Comment by rossdavidh

The core issue is that p-values are cheaper to get than replicating the study, but replicating the study is the only reliable way to see if it's true or not. Sometimes the expensive/time-consuming way, is the only good way.