Scientific integrity. Who is responsible when things go wrong?

What should we do when practitioners ignore ethical standards in animal experiments? For example, if a journal has a policy that it will not print articles from experiments that cause "unnecessary pain and discomfort," how far do they have to go to ensure that such articles are not accepted? What about the people who did the experiment? Or the institution where those people work? What happens if editors and publishers later discover that there was pain and discomfort?


From the blog of the Editors of AJB: http://blog.bioethics.net/2005/12/teaching-scientific-integrity-might.html

"Science depends upon trust in the honesty and integrity of its practitioners, perhaps more than any other human endeavor. The peer-review process means that experts review the data and the methods that are reported but, ultimately, scientific articles are a form of testimony. Other scientists then attempt to repeat research results -- so that if something is fabricated it will typically be discovered over time. But oversight committees like human experimentation committees have to trust what their investigators tell them. Journals also must rely on scientists to tell the truth to them. And researchers themselves often have to rely on the honesty of their graduate students and post-doctoral students. When trust breaks down, the very possibility of science is threatened. "