My favorite is their trick of using zero exposure as a reference point to zero risk. Straight line from zero/zero to an effect level, cut that by a factor of 1000 or more for a safety cushion and you just got yourself a cancer causing chemical in very small doses...
Nobody really looks to see what kind of data supports a 1 in a million risk. They look even less closely at exposure type, duration, etc., that is needed to support a risk level.
Ignorance is grounds for panic, and regulation I guess...
Even if you have a real risk of 1/1,000,000 how do you ever sort that out when the background rate is 300,000/1,000,000?
Do you ban a substance that may save lives such as clorinated water. A small amount of organochlorines is produced which may produce a 1/1,000,000 lifetime risk of cancer but has saved millions of lives since its introduction?
It is really a shame that epidemiology is not tought to the vast majority of students.