Ha … yes indeed I would. The clue for you is “exploring’ for...

Ha … yes indeed I would.

The clue for you is “exploring’ for something versus “confirming” something.

Its lunch time … so a good challenge for me … explain this dogma … without having to write a book.

We are playing billiards.

I hit my white ball into a big clump of coloured balls. The black ball goes in right side pocket. Was this a good shot? Its hard to say … we would call this an exploratory result.

Now imagine prior to taking the shot I pre-specified black ball right side pocket. And that’s what I did. We would say this looks like skill; this is a confirmatory result.

If I did it two times in a row (like the FDA required two positive on the primary endpoint confirmatory trials) we would be very confident about my skills.

But what if after pre-specifying black ball white side pocket instead I put the blue ball in the left bottom pocket?

I turn around and say well arguably the blue ball is just as good as the black ball …maybe even better, And the chances of me putting the blue ball in the left bottom pocket were less than p<0.05.

This sounds like nonsense … the p value on the blue ball is meaningless because what being tested was the p on the black ball.

And thats my best shot at a cringe-worthy explanation of something that is far more complicated.

And its not actually revolutionary thinking on my part.

It stems roughly from 1955 when two guys Neyman and Pearson spotted the importance of the apriori nature of hypothesis testing. MSB need to hit the history books I think.

If you are still not happy with the explanation … you could try writing to the FDA and pointing out the error of their ways.