Monday, April 16, 2007

The Rays from Outer Space Strike Canary Wharf

The Sindy has been getting a lot of blogosphere points for this article, which alleges that a mysterious ailment of bees is caused by "radiation" from mobile phone networks. Nowhere is it mentioned that an identical, and unexplained, condition has been documented as early as 1896, before the invention of radio and a hell of a long time before mobile phones were common.

It gets worse, though.
But an official Finnish study found that people who used the phones for more than 10 years were 40 per cent more likely to get a brain tumour on the same side as they held the handset.
No, it didn't. The study found that people who used the phones for more than 10 years were no more likely to get a brain tumour than anyone else. But, in the event they did, they were 40 per cent more likely to report that it was on the side they held the phone - or at least to think that they held the phone 10 years ago on the side of their head the tumour was on, as this was self-reported and clearly subject to confirmation bias.

There is no mention of the Royal Danish Cancer Institute report, the biggest (n=420,000) and longest (25 years) epidemiological study ever undertaken into the subject and the only one to use network operators' billing data, rather than self-reporting, to find out how much the patients used their phones over a period of 20 years. So what did they find?
We found no evidence for an association between tumor risk and cellular telephone use among either short-term or long-term users. Moreover, the narrow confidence intervals provide evidence that any large association of risk of cancer and cellular telephone use can be excluded.
You can read the paper in the Journal of the U.S. National Cancer Institute.

All in all, an exercise in bad newspaper science worthy of the Daily Hell - even the headline sounds Mailesque. Why? According to media sources, the Indy recently hired two chief sub-editors from the Mail. Do Associated Newspapers Staff Wreck Your Journalistic Standards? - not a bad headline, eh?

Update: I've managed to find the German study they referred to (pdf). It consisted of plonking a DECT cordless phone base station inside the bees' hive, either with or without homemade shielding, then catching some bees, marking them, and counting how many returned within a given time period. It's not clear from the paper whether they put anything in the control group's hive, which raises the question of whether the results are an artefact of the experiment.

In two rounds of tests, they got one marginally significant result and one nonsignificant. Apparently, 54 per cent of the bees with the DECT station on returned on time, compared to 63 per cent without, in six tests. Reading some of the other papers, the initial hypothesis appears to be that "GSM TDMA time slots change over at about a frequency of 217Hz, and that's nearish one cycle in the bees' dance, so it must be connected". They do mention that the pulse cycle in DECT is 100Hz, but do not discuss the fact that this isn't the same.

Neither do they mention that timeslotting doesn't mean there is no signal from the BTS, just that it communicates with a different user. Ho hum.


DBK said...

Thank you, Alex. I was curious about this.

You just got yourself a blogroll link, my friend, for being objective and smart.

fido said...

Well I tried to chase a bee out using my mobile phone, gave up and used fly spray in the end.

kostenloser Counter