Witness the bizarre Ann Coulter vs. Bill O'Reilly exchange in which Billo comes off as the voice of reason.
I think there's just a lot of honest ignorance out there. People want to know, am I safe (given some level of exposure)? The only way to say yes unequivocally would be if a threshold model were true (or possibly even one of the "a little radiation does you good" theories). 90% of the serious attempts I've seen to answer this question in the media do assume a linear no threshold model, but the concept of probability is a serious challenge to public understanding. So the true answer is twofold; on one hand, as an individual, given the exposures seen so far, unless you've been working to fight the disaster the odds are still overwhelmingly against this killing you. On the other hand, expose millions of people to a enough radiation and you can bet on a certain number of additional deaths thanks to the exposure.
So people are triangulating, because it can be genuinely harmful to cause people to overreact (and not just to the finances of the nuclear industry). You need to do some risk assessment (I'll be grabbing numbers from the
http://en.wikipedia.org/wiki/Micromort">Wikipedia article on the "micromort". According to a standard LNT dose model, exposure to 10 millirem of radiation increases your risk of a cancer death by one part in a million. Expose 10 million people to 10 mrem each and you'd expect that on average you'd get 10 additional cancer deaths - in practice, impossible to measure, but very real nonetheless.
So let's say you want to evacuate a huge area, and do so by car. Let's say you move your hypothetical 10 million people an average of 100 miles. It turns out traveling 230 miles by car carries a 1 in a million risk of death due to collision. This hypothetical evacuation results in an additional 1000 million passenger-miles of travel; you'd expect there to be an additional 4 deaths because of this extra travel vs. the 10 people you may have saved from fatal cancer. Of course, this further assumes that you've dropped the radiation exposure from 10 millirem to nil; that probably isn't going to be true, either. That's even more true if you're evacuating people by plane; you're automatically giving them a radiation exposure just by flying.
So whether a proposed action makes sense really depends on the numbers. Just how much exposure can you anticipate? What are the effects of reactions to the threat? And most of all, how much confidence can you place in your risk assessment (i.e. how bad can it get if things go worse than you'd expected)? I think it's on this last point that we've seen the biggest divergence between US and Japanese authorities' assessments of this incident. Unfortunately, TEPCO has a horifically bad track record with respect to honesty, and I'm more inclined to accept the more pessimistic view the US government seems to have.