The problems with one data point

Is this proof that WiFi radiation is dangerous? From ABC News.

A recent report, shared around the World, described an experiment conducted by a group of Danish girls that seems to demonstrate that WiFi radiation stunts plant growth. This report has supposedly caused quite a stir in the scientific community. I write “supposedly” because, to be honest, I have not seen much chatter over these results outside of the health section of mainstream media reports and some online channels that already want us to believe that WiFi signals are dangerous to humans.

This group of five 9th-grade girls from Hjallerup, Denmark planted seeds of a garden cress herb into six trays, and set them in a room without a WiFi router. Similar seeds were planted into six other trays, in another room near two broadcasting WiFi routers. After 12 days, the seeds in the trays away from the routers grew normally, while the seeds in the trays next to the routers turned brown and died. The biology teacher who mentored the girls insists that the students kept the cress seeds in both groups sufficiently moist during the whole experiment, and the temperatures were kept constant to offset the effects of heat from the routers.

This is the kind of story that the media loves to report. A scientific experiment that simply confirms what is already known is not newsworthy. But five teenaged girls challenging what scientists (cue picture of old, white, bearded, smug men) have been telling us to believe? Now that’s a story. It plants a seed of doubt (pun intended) in the minds of those who think that WiFi usage is safe. It creates debate and conflict, a bit of a David injuring Goliath narrative, which is exciting.

In the end, though, a single experiment can never be used to draw wide-ranging conclusions.  One child who was born with HIV but is now clear of the virus does not mean that a cure has been found.  One research group detecting particles travelling faster than the speed of light does not mean that Einstein was wrong. When a result challenges widely-accepted knowledge, it is most important to take all steps to ensure that there isn’t another explanation, and that the evidence truly points to the unexpected conclusion.

Some logic should be thrown into the discussion.  If WiFi keeps plants from growing, how do gardens thrive in cities and suburbia despite 24/7 bombardment from WiFi radiation in homes and businesses around it?  These girls explained that they conducted the experiment after noticing that they had trouble focusing in school after they slept with their cellphones near their heads overnight.  Is it possible that on those nights, they spent more time online just before going to bed, and the artificial light and stimulation causes them to have a bad night’s sleep?

I know, I know – some people would expect me to answer in this manner. After all, these girls found a result that contradicts my past writings that exposure to WiFi radiation is not dangerous. Why can’t I just accept their results, instead of poking holes in their experiment?

Because that is absolutely necessary to ensure robust scientific research. I’m not saying that their results were fraudulent; they may even be onto something that most scientists have missed. Of course, this was a science fair experiment, so the group likely only had the time to do the experiment once.  But before going to the media with these results, the least that could have been done was to see whether the findings were reproducible and predictable.  It is preferable that some of these trials be conducted by other researchers in other locations, so that any inadvertent bias introduced by the girls can be avoided.  If they expected, or desired, for the plants exposed to WiFi to show lesser growth, they may subconsciously offer a bit better treatment to the plants in the control group, not exposed to the radiation. (In case you think this is a comment on their inexperience, I can assure you that even the most experienced scientists fall into these traps.) In particular, it would be ideal that the experimenters not know which seeds are exposed to the WiFi radiation, to ensure that those plants are not treated any differently.

Am I being overly harsh on these girls? After all, I could just be bitter that they put forward a result that contradicts what I’ve been writing for years – that WiFi radiation is not dangerous. But in fact, I want to help them practice strong, robust science. I would like to see them apply the rigour necessary to be certain that any conclusions are supported by the evidence. If they want to access the impact of WiFi radiation, they must isolate the effect of only that radiation, making certain that any other variables – temperature, exposure to light, humidity, condition of the seeds and soil, etc. – cannot explain what happened. Furthermore – and this is where many new medical treatments stumble before reaching the finish line – we must be certain that the proportion of people healed by the proposed treatment is significant, not just a random occurrence. For this reason, pharmaceutical trials are done on hundreds of people, not just on a few of the researcher’s friends. This is also the reason why placebo treatments are given to a portion of the population, with neither the doctors nor the researchers nor the patients knowing if they have the actual product being tested or the placebo.

All other variables and potential explanations must be accounted for, so that the final conclusion – that a change in X is clearly caused by Y – can be stated with complete confidence.  It is perhaps the most frustrating aspect of science for young researchers.  It gets very depressing when every seemingly positive result gets torn apart by supervisors, colleagues and other scientists.  But that robustness, that continuing questioning of the results, is essential when applying the scientific method.  It is much better for questions to be asked and questionable conclusions addressed in the privacy of the lab or office, rather than being exposed after being broadcast around the World.

Be Sociable, Share!

One Comment

  1. Comment by Paul:

    The lack of any kind of blinding is, in itself, enough to completely invalidate these results.

Leave a Reply