Hypothesis Testing

A new drug lifts test scores by 4 points on average. Triumph — right? Not so fast. Maybe the lucky group happened to be sharper. Maybe a different sample would show zero effect. How do you tell a REAL effect from the noise of chance? Science can't run the experiment infinitely many times, so it runs it once and asks a sharper question: if there were truly NO effect, how surprising would this result be? That question — and the number that answers it — is the backbone of every experiment you've ever read about.