How Dangerous Are Cell Phones?

A new report has come out that says that cell phones probably do cause cancer, as several people, such as David Servan-Schreiber, have argued. But the news is not all bad:

The design of the study is fundamentally flawed, as well-documented by “Cell Phones and Brain Tumor.” For example, users of cordless phones only were treated as unexposed. But, two independent studies found users of cordless phones had an increased risk of brain tumors. So, excluding such users underestimates the risk of brain tumors. This flaw suggests either ignorance or dishonesty on the part of the researchers running the Interphone Study. Then, there’s the suspicious finding from some parts of the Interphone Study which concluded the use of a cell phone for less than ten years lowers your risk of brain tumors. This suggests the bias was so strong it eliminated enough tumor risk to show decreased incidence. The Interphone studies did find more brain tumor risk after more than ten years of cell phone use. The report notes that the risk was so great it could not be camouflaged even by the bias of the study.

Emphasis added. The person who wrote that hasn’t heard of hormesis.

3 Responses to “How Dangerous Are Cell Phones?”

  1. Caleb Says:

    What’d be interesting is to see the data broken down by how often a person uses a cell phone. I imagine we’d see those who use their phones only every now and then for short calls would have less brain cancer due to hormesis. Here’s to being an antiosocial guy:) Now if we just knew where the dose response peaked: “Sorry honey, I’m going to cut you off; I’m over my allotted radiation regime for today.”

  2. Dave Says:

    Forget hormesis. What about the irony of someone suggesting that a study must be biased because its results don’t fit his pre-conceived notions?

  3. seth Says:

    Dave, that’s a very interesting comment. In my experience, to test Theory X you can never arrange a pure test: a test of Theory X and nothing else. You need several additional assumptions, which I’ll bundle together and call Theory Y. So what’s tested is Theory X and Theory Y taken together. If the prediction of the combined theories is false, you tend to say Theory Y is at fault. At least you consider that possibility. Only after the predictions of Theories X and Z, X and A, and so on (tests based on other supporting assumptions) turn out to be false do you start to believe Theory X is wrong. It might be ironic, but it’s reasonable.