16. Design experiments that can answer cause-and-effect questions
Use random assignment, control groups, blocking, blinding, factorial designs, and replication. You will see how good experiments create stronger evidence than analysis alone can rescue.