Question

If I had a normal population, and I went out and got bunch of sample pairs (lets pretend it was a thousand)... If were to do an F test for an variance eqaulity... and lets say I wanted a 95% confidence... how many times do I expect to reject my null?

I'm not sure how I would go about this? How would i know with out actual data to plug into my numbers?

Answer 1

P(reject null hypothesis when it is true)=level of significance=alpha(%)

%confidence level=100-alpha

here confidence level=95%

so alpha=100-95=5%

so 5% times we reject the null hypothesis when it is true.

so if you repeat the experiment 500 times then 5% of 500=25 times you expect to reject null.