Scientists are normally happy to find regularities and correlations in their data—but only if they can explain them. Otherwise, they worry that those patterns might just be revealing some flaw in the data itself, so-called experimental artifacts.