Something that you assume to be correct based on the data from an experiment is a(n)