Either way, though, that seems surprising to me that there'd be a 35% increase in infant mortality.
I'm sorry if I didn't make it clear, but that's my point, it's so implausible. US is very far from Japan, so if there was a 35% spike there then the researchers' first reaction should have been to check the corresponding stats for Japan. If they didn't it smacks of intellectual dishonesty.
One of the most frequent complaints that comes up is the lack of safety mechanisms built into the reactor. Sure, they could have done a better job (the fact that electricity is needed for a cool-down is particularly disturbing) - but really, /all/ industrial mechanisms are going to start to experience a significant problems when hit by a historically significant earthquake. Sure, the engineers should have expected earthquakes (and they did - this is Japan, after all), but it's unreasonable to expect them to be prepared for a 9.0, just as it's unreasonable to expect them to be prepared for a direct meteor strike.
Would you build a plant that could only handle a 100 year earthquake, but not a 1000 year one? You'd give your children and grandchildren a 1/10 chance of catastrophe... in exchange for what, supposedly cheaper power?
Again, not that the engineers didn't screw up. But the fact that a nuclear power plant got destroyed by a magnitude 9.0 earthquake should come as a surprise to nobody, and should not reflect poorly on nuclear technology in general, any more than the fact that cars crash when driven by drunks should reflect poorly on motors.
"But sir - the question only said to shut down the power plant. It didn't mention doing without electricity!"
I'm surprised by both those things.