Even Fukushima levels of radiation, spread over the globe, aren't too horrible.
There are all kinds of reasons to reject coal-fired power. Radiation isn't one of them.
See the xkcd radiation chart for relative measures.
A nuclear power plants regular radiations are below the level of a coal plant. But a worst-case scenario, of which we have multiple extant examples are unbelievably worse: 9 trillion times more radiation, using the xkcd reference I posted (adjusting the intervals for equivalence -- 1 year rather than 10 minutes as the Chernobyl core instance is given). And it's been non-engineering factors, for the most part (management and oversight) which are the failure modes driving these. Which is to say, the problems cannot be fixed by engineering around them.
The fundamental error, though, is that this is a false dichotomy. There are alternatives to both coal and nuclear. And they bypass this problem entirely.
That's so cute. A nuclear plant would ideally emit zero radiation if everything is designed and working correctly. A leak is a warning sign and the potential danger is an event like the two that have happened so far. In other words, that level of radiation leak from a nuclear plant isn't actually a problem. They shut it down to make sure something really bad doesn't happen. With coal you get the leak without the catastrophe.
I havent seen anyone do the math, but given the amounts of coal involved i would expect more radiation comes from coal plants each year than a yearly fukashima-type event.