Data snooping is a major concern in empirical asset pricing. We develop a new framework to rigorously perform multiple hypothesis testing in linear asset pricing models, while limiting the occurrence of false positive results typically associated with data snooping. By exploiting a variety of...