On question 4 when we are asked to test the hypothesis about proportion of hair colors, personally I calculated the (observed-expected)^2/expected test statistic for all 4 rows, and had a degree of freedom of 15. In the mark scheme they just did the totals, and although we came out with the same result, I feel their method of only using totals disregarded a lot of data? Any advice about the best method, and I'm really interested to know if you think I would get the marks for doing it my way.
Thanks for any advice, and good luck for the exam!!
Yeah, so there was two papers. Some sat the replacement, and some the other normal paper. I think it is to do with the original paper being leaked. It happened with all models
On question 4 when we are asked to test the hypothesis about proportion of hair colors, personally I calculated the (observed-expected)^2/expected test statistic for all 4 rows, and had a degree of freedom of 15. In the mark scheme they just did the totals, and although we came out with the same result, I feel their method of only using totals disregarded a lot of data? Any advice about the best method, and I'm really interested to know if you think I would get the marks for doing it my way.
Thanks for any advice, and good luck for the exam!!
I did it the same way that is written in the mark scheme, I don't really see how you think it disregarded a lot of data, since every value is used in the calculation. As far as I can tell no data has being disregarded. But obviously the method you chose to use worked for you so my advice would be to continue doing what you're comfortable with in the exam, there's more than one route of getting to the answer. For the same paper, I don't really understand question 6a), any chance you could try and explain it to me?
On question 4 when we are asked to test the hypothesis about proportion of hair colors, personally I calculated the (observed-expected)^2/expected test statistic for all 4 rows, and had a degree of freedom of 15. In the mark scheme they just did the totals, and although we came out with the same result, I feel their method of only using totals disregarded a lot of data? Any advice about the best method, and I'm really interested to know if you think I would get the marks for doing it my way.
Thanks for any advice, and good luck for the exam!!
Just did this paper, you're right it was quite challenging.
The question states for you to only use Table 1 - so you essentially ignore all the stuff about eye colours and focus on the hair colours. From this you get your observed values as 105, 282, 48 and 165 respectively. Using the ratios they give, you can calculate your Expected values (100,300,50,150) and then use the normal method. There are only 3 degrees of freedom (4 parameters and 1 constraint)
It was definitely a question that required you to thing outside the box - you just need to ignore all irrelevant information (eye colours) and focus on the distribution they're asking you to test.
I doubt that tomorrow's paper will be as difficult as this, however.
on june 2013 replacement paper, on question 4, can anyone explain to me why the null and alternative hypotheses were written like that? Its because every other past paper ask you about association. Thanks
We've got to wait 24 hours, so all im saying is that I was quite pleased. A few dodgy definition questions and weird mean and variance questions but overall, happy. Quite similar to most past papers.
We've got to wait 24 hours, so all im saying is that I was quite pleased. A few dodgy definition questions and weird mean and variance questions but overall, happy. Quite similar to most past papers.
Nope, you're welcome to discuss it straight away, TSR introduced new rules the other day meaning that we can discuss the exam as soon as it has finished, provided we don't post a full unofficial mark scheme until 24 hours have passed.
Yes, I found it quite a good paper on the whole, that defining statistics question was a bit annoying though, I put yes, no, no respectively for the three, did anyone else get that?
Yeah I thought it went quite well too. The only thing I think I did wrong was forgot to square the 2 on the bottom of the statistic fraction when calculating the variance :-((( besides that pretty standard I wondered why there was 7 mark for just an S2 statistic question