In an article yesterday the HuffPo posed the question ‘Will Cell Phones Prove as Cancerous as Cigarettes?‘ and answered it in a truly dizzying game of Logical Fallacy Bingo. There have been so many fearmongering stories about mobile phones that I lose count – apparently the radiation they emit causes all sorts of cancers, makes your nuts drop off, stops you sleeping…

The first canard of quite a few in the piece is that Motorola released the house brick posing as a phone – the DynaTac 8000X – 30 years ago, and it took 30 years to establish a causal link between cigarettes and lung cancer (it didn’t), ergo are we about to discover the same of mobile telephones?

There are two ways to look at this: basic plausibility / lab experiment, and epidemiological studies to see if there’s a correlation between mobile use and cancer.

Plausibility

First off, radiation from cell phones is non-ionising radiation and so is too weak to have any plausible biological effect. Unless some as yet undiscovered phenomenon is uncovered – but this is unlikely and here’s why:

All electromagnetic radiation is made up of photons, and their energy is determined by Planck’s law: the energy of the photon increases as the frequency increases. Yellow light is around the middle of the visible spectrum and has a wavelength of about 600nm which equates to a frequency of 5×1014 Hz. A mobile phone uses a frequency of 1×109 Hz – so a photon radiating from a mobile phone is about 500,000x less energetic than a photon of yellow light. For comparison a photon from an x-ray machine has a frequency of 3 x 1017, or energy a thousand times larger than that photon of yellow light. The nuclei of atoms can only absorb the most energetic photons (gamma rays – irrelevant here) and only ultraviolet photons (which have higher energy than visible photons) have the required energy to break chemical bonds – which explains the link between excess UV exposure and skin cancer. (But this is also responsible for the formation of vitamin D by photolysis of 7-dehydrocholesterol when skin is exposed to UV – visible light can’t do this as it is reflected by the skin – you win some, you lose some). So, basic ‘O’ Level scientific plausibility is absent from the get-go.

Epidemiology

“Cell phone radiation isn’t good for you. But is it bad for you? If it is bad, will the bad turn into fatal? And even if cell phones proved harmful to our health — then what?”

Slippery slope fallacy, begging the question, non sequitur… And that’s ignoring the hint of false dichotomy and straw man in the first bit: “cell phone radiation isn’t good for you (wrong) but is it bad?”

The reference cited for ‘could cause cancer’ relates to a May 2011 press release from the International Agency for Research on Cancer (IARC) in which it classified cell phones as Category 2B, which is “possibly carcinogenic to humans”. Check the Wiki link for some of the many deep flaws in this organisation’s credibility – but the science behind that release was roundly criticised at the time (and seized upon by all the usual Tin Foil Hatters). Since then it has been pretty widely disproven. Also for interest 2B contains about 300 substances including talcum powder and pickles.

Rather inconveniently for HuffPo’s argument some very large datasets have been analysed looking at cell phone usage and – for example – brain tumours. The logic is quite simple and easily tested: mobile phone usage has increased massively over the last two decades so were there a link one would expect to see a rise in malignant brain tumours. No such rise has happened. Even if there were, correlation is not necessarily causation. Another central nugget of the argument posited is that “no one… has been able to conduct a controlled study over the decades it would take to prove a definitive cause-and-effect”. Again, Utter, Utter Bollocks (µ²B) – there have been many. These studies come in two flavours (with apologies to any epidemiologists reading this for the gross oversimplification): case control studies and cohort studies.

In case control studies people with brain cancer answer a questionnaire about their cell phone use and the answers are compared with a control group (who don’t have brain cancer). You then work out an ‘odds ratio’ of developing brain cancer. These studies have some serious flaws, the main one is recall bias; you are relying on the accuracy of people’s memory. A more robust measure is the other flavour: cohort studies. This is where you follow a population over time and see how many develop brain cancer. This is far more accurate as you’re looking at records rather than relying on folks’ memories. There have been some pretty big studies; a Danish one looked at over 400,000 people, some of whom had been using mobile phones for over 20 years. No cancer risk was found.

“So, my phone will kill me. Eventually. And…?”

This is the logical fallacy of Begging the Question: the assumption that a statement is already proven without any logic or proof to show the statement is true in the first place. This is NOT the same as raising a question (“this begs the question…”)  – which is what many people think begging the question is. There’s a great example here regarding popularity of exclusive clubs: “the reason there’s such a big demand is because everyone wants to get in them“.  Or there is a big demand because there’s… errrrr… a big demand.

OK. The article then goes on to make a false comparison with lung cancer’s correlation with smoking:

“I unhappily suspect we will reach the same eventual conclusion where cell phone use is concerned… Replace the word “smokers” and “cigarettes” with “cell phone users” and “cell phones” and this human behavioural observation could be earily precient.”

Sic erat scriptum and sic! again on those last two words!!

This really is Utter, Utter Bollocks (µ²B) of the first order. It makes the assumption that because it took time for the causal link between smoking and lung cancer to be accepted, the same is true of cancer and mobile telephones. So there are several false assumptions here; first it begs the question there is a link between mobile phones and cancer (no evidence) and further that epidemiology hasn’t improved since the 1930s. Which really is Utter, Utter Bollocks (µ²B): it has, as evidenced by the studies the writer couldn’t be arsed to look up. We’ve come a long way since Dr John Hill and cholera and even further since Richard Doll and Archie Cochrane. I think the (non-) argument above is best summed up with the proverb:

If my grandma had bollocks she’d be my granddad.