I'm not saying that I advocate their use, I think the privacy issues and the fact that they're plain ineffective is enough to warrant their removal from service.
We're not talking about the radiation fluxes through an arbitrary area at 30,000ft and in the machine, we're talking about radiation absorbed by the body and the figures seem to correlate with what Schneier's article quoted.
Most of the quotes for the backscatter machines are in Rontgens which is an equivalent dose, just non SI. Ideally we'd be using Sieverts for standard's sake.
Frequent flying nets you 3uSv per hour (from a Google/BBC), a back scatter x-ray nets you 0.1uSv. So the effect of flying is an equivalent dosage of 30 times the x-ray. A three hour flight, e.g. across the US, is therefore around 150 times that from the machine. Neither are specifically pointed at one part of your body and the anatomical weighting system is such that a whole body blast is weighted to be 1, i.e.
Effective Dosage = Sum (H W)
where H is the whole body dose and W is the specific body part weighting.
Yes, if your 2.6Sv annual limit is solely from absorption by your testicles you should be worried, but in terms of weighting, gonads are somewhere in the middle. Skin requires more radiation for the same equivalent does (8 times more) and your colon requires just under half as much. The weighting factors are currently from the ICRP 2008 guidelines.
That said, there seems to be a lot of conflicting arguments going around. Some researchers say that it's not an equivalent argument because the radiation is concentrated - but surely they scan your whole body? - compared to the diffuse exposure you get at altitude. However, equivalent doses are supposed to account for this and are designed for direct comparisons.
The US government (heh, unbiased much?) said this:
http://www.fda.gov/Radiation-Emittin.../ucm231857.htmThe recommended limit for annual dose to the skin for the general public is 50,000 µSv. The dose to the skin from one screening would be approximately 0.56 µSv when the effective dose for that same screening would be 0.25 µSv. Therefore the dose to skin for the example screening is at least 89,000 times lower than the annual limit.
http://www.radiationsafety.ca/airport-backscatter-x-ray probably pretty biased, but anyway. Even if the figures are too high, the risk of flying still vastly outweighs a quick blast from a backscatter machine and any deaths due to the machines are indistinguishable from the statistical fluctuations.
CT scans on the other hand which people bring up, do give you a big dose and should probably be cause for concern. However, CT scale dosages are only used for luggage. I think people misunderstand this and assume that they're getting a CT scan every time they walk through the gates. X-ray backscatter is fundamentally different radiation.
What is definitely clear is that there is a lot of misinformation on both sides of the argument. Ionising radiation is generally bad business and I don't see why metal detectors can't remain.