1
$\begingroup$

I have been really interested in P - values lately and I have been reading various journals and websites on the subject. I would like to clear up some confusions that i have regarding P-values, CI and false discovery rates.

  1. Is a confidence interval set before a p-value is calulated?

  2. Does the confidence interval change when the p-value changes? For example, if the p value was 0.02 the confidence interval would be 80% or does that only occur for the alpha level.

  3. There is a lot of confusion over FDR(False Discovery Rates). If you get a p value of 0.047, on some websites they suggested that the chance of a type 1 error would be at least 26%. On some websites/Journals it says that a smaller p value would decrease the FDR. However, i thought that smaller p values have no more relevance in hypothesis testing after finding if it's significant or non significant. For example, would 0.05 be the same as 0.02 or 0.000001. Do they mean that increasing the confidence level would decrease FDR ?

0 Answers 0