s2 help

https://c08863e2f4be4cccd221320d0030b64a3346fa13.googledrive.com/host/0B1ZiqBksUHNYd29nZE5zT19PTTQ/June%202005%20QP%20-%20S2%20OCR.pdf

Can someone explain 6ii.

What do you think the answer is/why?

(It has something to do with significance levels and what they actually mean).

The null hypothesis gets rejected at 0.0468.
so 20*0.0468. But what does 0.8 show and what does 0.468, thats the bit im slightly confused with.

The 0.8 bit just shows that the probability of rejecting it is indeed the significance level you found earlier. If the probability was 0.6 or anything other than 0.8 then you can't use your significance level to answer thet question.

And the significance level (if it were 5%, for example) means that you would expect 1 in 20 tests to not lie in the confidence interval of whatever you have (and so you would expect to reject the null hypothesis once every 20 tests).

And since you have 20 tests, you're just multiplying by 0.0468 (the significance level, which is the same as the probability of rejecting H0) to see how many you'd expect to reject in 20o fthem.

But isn't 0.8 the probability. So why isn't it 20*0.8?

0.8 isn't the probability of rejecting H0. Remember for the test, you take 25 chocolates and calculate a statistic, and find that the probability of rejecting H0 is 0.0468. So you can think of the tests themselves as independent trials with p = 0.0468, and the distribution represents how many tests are failed, so if you have n = 20 and p, then you find 20p for the expected value.

It just tells that p=0.8 so you can use the significance value. If it were p=0.6 for example then the probability of rejecting the test would be much higher.

The most confusing bit about all of this is probably how the significance level of the test represents the probability of rejecting H0, and how you can think of it as
~B(20,p) where p is the probability of rejecting H0.

Cheers. For 7iv, how do you know its a normal distribution?

CLT, I suppose