# Thread: Hypotheses testing - apparently crying isn't going to help

1. ## Hypotheses testing - apparently crying isn't going to help

So that is the entire day wasted.

I have no idea what I am supposed to be doing here, an assignment due in next week and only in the second question (which is the above). I distance learn so can't just pop and see a teacher.

I am just reading this and don't even understand half of the words and can't find anything similar anywhere.

Basically, I have two sets of information - comparing the effects of alcohol on two similar illnesses - I am given the sample number for the first, and the number of drinkers- and the same for the second illness.

I have to create distribution models for both - I created Binomial distribution for both of them.
Then I need to write the null and alternative hypotheses used to test the proportion differences/similarities.

I am sorry to be so emotional haha - but if I can't do this, I have no hope with the rest of the assignment and I don't know what to do :-(

2. ## Re: Hypotheses testing - apparently crying isn't going to help

Originally Posted by SamanthaJane
So that is the entire day wasted.

I have no idea what I am supposed to be doing here, an assignment due in next week and only in the second question (which is the above). I distance learn so can't just pop and see a teacher.

I am just reading this and don't even understand half of the words and can't find anything similar anywhere.

Basically, I have two sets of information - comparing the effects of alcohol on two similar illnesses - I am given the sample number for the first, and the number of drinkers- and the same for the second illness.

I have to create distribution models for both - I created Binomial distribution for both of them.
Then I need to write the null and alternative hypotheses used to test the proportion differences/similarities.

I am sorry to be so emotional haha - but if I can't do this, I have no hope with the rest of the assignment and I don't know what to do :-(
Hi SamanthaJane!

It sounds a bit like you're supposed to do a 2-way ANOVA test.
But it depends a bit on the actual data you have.

What is it that was actually measured?

3. ## Re: Hypotheses testing - apparently crying isn't going to help

I have managed to calm myself down a bit - temporarily!
I can't really say anymore because it is a marked assignment. But I look into ANOVA testing - it is only 2 points, but the rest is on minitab so will have to try figure that out!

4. ## Re: Hypotheses testing - apparently crying isn't going to help

Sure!
Let us know if you have any questions.

5. ## Re: Hypotheses testing - apparently crying isn't going to help

Actually, I do have one :-)
I thought I was doing really well with this - took a few days - but seem to have got there.

What I missed in the question was, although n=10 they told us to 'assume normal distribution'

So back to square one!

I have used xbar-mu / s/sqrtn to get the test stat - I also need the null distribution which I think is;N(mu,sigma^2) - So I have used mu from my null hypothesis and s^2 for the pop variation - is that correct?

But now I am trying to get the rejection zone - and I keep getting +/-1.96 as it is a 5% confidence interval, but when I use minitab I get totally different answers, and have no idea where they came from.

Any clues? I am fairly sure I am making errors all the way through!

6. ## Re: Hypotheses testing - apparently crying isn't going to help

When you have n=10 and no information on sigma, you are supposed to do a t-test instead of a z-test.
Your formula is correct, but you need to look it up in a t-table with n-1=9 degrees of freedom.
This is what minitab will have done.

7. ## Re: Hypotheses testing - apparently crying isn't going to help

oh!!! That is great, so I can perhaps salvage some of my earlier work?
Even though the question states specifically that it can be modelled by a normal distribution?

8. ## Re: Hypotheses testing - apparently crying isn't going to help

The t-test also assumes a normal distribution - it's just with an unknown sigma that is approximated by the standard deviation of the sample.

9. ## Re: Hypotheses testing - apparently crying isn't going to help

Oh - I thought it was two totally seperate tests - one for below 30 samples and one for above 30 samples.

So actually, my original answer should be correct - although in that I showed a T distribution - T~T(n-1) rather than an N one - N(mu, sigma^2)- which one is correct?

10. ## Re: Hypotheses testing - apparently crying isn't going to help

(as a null distribution?)

11. ## Re: Hypotheses testing - apparently crying isn't going to help

When n is large enough the corresponding t-distribution is equal to the normal distribution.
You can see this if you check the t-table.
Btw, n=30 is still a bit low to switch to a z-test.

In your case, the assumed null distribution would be the t(n-1) distribution.

12. ## Re: Hypotheses testing - apparently crying isn't going to help

Btw, if n > 30, you can assume that sample mean of any distribution behaves like a normal distribution (central limit theorem).
To do a t-test, the assumption is that the sample mean is normally distributed.