Here are questions to ponder at some time when the definitive data are available:

What was the fatality/injury rate among those who elected to stay on the keys during Irma? How does that compare with those who evacuated?

So far, I have not seen anything about large numbers of fatalities among those who remained, nor, for that matter, anything about incidents among the evacuees. Indeed, the number of fatalities (somewhere in the mid-50s) seems quite low considering the population at risk in this storm.

Presumably one would leave because remaining would expose one to a significant chance of dying. If that is not the case, there is a significant advantage to sheltering in place - nothing works like home field advantage...