OCR statistics 3 23 june 2011

Ahhh I see... Sorry!

I took the exam, thought it was quite difficult compared to previous exams because there were fewer questions with 7+ marks which is what I'm best at! Not really confident with many of my answers, can you remember any of yours?

1) 22 and 56 (although ive been told the 56 was wrong because it should have been independent)
for the second part I think got them both as 3/2
I think I got a as 2/3 :/
I remember having 4/3 for the probabilty distribution of Y
think i had answers like 31/18 and 19/36 for somthing?
think i got n=150 for something
X^2 = 1.45 for p(x=6)
got roughly 4.5 for the last question
i think one of my confidence intervals was [0.017,0.134]
got the sample mean as 26.05 and a confidence interval of roughly [25.??,27.??]

Really struggling to remember, but like I said im not too confidence in any of those answers!

Hey guys, I took this exam today too. A have similar answers for the most part with the following additions:

1)(i) E[X] = Var(X) = 22, since there is no amplification involved.
(ii) Both equal 3/2
(iii) Both can be modelled by a poisson distribution (mean = variance)

5)(i) F(m) = 1/2... can't remember the exact answer.
(ii) Y = 1/(X^2) >> G(y) = 1/3(4y - 1) for 1/4 < y < 1
or G(y) = 0 for y<1/4, and G(y) = 1 for y>1.
Hence g(y) is 4/3 for 1/4 < y < 1, and zero otherwise.
(iii) I personally got 3/4 as my value for E[2+2X^-2] since I rewrote it as E[2+2Y], then solved the integral between 1 and 1/4 of (4/3)*(2+2y) dy


7(i) I put that two sample t test is not appropriate since each data subject / student gives two data values.
(ii) t(crit) = 3.250, t(calc) = 4.699. Hence reject Ho etc.
(iii) t(crit) = 3.250, t(calc) was 2 point something. Hence accept Ho >> insufficient evidence to support the claim.

Other random answers since I can't remember the question numbers:

PDF question:
a = 6/5
0.9, 0.95

Chi squared question:
The x=6 (or whatever value it was) contributed 1.45 to chi squared
Calculated chi squared was about 13, critical was about 15 so accept Ho and conclude that X~Po(3.47) fits the data well.

Confidence intervals:
[0.017,0.134]
[25.x,27.x]

I will edit this if I remember any more.
Hope this helps.
Andy

Anyone know if q6 (bmi question) requires x bar? For part i. I did xbar +\- z (Square root variance/n) for confidence interval but didnt use xbar for p(x>30) (part ii). Were we supposed to use xbar? And therefore divide the variance by n?

Oh and I think Q7 I got accept for part ii and reject for part iii. And I put that the 2nd test results are not independent from the first so the other t test can't be used :/.

Anyone know if q6 (bmi question) requires x bar? For part i. I did xbar +\- z (Square root variance/n) for confidence interval but didnt use xbar for p(x>30) (part ii). Were we supposed to use xbar? And therefore divide the variance by n?

Oh and I think Q7 I got accept for part ii and reject for part iii. And I put that the 2nd test results are not independent from the first so the other t test can't be used :/.