OCR MEI S4 Jun 2010

Yeah, it was quite hard.

Which questions did you find hard?

And had you done any practice papers?
I hadn't and found it whooooa different to the stuff in the (slightly obese) textbook

I could only do fragments of question 1 (iv?)
Question 2 was quite killer, and question 3 (i), I didn't know what you had to do.

The book is extremely TERRIBLE, they overcomplicated EVERYTHING. I hope the grade boundaries are very low.

Pretty much had the same experience.
Question 1, I just didn't get at all - the most I got was writing down the equation for unbiased estimators
Question 2 was just bluh - I knew what was going on, just couldn't do it!
Question 3 was good, pretty standard, and I just so wish I'd studied for question 4!

I utterly agree with your second paragraph there: I read a bit, think WUT, get depressed, shut book, get depressed because I am not reading book and repeat

In seriousness though, it is the least accessible book I think I have ever read - and I thought I finally understood most of the stuff in there.
Then walked into the exam and ...

Still, I have satisfaction in that I wrote to the examiner telling him/her/it about how I thought all the hard papers this year (the '4' units, FP3 and even FP2) were made especially abstract and difficult, in order that OCR and MEI will get feedback about how "unpopular" those units now are, and will then have no reason to continue to offer them, when they launch the new "better" specification.
It's all a conspiracy

I thought I was well prepared for the exam, I had S2 and 4 and onthe same day. After S4 I felt blank. I didn't like it at all, I really hope the grade boundaries are low. BTW what did you do for question 4, particarly 4i and 4iii?

I knew I wasn't prepared for the exam at all.
Physics was on the same day for me, so I concentrated on that.

Unfortunately, I didn't do question 4 since I hadn't prepared it.
Wish I had though!

I hope the grade boundaries ar!e low too.
But I've had bad stats scores even when I thought I did well, like with S3 last time I got 57 (D) and just thought wt-actual-f because I'd not got below an A on any other maths unit .
So on the back of that, I'm expecting far lower than 57!
Maybe even a U
At least I'm expecting it this time, though.

By question for 4 I meant question 3 (i) and (iii)... it has been a long two days for me.

Oh right!
I can't remember what I got for the confidence interval stuff (I assume that's what it meant by 'acceptance region' or whatever it was).

For the hypothesis test, I used the Wilcoxon rank sum 2-sample test (hope that was the right one ), and got (surprisingly) that $H_0$ was acceptable.

What about you?

Crap, I had a feeling it was Wilcoxon but I didn't do it Damn, I wonder how many marks I will lose?

M4, FP3 and S4 have been quite brutal (but M4 is mainly due to how ridiculous the book is...)

How did you go about answer the first question of finding E(X) and the question on proving the when lambda -> infinity Mz(th) -> e^(th^2)/2

I agree - and even FP2 was pretty 'abstract' (i.e. hard )in the sense that one of the questions was about a matrix which you had to analyse, but you never actually found out what that matrix was.

I think the M4 book is pretty good to be honest, way way better than stats at any rate

Finding E(X), I used integration by parts twice
I just got completely confused because the textbook says that you can't integrate the normal distribution stuff (hence the tables), so I thought well if this is about the normal distribution they must be looking for something else
As for the lambda tending to infinity ... did it really tend to $e^{\frac{\theta^2}{2}}$? - it didn't seem to tend to anything like that at all. Especially since the whole power was multiplied by lambda (i.e. raised to the power of lambda) which to me would suggest it would tend either to 0 or infinity.
Mind you, there again, 'e' can supposedly be found by $(1+ \frac{1}{n} )^n$ as n tends to infinity.
But bluh?!

By 'M4' I meant S4 (it really has been a long two days for me). I agree with what you said about f(th) either being zero or infinity. I had no idea what exactly they meant by that. BTW were you able to find the variance of lambda. I am not sure if what I did was correct (as I integrated.) I really didn't like that paper

I wonder how the other's found it, because if the other people who did S4 found it very good then the grade boundaries will be quite high. S4 isn't very complicated, it is just that the book didn't help me to get any sort of information. I have S3 on tuesday, so hopefully that will be much better (the text book is also much much much better for S3 that it is for S4.)