becomes … a normal variation.

    So, if 1 child in 58 is “autistic”, then “autism” is approaching the diagnostic relevance of “blonde”.

    After all, if so many people meet the criteria, it can’t really be a disorder.

    I think that the “autism epidemic” crowd may have bitten the poison apple with this one.

    Prometheus

  4. Interesting point, Prometheus.

    From a statistical perspective (based on psychometric approaches to behaviour), it’s the extreme outlying scores that are seen as evidence of ‘disorder’, in the sense of being ‘extremely abnormal’ (e. g., standard scores of 1 or 19 on a Wechsler Scale, although there is certainly a value-laden judgement used in the interpretation of these types of scores). Such scores are statistically rare: usually <1% of examinees would get such scores (each one being 3 s. d. from the mean, this would give a probability of 0.0026 of such scores – either one or the other – occurring), as far as I’m aware.

    The tendency, though, in psychometric approaches is to use one of two ‘normal limit’ definitions: a) within 1 s. d. either way of the mean; or b) within the middle-most 50% of scores.

    1 in 58 is approaching the 2-and-a-bit % of the extreme outliers…

    Just an observation.

  5. The internet is a real Pandora’s Box because of the speed of how information, accurate or inaccurate, gold or garbage, circulates and then persists in the cybersphere. Your virus analogy is quite apt.

    My belief on that basis is that it behooves news outlets and journalists to exercise responsibility, closely fact-check and step up sharply when a mea culpa is called for. They probably won’t, but wouldn’t accountability be a fine thing?

    Even if pragmatically it doesn’t make a darn, it would seem to be the ethical and correct thing for the Observer to publish a retraction with at least equal visibility to the original headline. Headline or frontpage would suit.

Comments are closed.