The traps of statistical significance
Over the course of two decades, The American Economic Review—the most highly regarded journal in its field—published articles in which 70% of authors conflated real-world significance with statistical significance.