Slashdot Mirror

← Back to Stories (view on slashdot.org)

Possible Cellphone Link To Cancer Found In Rat Study (nbcnews.com)

Posted by BeauHD on from the significant-findings dept.

An anonymous reader quotes a report from NBC News: A giant U.S. study meant to help decide whether cellphones cause cancer is coming back with confusing results. A report on the study, conducted in rats and mice, is not finished yet. But advocates pushing for more research got wind of the partial findings and the U.S. National Toxicology Program has released them early. They suggest that male rats exposed to constant, heavy doses of certain types of cellphone radiation develop brain and heart tumors. But female rats didn't, and even the rats that developed tumors lived longer than rats not exposed to the radiation. The National Toxicology Program, part of the National Institutes of Health, is still analyzing the findings. But John Bucher, associate director of the program, said the initial findings were so significant that the agency decided to release them. A 29-year-old study published earlier this month from Australia reassures us that cellphones are reasonably safe, and do not cause cancer.

5 of 113 comments (clear)

  1. More likely to die like a bug on a windshield by TigerPlish · · Score: 2, Insightful

    People who walk about with their noses in their screens will stand a much higher chance of dying like a bug on a windshield than from radiation.

    This debate is getting as old as climate change and "does this belong in slashdot or not?"

    --
    The "Civilized World" jumped the shark ca. 1973.
    1. Re:More likely to die like a bug on a windshield by Ol+Olsoc · · Score: 4, Insightful

      People who walk about with their noses in their screens will stand a much higher chance of dying like a bug on a windshield than from radiation.

      This debate is getting as old as climate change and "does this belong in slashdot or not?"

      Agree on the first point, but not so much on the second one. Reading the paper, I'm a little surprised by the preliminary results. But if it can be reproduced, they might be on to something.

      I've always been a proponent of expecting no cancer relation from these things, because they aren't sending out ionizing radiation. That being said, these little devices put out RF, and you are using it in the near field, so unless cell phone RF is unlike any other form of RF, there will be some tissue effects.

      So my interest is keen in this study.

      --
      The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
  2. A more accurate headline by dlleigh · · Score: 2, Insightful

    A more accurate headline would be "Cell Phone Links to Cancer Only Found in Shitty Studies".

    1. Re:A more accurate headline by Anonymous Coward · · Score: 2, Insightful

      What I didn't like about the study is that they drew the conclusion that brain lesions are likely caused by radiation in males, although the results showed that 3W/kg GSM is more harmful than 6W/kg GSM and the 3W/kg CDMA group suffered no brain lesions although 1.5W/kg CDMA group had 2 cases. These counterintuitive results (and the very low nominal value of positives) mean we need much bigger sample sizes to draw conclusions.

      The Schwann cell tumors in non-irradiated females (clamed to be non-susceptible) are as likely in as irradiated males (claimed to be susceptible).

  3. Lies, Damn Lies and No Statistics by Roger+W+Moore · · Score: 4, Insightful

    People who walk about with their noses in their screens will stand a much higher chance of dying like a bug on a windshield than from radiation.

    I doubt you would get that result if you let the people who did this study test that hypothesis. There is not a single uncertainty on a measurement shown in the paper as far as I can tell and they are dealing with tiny statistics which are prone to large fluctuations. Their most statistically significant result seems to by about 5% likely to occur by random chance (based on their own statistical calculation which frankly I would not trust at this point) but with just over 100 measurements it seems very reasonable that this would occur by chance. Indeed they even point out that this rate was achieved in one of the 13 control samples they list in the appendix D!

    With a sample size of 90 differences of a few incidents are not statistically significant when you are making lots of measurements and there is a high degree of correlation which has to be taken into account since all comparisons are made to a single control group so a statistical fluctuation there affects all measurements. As the saying goes there are lies, damn lies and statistics and this paper is very much lacking in statistics...not that the authors are deliberately lying but their conclusions do not seem statistically valid.