A classic bedtime story teaches a confusing statistical concept. Don’t believe me? Follow along.
Given the following hypothesis:
Ho: no wolf present
Ha: wolf present
What error was the boy who cried “wolf!” committing? A type I error, of course. He rejected the true null hypothesis and concluded the alternative.
The boy can commit a type II error by not crying for help when there is a wolf present or fail to reject a false null hypothesis.
Remember, type I errors require action or rejection while type II errors require inaction or failure to reject. But what about correctly rejecting a false null? That’s what the boy did last, correctly rejecting a false null hypothesis or crying wolf when there really was one.
Spouting significance will draw attention, but if your methods are phony, your credibility evaporates like rain on North Carolina summer streets. Once your credibility is gone, no one will come running no matter how significant your findings, and Old Tuffy will devour all the sheep.
Columnist: Nathan Jones