When an experiment is repeated and the new results are different from the original results, what could you assume to be true?