The discussion of whether it’s true that 3,000 people were killed by Hurricane Maria in Puerto Rico hinges on how much you trust statistical models created by university professors.
Obviously, there is no list of specific 3,000 people who died because of the hurricane, and it’s impossible to make such a list. I hope at least here on the blog we all understand why it’s not how such estimates are made.
The 3,000 dead (and previously 5,000 dead) was a provisional number created by using a statistical model. There were actually two different models applied. The first one produced the number of 5,000 victims. It was later discredited and a clearer model was created. Based on the newer model, the result is 3,000.
When you actually look at the model closely, I’d say that 1,000 deaths are pretty much indisputable. Another thousand can be disputed but not very fruitfully. The last thousand is maybe a bit more doubtful. I’ve read that the authors of the model might have assumed that the people who evacuated from the island before the hurricane were less healthy than those who stayed. And we all understand that it was probably the other way round. Those who evacuate tend to be richer.
So I’d say the death toll is definitely over 2,000 but if you don’t understand the basic functioning of statistical models, you won’t get it. To me, the model used by the scholars at George Washington U to arrive at this number looks very convincing.