It's simply interpreting the data differently. Statistical conclusions always depend on your priors. We may also have a difference in risk tolerance. I'm willing to accept "accidents almost never happen" for something like air travel because the consequences are much less severe even if a worst-case accident does happen. I'm not nearly so willing accept even a small risk of nuclear accident because the consequences are so potentially awful (the worst case is not death IMO, but slow, painful radiation poisoning).