2 Comments
Felipe
9/1/2017 07:03:50 am
Shouldn't be the probability of the error alpha instead of p for the alpha error? The argument is that one sets a statistical procedure to have a certain probability of failing to not reject a null hypothesis (one controls that probability). The p-value is just a random variable used in the same way than the test statistic: if p is below alpha, then I reject the null hypothesis (and thus, if that hypothesis is true, I'd just be wrong alpha% of the time). What you think about this reasoning? Great blog btw. Cheers from Chile! 9/1/2017 11:22:38 am
I think that putting statical concepts into words can be misleading if people read the same thing in two different ways. I have YouTube videos on p and and on Alpha, as well as how they work together. I'd suggest taking a look at them. For links, see the "Videos" page of this website. Leave a Reply. |
AuthorAndrew A. (Andy) Jawlik is the author of the book, Statistics from A to Z -- Confusing Concepts Clarified, published by Wiley. Archives
March 2021
Categories |
RSS Feed