Hypothesis Tests - The practical process...

A hypothesis test is part of a process of understanding, it is not a end in itself....
If you like these video lessons. Here is my buy me a coffee link.
www.buymeacoff...
Here is a free D.M.A.I.C users guide - Click the link and download it now! paul-allen.myk...
Design of Experiments for 21st Century Engineers - www.lulu.com/e...
Drink tea and read the paper - www.lulu.com/e...
Statistical Process Control for Small Batch Production - www.lulu.com/e...

[@MohamedAmarch]

First like First comment Keep doing a good job!

[@alexeykotov6650]

Hi Paul, as for me, this is quite contradictory. If you really had a lucky day and got 0.07, you repeat and receive 0.04 - why should be think that exactly 0.07 was lucky day, why not 0.04? I understand your point that we should be driven by practical experience, not math rules, but why then use math at all? As for me, this is very simple. If we agreed before testing that 0.05 is the limit, it definitely means 0.05 is the limit, even if we got 0.06. This is not throwing any knowledge away, it actually says that we don't have enough knowledge. More analysis should be put in place and hypothesis should be reviewed. Actually, receiving 0.07 doesn't limit us from any decision, it's still a business decision, we just have to be sure that everybody understands that data doesn't confirming our theory well enough. So from my POV, I'm totally against the idea of repeating the test to confirm/find statistical significance (until we do it 20 times to really get 19/1 distribution).

[@paulallen5321]

Hi Alexey - 0.05 is not a hard and fast rule. Fisher who invented hypothesis testing never used this rule. But as the sample size increases the estimate improves but not unlimited way an extra sample will improve the result. I wasn't saying test after test after test. I was saying increase the sample size. Perhaps I did give the wrong impression in the video...

[@kareemcurwen891]

🙋 Promo sm