[archgoon]

> but if the universals that are now commonly used have turned out to not actually be true wouldn't that mean huge swathes of researchers have to retract/redo their research?

Reproducibility is a major concern. Many experiments are poorly designed.

http://science.sciencemag.org/content/349/6251/aac4716

So, yes, about 60% of research needs to be redone.