I always get confused when to reject or support/fail to reject the null hypothesis given values for both. Is it like, whenever p values aree less than significance level we reject the null? Does this statement always hold true?
Thnks
p-value < (1-confidence intevall = significance level) than reject the null otherwise fail to reject the null. For visualization draw a schematic bell curve with a confidence intervall and then draw a p-value inside the confidence intervall (=fail to reject the null) and a p-value outside the confidence intervall (=reject the null). Best, Oscar
thnks
I agree with Oscar.
Does rejecting null mean you automatically accept the Alternate?
Yes, that’s the consequence.
I had a candidate give me a great way to look at p vs. α:
- α is the chosen level of significance
- p is the observed level of significance
You reject the null hypothesis (in favor of the alternative hypothesis) if the observed level of significance is less than the chosen level: p < α.
You fail to reject the null hypothesis if the observed level of significance is greater than the chosen level: p > α.
I always over thought this one, but it is really simple. If the p value (decimal from 0-1) is less than 1- confidence interval, you reject the null.
So if confidence interval is 95% you reject all p values below 0.05.
Unless I am wrong, then this was not helpful
do you mean you fail to reject the null if the observed level of significance is GREATER than the chosen level: p>a? otherwise, wouldn’t your statement be the same as the one right above it?
Yup.
I’ll correct it.