There's a ton of studies with these problems. Researchers simply engrossing the test subject in the material until something bad happens. Unless you're researching on a test group of humans, then suddenly all the levels are actually less than typical.

It all depends if you're looking to prove that it's harmful or not. Want to find it's harmful? Get a bunch of mice and expose them to as much of whatever substance you need to in order to find a problem.... Want to prove something is safe, set up a "double blind" study of the effects on humans, and give half of them regulated and limited doses of it for weeks or months until you can convince everyone that "nothing bad happened".

I have a problem with research done in either way. Researchers should be neutral, and just test and let the data speak for itself. (With limited interpretation for the people who read it)

Instead, almost all research is funded by someone with an agenda who is trying to find out if x is good/bad, and prove or disprove a specific stance. Argh