Bees, Cell phones, and BS (again)

The interwebs are all abuzz (ha!) with a new report that cellphones might be responsible for the losses in honeybee populations. ResearchBlogging.orgSpecifically, the news stories reference this paper:

Ved Parkash Sharma and Neelima R. Kumar (2010). Changes in honeybee behaviour and biology under the influence of cellphone radiations Current Science, 98 (10), 1376-1378

It’s a peer reviewed paper…but frankly, I think the journal editor has a lot of questions to answer about letting this paper be published.

There are many things wrong with this paper that make me discount all of its findings. A short list:

1. They tested only FOUR hives–and each hive had a different treatment. This means that there was NO replication for their treatments!!
Basically, they put cell phones in two hives, cellphones with no batteries in another hive, and left another hive alone.  Because multiple hives were not tested, this calls all the measurements reported into question. For example, one of the things measured was the egg laying rate of the queen bee in each hive. Because there is only one queen per hive….and one hive per treatment….if one of the queens happened to be substandard, or just have a bad day, that would skew the results.  And there is no way to say what really caused the differences without a SAMPLE of many hives, each with the same treatments, or lack of treatment.

2. Because the sample size was so small, the claim of “significant” statistical results is invalid. Essentially, they had a sample size of one. You can’t do statistics on that. No statistical values were presented in the paper, although the word “significant” was used and the phrase “(p<0.05)” was in the abstract.  This is an inappropriate use of these numbers.

3. They used an EMF monitor in a way that (as best as I can tell) it was not meant to be used (see photo). EMF monitors are finicky things, and are supposed to be used in carefully calibrated conditions. Walking around with it and plopping it on a beehive are NOT part of standard procedure.

Additionally, there was no shielding on the other two hives that were NOT treated with cellphones–so they really received a background dose of whatever the cellphones in the region were producing, not a true control.

4. They put the cell phones in the freakin’ hive (see photo). Now, granted, maybe that is where Eddie Izzard takes his phone calls. But most people do not stand inside–or next to–active beehives when they are chatting about what to get for dinner.  This design is rather analogous to strapping cellphones to your scrotum. I’m betting you’ll get an effect–but is it a real one that the average scrotum owner needs to worry about?

5. The references cited include newspaper articles and advocacy websites, which are not authoritative sources for a scientific paper. An article on X-rays, (which are not the same as cell phone emissions!), was cited as a reference showing cell phones could cause elevated drone production.  Additionally, that paper was published in 1986, much earlier than cellphones were in common production or use.

This paper (which for a student research paper would be questionable) should not have been in a journal.
It definitely should not have postulated a connection to Colony Collapse Disorder.
And it should never have made the levels of press exposure that it did. Shame on all of you newsies.

This is a classic example of Bad Science Reporting.  OMG RADIATION IN MAI BEEZ!!!

(I’M IN UR BEEZ, NUKIN UR DNA?)

The idea that cellphones affect bees has been around since 2007–and it wasn’t legit that time, either. You might also find this discussion of electromagnetic fields at the Skeptic’s Dictionary helpful.

BTW, this paper would be a great exercise for a classroom–take one of the press reports I cited above, and have your students evaluate this paper. Can they spot the ways in which it’s not quite proper science?

18 Comments

Add yours →

  1. From the sound of it, this paper doesn’t even have anywhere near the rigor of a Mythbusters episode!

  2. Hey! Nice post!

    Current Science might be peer reviewed but it’s obviously not very rigorously so.

    Its impact factor is (according to wikipedia) 0.774 — I think that means it gets about 7 references for every 10 articles it publishes… (But I’m not sure.) Hopefully this will be one of the other three!

  3. They are using the EMF meter properly, the problem is that they are using an EMF meter instead of a spectrum analyzer and calibrated antennas.

    EMF (aka RF woo) meters are generally only suitable for ghost hunting, detecting electrosmog/electropollution and other RF woo sports. They don’t specify what brand/model they have there, but it looks like a TDM-200, +/-25% accurate and non-frequency discriminating, therefore not very useful for science and engineering. The lack of frequency discrimination means that AM, FM, TV, cordless phone, cell phone, WiFi and microwave oven signals will all read as values with no way to know the actual source.

    A standard radio receiver with a signal strength meter would be a more appropriate inexpensive instrument to show that the phone is on and transmitting. The tuning of a radio receiver provides the discrimination so that you’d know it was a cell signal rather than from some other source. Of course in the intended woo applications this doesn’t matter because all radio signals are considered equally dangerous to the electrosmog woo purveyors.

    On the basic premise of cell radio signals being the source of CCD, if this is plausible shouldn’t we be seeing significantly more CCD in urban/suburban areas than in rural areas. IME, cell signals in rural areas where most large scale farming is done are far weaker than in urban areas. AFAIK, we are observing less CCD in urban/suburban areas.

  4. Thanks Paul! I was hoping one of my more physics-oriented friends would comment on that very issue.

  5. Here in Scotland, we’ve noticed both honey bees and bumble bees have almost vanished this year. Our croft usually has large numbers of both as we leave a lot of “weeds” to flower, grow green manures that also serve as bee food, and grow many flowering crops that are attractive to bees (from broad beans to apples).

    Normally, on any given sunny day from mid-spring through summer we’d expect to see scores of our more common bumble bees, a few dozen of the more rare ones, and thousands of honey bees. And when the clover is in flower we’d normally have to tread carefully to avoid crushing numbers of bees.

    This year, we’ve seen less than a score of one species of bumble bee—over the entire season to date—and no honey bees whatsoever. It’s barely rated a mention in the media despite it being of enormous concern—I’m spending a lot of time fertilising flowers with a paintbrush to ensure we get halfway decent crops. And that’s not something that can be done on large scale.

    Very worrying.

  6. so how can this be properly tested? Rhere may be no pristine MW free environmets in the world anymore…

  7. Whenever it inconveniences us, we find reasons to disbelieve. We should be damned worried about the bees disappearing, and doing better studies instead of putting our efforts into disproving what they’ve done. If we don’t find out what’s happening to the bees and do something about it, we won’t be needing those precious cell phones anymore.

  8. @dannybill: It’s called a “Faraday Cage.” If you enclose the hive in a metal cage (the holes can still be big enough for the bees to fly through) you will block all EM signals. The cell phone can be placed inside and transmitting; it doesn’t matter if it doesn’t connect, because it’ll keep sending out a signal trying to find a base station. That’s what they should have done.

  9. Sage, disproving faulty studies is a PART of the process of finding out what’s happening. Think of it like this: if we’re convinced by ridiculous studies like the one above that something like cell phones is killing all the bees, and start acting as if that’s the truth, we’re going to completely fail at finding out what’s ACTUALLY killing them — which can then keep doing so unhindered.

    If something is indeed killing them, that is. I’m still unconvinced, but admittedly I haven’t read up on the subject (shame on me). I know bumblebees aren’t doing well because of habitat loss but that’s rather a different problem than domesticated honeybees dying.

  10. I’m more convinced that it is “bee’s AIDS”, this virus is caused by varroa. see e.g. http://www.daff.gov.au/animal-plant-health/pests-diseases-weeds/animal/varroa-mite

  11. The only real problem with this study is sample size. If they had gotten these effects with 400 hives you wouldn’t need statistics to analyze it.

    It is hard to do studies with large numbers of hives because bees need a large territory to gather honey and pollen over, and so you can’t put an arbitrary number of hives in an area.

    The only thing that is needed to do a larger sample size is more funding. If you don’t fund research like this until they have a large sample size, you will never get a large sample size.

    Bees navigate and probably communicate using electromagnetic fields. Precisely how is not well understood. A number of insects do use cryptochromes to sense magnetic fields. These utilize a photon to generate an ion pair which depends on the local magnetic field. This is how birds reset their magnetic compass each day. Look on page 22. Bees are adversely affected by electric fields on the order of 100 V/m, the same level measured in this study. Maybe it was a crappy measurement, but that only increases the uncertainty, it doesn’t negate the study.

    http://naturalscience.org/fileadmin/portabledocuments/microwaves_bees_birds_and_mankind.pdf

    In any case, no matter how crappy this study is, it provides zero evidence that bees are not adversely affected by cell phones. Because it is a positive study, even though it is crappy, it does provide some evidence that bees are adversely affected by cell phones. That evidence may be small, and because it is a crappy study we don’t know how much evidence it is, but that evidence is not zero, and it certainly is not negative.

  12. I’m surprised you don’t have a category of “Frass”!

    Faraday cages are awesome; I spent hours using one for sensitive recordings. Nowadays I just wish we had classroom-size Faraday cages to keep out the social & rudely inattentive & cheating mobile usage. (It’s one thing to not pay attention in class because you’re exhausted or bored, and quite another to sit up near the front and play games on your laptop or be constantly texting throughout the class.)

    andrea

  13. I’m amazed that up to this point people seem willing to blame it on everything but on the fact that it’s one single species being bred for such a long time. Could be varroa. Could be some case of a giant inbreeding (???) Or it could be that finally honey bees realized how much they suck and decided to leave other native bees do their jobs. Is this collapse disease affecting anything else other than Apis?

  14. Maybe you’re standing 10-30 feet near some bees that can get affected by the radiation, then they get lost and die? I don’t know, just proposing an idea of how they think it affects them. But they probably stick the phones IN the hive to be able to see a large distinction since they only had 4 specimens. Anyway, you gotta make some ideas of what has changed recently that could affect bees, apparently cell phone usage is one of them.

  15. Thank you, thank you for this post. This is exactly how I would have addressed it if I was in any way an authority on insects (in my work, insects are good food for vertebrates in wetlands!).

    The whole premise (i.e. non-replicated study, study not repeatable) seemed dubious at best. Thanks for tackling it – I’ll be sending it around!

  16. And Daedalus, that’s a bit off. YES – 400 hives with the same treatment – a treatment that replicates the amount and frequency and amplitude and duration of background transmissions that exist in the world today (somewhere), those 400 hives could be subjected to rigorous data analysis.

    Under those academic-level study conditions, they would still need data AND analysis. Otherwise their findings would be limited to, “Cell phone radiation has an effect on honeybees.”

    What effect? Is it statistically significant, compared to the natural mortality rate of those 400 hives. How soon are the effects notable? How soon are the effects irreversible? What, exactly killed the bees – the frequency, amplitude, duration of transmissions? Which bees died first? Workers? Queens? What happened next? Is this a fertility issue for the bees as well?

    Those are reasonable research questions, the answers to which could actually help guide government policy (and industry standards). Without that, causing an uproar based on “there is definitely effect on the bees. A bad effect!” is just tilting at windmills – or cell towers in this case.

  17. As a scientist I must agree that the study did not even come close to meeting “scientific method” criteria, so the results are poor at best.

    However, sometimes bad science does lead to interesting questions and this “study”, although deeply flawed, is worthy of an intelligent, well designed, follow-up.

    It might be that the follow up study finds no correlation to cell phones and bee decline, but at least we will have potentially answered some questions that this poorly orchestrated study raises.

    bill:www.wildramblings.com

  18. Out of place but you closed the comments on National Pollinator Week. Did you know that there is a community garden just steps from the mall between the Aerospace Museum and the Department of Education? It is open to any pollinator who flies by, pass the word

Comments are closed.