Methodological errors can be exploited to produce null results researchers want

Evidence

Regrettably, one can get quite a variety of results by making enough methodological blunders. Ask me how I know. It is particularly easy to screw things up and find nothing. Sometimes, believe it or not, that is what a researcher hopes to find. For example, “See, my test is a good as hers; there’s no significant difference in the validities.” Or, “Manipulating the key variable in Theory X—a theory I frankly have little time for—had no significant effect upon the outcome.” — Bob Altemeyer, The Authoritarian Specter, 648