Most the death tolls, especially the "1 in every X seconds". They're usually such a catastrophically newsworthy rate that they're impossible to be true. and don't account for fluke spikes and the like that might work to their advantage to skew results. Advocates against smoking, for example, might take all lung cancer death totals to use in their favor because smoking is linked to lung cancer. There's no real context behind these stats yet they just get posted up as "shocking" results.