Adventures in Error Bars, Cell Phone Edition

So if you're worried about an exposure, and the exposure is difficult to measure (because, oh, you have no idea what the causal link would be between the exposure and the disease, so you don't know what you're measuring), and you keep doing studies about the exposure, you are eventually going to find a positive result (or weakly positive) because of the nature of chance.

That doesn't mean cell phones cause cancer. It means epidemiology is inexact.