Concerns about finding effects that are actually spurious

During the course of data analysis one must be concerned about 1) failing t o detect real effects that are present in the data and 2) finding effects t hat seem supported by the data but are actually spurious. Our paper deals w ith the latter issue and outlines 5 scenarios in which the probability of f inding spurious effects is high. We provide some guidelines to avoid findin g and reporting effects that are spurious. It is unfortunate that there see m to be rewards but no penalties for finding and reporting on results that have a high probability of being spurious. We conclude that there is a need for more theory to guide empirical studies and warn against analysis strat egies that are especially prone to elicit spurious results.