Watch on YouTube

7:16

What is statistical inference?

A large trial found 30% lower cancer risk among older women taking vitamin D and calcium, and the authors still called it inconclusive. That is not timidity — it is statistical conclusion validity, the question of whether the numbers support the conclusion drawn from them. Frequentist inference asks how surprising this result would be if the treatment did nothing. Bayesian inference asks what we should now believe, given what we believed before. Both are reasoning under uncertainty; they disagree about what the uncertainty is attached to.

Also covered: statistical conclusion validity · frequentist and Bayesian inference · inconclusive results