FROM THE EDITORS BIG SAMPLES AND SMALL EFFECTS: LET'S NOT TRADE RELEVANCE AND RIGOR FOR POWER

摘要

My conclusion, obviously, is that sample and effect size matters a great deal. Rephrasing Robin Williams's genie, large samples potentially give us "phenomenal statistical power! . . . itty-bitty effect sizes." Certainly, this does not need to be the case if underlying relationships are large and we researchers pay close attention to construct validity. The statistical power rendered by large samples does, however, give us the ability to find small effects, and it seems to me that such power increases our responsibility to insure that other research design problems are not masked, and that we look less to statistical significance and more to theoretical and managerial significance in evaluating our research.