OCR (not MEI) S2 - 22/06/2011

Contrad!ction.

15

Who are the other lucky ones taking Stats 2? How are the distributions and tests going?

Reply 1

Contrad!ction.

OP

15

Just thought I'd bump this as we're nearing the exam now. I don't think I'm the only one taking this from TSR - or am I? :eek:

Reply 2

Mousebudden

3

I am taking it.. Not at all looking forward to it

Reply 3

Contrad!ction.

OP

15

Original post by Mousebudden

I am taking it.. Not at all looking forward to it

What's worrying you? For me it's getting the corrections right on hypothesis testing. And writing things.

Reply 4

Gamma

I am taking it but I only did 3 hours revision today which isn't a whole lot considering I feel a bit unprepared for this...

Reply 5

Contrad!ction.

OP

15

I feel sort-of-prepared-but-at-the-same-time-not-getting-as-many-marks-as-I'd-hope.

I'm too busy focusing on Monday's M1 exam to work on the S2 one at the moment... =/

Reply 6

Freakonomics123

10

Original post by Gamma

I am taking it but I only did 3 hours revision today which isn't a whole lot considering I feel a bit unprepared for this...

As long as you dont have any exams from now till wednesday you should be fine.

My plan is to go over all the content tomorow, then revise that a bit more on monday, and do two 2007 past papers then on tuesday do the rest of the ocr papers.

Big challenge will be learning all of physics unit 5 in 4 days....sigh

Reply 7

Morkike

This exam will be fine unless they bring in an obscure question -which is likely considering they are trying to make grade boundaries lower.

Reply 8

Contrad!ction.

OP

15

Just got my mock paper back - 63/72, and a couple of silly mistakes. I'm much more confident about this one now, and definitely think I'll do better in this than in S1/M1/C1.

Rolling out the last 5 past papers tomorrow, let's see how they go...

Reply 9

mgmt11

Anyone have the mark scheme for jan 2011 ? :|

Reply 10

Freakonomics123

10

Original post by Contrad!ction.

Just got my mock paper back - 63/72, and a couple of silly mistakes. I'm much more confident about this one now, and definitely think I'll do better in this than in S1/M1/C1.

Rolling out the last 5 past papers tomorrow, let's see how they go...

Original post by mgmt11

Anyone have the mark scheme for jan 2011 ? :|

is the exam in morning or afternoon??

Reply 11

Contrad!ction.

OP

15

Original post by Freakonomics123

is the exam in morning or afternoon??

It's a 9am exam - morning. :smile:

Reply 12

Sawasdee

Shall we make some revision notes together? e.g.

For Binomial distribution X~B(n,p)
Approx X by Normal distribution IF np>5 , nq>5 , and n is large (how big? I can't remember) Y~N(np, npq)
Approx X by Poission distribution IF np<5 Y~P(np)

For Poisson distribution A~P(lamda)
Approx A by Normal IF lamda>15 Y~N(lamda,lamda)
Approx A by Binomial ???? I don't think so.

For Central Limit Theorem
size of the sample, n, has to be greater than 50 (am I right?) so it has a Normal distribution

phewww what else we are supposed to know?

Reply 13

warrior_22

Can someone explain to me how to answer Chi squared questions when they ask to compare the data using the contributions to the test statistics. When I check the mark schemes, they have two ways of answering, they use strong positive /strong negative association and the other one is more frequent and distributed.

Reply 14

MH1992

Can someone explain to me the simplest way possible the two types of error.
Am i right in saying that type 1 errors are just the critical region like for example 10 %. ???

Reply 15

Freakonomics123

10

Original post by warrior_22

Can someone explain to me how to answer Chi squared questions when they ask to compare the data using the contributions to the test statistics. When I check the mark schemes, they have two ways of answering, they use strong positive /strong negative association and the other one is more frequent and distributed.

Original post by Sawasdee

Shall we make some revision notes together? e.g.

For Binomial distribution X~B(n,p)
Approx X by Normal distribution IF np>5 , nq>5 , and n is large (how big? I can't remember) Y~N(np, npq)
Approx X by Poission distribution IF np<5 Y~P(np)

For Poisson distribution A~P(lamda)
Approx A by Normal IF lamda>15 Y~N(lamda,lamda)
Approx A by Binomial ???? I don't think so.

For Central Limit Theorem
size of the sample, n, has to be greater than 50 (am I right?) so it has a Normal distribution

phewww what else we are supposed to know?

Original post by Contrad!ction.

It's a 9am exam - morning. :smile:

if you can be bothered dont have to, can someone explain why on june 09 5iii they divide the final value for lambda by 3?? i would post question but its a bit long

Reply 16

Rational Paradox

9

Original post by Sawasdee

Shall we make some revision notes together? e.g.

For Binomial distribution X~B(n,p)
Approx X by Normal distribution IF np>5 , nq>5 , and n is large (how big? I can't remember) Y~N(np, npq)
Approx X by Poission distribution IF np<5 Y~P(np)

For Poisson distribution A~P(lamda)
Approx A by Normal IF lamda>15 Y~N(lamda,lamda)
Approx A by Binomial ???? I don't think so.

For Central Limit Theorem
size of the sample, n, has to be greater than 50 (am I right?) so it has a Normal distribution

phewww what else we are supposed to know?

n has to be bigger or equal to 30 for a sample of means to approximate to a normal distribution

If n is 50 or above, that constitutes as large in terms of the distributional approximations (aside from the CLT)

And if approximating a discrete distribution to a continuous distribution, you always have to use a continuity correction

Reply 17

Xero Xenith

10

Pretty psyched for this TBH - really hoping for 90+ to secure my A* Further Maths. By "hoping" I really mean "hoping"!

Here's the Jan 11 paper - apologies for the not-scanned-ness but I don't have much time right now. It's readable anyway, and markscheme included. Just click the pictures for better quality :smile:

http://xeroxenith.imgur.com/ocr_s2_jan_2011

Just throwing this out there because I got given it today and I was having trouble finding it before!

(edited 12 years ago)

Reply 18

Sawasdee

Original post by Rational Paradox

n has to be bigger or equal to 30 for a sample of means to approximate to a normal distribution

If n is 50 or above, that constitutes as large in terms of the distributional approximations (aside from the CLT)

And if approximating a discrete distribution to a continuous distribution, you always have to use a continuity correction

But from this link, http://www.mathshelper.co.uk/OCR%20S2%20Revision%20Sheet.pdf
n>30 for CLT
and n>50 if you want to approx binomial by poisson.
there's nothing about the size of n for approxing binomial by normal

I have no idea which one is right, I don't have notes with me :sigh:

Reply 19

Rational Paradox

9

Original post by Sawasdee

But from this link, http://www.mathshelper.co.uk/OCR%20S2%20Revision%20Sheet.pdf
n>30 for CLT
and n>50 if you want to approx binomial by poisson.
there's nothing about the size of n for approxing binomial by normal

I have no idea which one is right, I don't have notes with me :sigh:

As far as i remember (did this all the way back in Jan so i may be rusty)

B(n,p) -> N(np,npq), for np>5,nq>5 (Continuity correction required)
B(n,p) -> Po(np), for n>50,np<5 (No continuity correction required)
Po(x) -> N(x,x), for x>15 (Continuity correction required)

So the size of n doesn't matter for the Binomial -> Normal approx, just as long as it satisfies the np, nq condition

and n has to be bigger than (or equal to 30) for a sample of means for it to be a good approximation to the normal by the Central limit theorem

(edited 12 years ago)

OCR (not MEI) S2 - 22/06/2011

Quick Reply

Related discussions

Latest

Trending

Trending

Articles for you