Not gonna pretend I'm an expert on EM radiation effects, but this statement: > Our main takeaway from the current review is that approximately 1,000 hours of l

Not gonna pretend I'm an expert on EM radiation effects, but this statement: > Our main takeaway from the current review is that approximately 1,000 hours of lifetime cellphone use, or about 17 minutes per day over a 10-year period, is associated with a statistically significant 60% increase in brain cancer. Is pretty misleading. He's talking about odds ratio here, not relative risk, because these were case-control studies. And because they seem to have been conducted specifically looking for people with cancer, they do not conform to the rare disease assumption (in this case, the paper says cancer = 24,717, control = 41,358, nowhere close to the typical 10% cutoff at which odds ratio is considered to reasonably estimate relative risk. Also they found no statistically significant relationship in any direction until they did subgroup analysis. That starts getting awfully close to p-hacking. "All studies overall don't find an effect, but if we limit to the studies that find an effect, we find an effect." Finally, this seems like an awfully hard thing to study using case controls. I'm not going to try and find these studies being meta-analyzed, but it would seem they're largely finding people with brain cancer and people without brain cancer, and surveying them on how long they've used cell phones and how much. Unless they're actually going back and consulting 10 years worth of their own phone records before answering, which I am sure they are not doing, how accurate is the recall seriously going to be? And there is no attempt to quantify the characteristics of usage relevant to the mechanistic reasons to believe a risk might present, that is, how strong was the signal, how far away from their ear do they keep the phone, are they typically going hands free, do they keep the phone or on near them when not in use? These are impossible to study using this kind of study design. I'm not some kind of shill for big telecom. I barely even use a mobile phone. I talk to my dad maybe once a week and that's my only regular call. My diet and workout tracking apps are the only apps I use aside from banking. I almost never bother to bring it with me when going shopping and don't sleep with it within arm's reach. So I'm following the recommendations here anyway, not because I'm afraid of cancer but because I don't see a need for having a phone nearby on any regular basis. At the same time, through no fault of their own, these researchers are studying something that is virtually impossible to prove anything about, and even what they might have found if you squint really hard at a specific subset of a non-random, non-representative scene, still wasn't a lot. I don't see how this warrants a press campaign aside from Joel Moskowitz apparently being politically active in getting California municipal governments to warn consumers about anything and everything.