Also a long history of spurious statistical correlations where you find "possible" links. e.g. for a while there was a scare for RF from overhead powerlines ca

Also a long history of spurious statistical correlations where you find "possible" links. e.g. for a while there was a scare for RF from overhead powerlines causing brain cancer. Turns out, the issue is the statistics used to try and isolate an effect from a general population sample have bias: plug in any effect, and any grouping factor, and you'll probably observe a doubling of the risk. This effect was the cause of a whole spate of "X may cause cancer" reports in the 2000s.