The Student Room Group
Uni students - Complete our survey >>
Uni students - Complete our survey >>
Uni students - Complete our survey >>

Edexcel S3 - Wednesday 25th May AM 2016

Scroll to see replies

Original post by Random1357
This paper also had several tricks (CLT not at assumption, lose 2 DoF for a chi squared test)


This question was 2 marks. I wrote two assumptions, one that

"the sample size was sufficiently large for the central limit theorem to render the sample mean approximately normally distributed"

and two

"s and mu are unbiased estimates of the population standard deviation and mean respectively"

Is this not correct?
Original post by Euclidean
This question was 2 marks. I wrote two assumptions, one that

"the sample size was sufficiently large for the central limit theorem to render the sample mean approximately normally distributed"

and two

"s and mu are unbiased estimates of the population standard deviation and mean respectively"

Is this not correct?


Neither sound like assumptions, to be honest.

For the first: you don't assume the sample size is large enough. The question either tells you that it is or isn't. i.e" if the question says the sample size is n=80, you don't "assume the sample size is n>50 so CLT applies)

[edit:]

Second, it's again not an assumption that s and mu are unbiased estimates of population standard deviation. They are unbiased estimates, I can prove it if you'd like. The assumption is that we assume they are equal to the population standard deviation and mean which isn't true in general.
(edited 7 years ago)
Original post by Euclidean
This question was 2 marks. I wrote two assumptions, one that

"the sample size was sufficiently large for the central limit theorem to render the sample mean approximately normally distributed"

and two

"s and mu are unbiased estimates of the population standard deviation and mean respectively"

Is this not correct?


Since the normal condition (no pun intended) is that n>50 for the central limit theorem to hold, I think it's reasonable to say that the CLT is a bit of an assumption in this instance (n=36 and 42 iirc) since it's by no means (no pun intended) ensured to be dead-on accurate.

Posted from TSR Mobile
Original post by Krollo
Since the normal condition (no pun intended) is that n>50 for the central limit theorem to hold, I think it's reasonable to say that the CLT is a bit of an assumption in this instance (n=36 and 42 iirc) since it's by no means (no pun intended) ensured to be dead-on accurate.

Posted from TSR Mobile


You totally intended those puns.
Original post by Zacken
For the first: you don't assume the sample size is large enough. The question either tells you that it is or isn't. i.e" if the question says the sample size is n=80, you don't "assume the sample size is n>50 so CLT applies)


Original post by Krollo
Since the normal condition (no pun intended) is that n>50 for the central limit theorem to hold, I think it's reasonable to say that the CLT is a bit of an assumption in this instance (n=36 and 42 iirc) since it's by no means (no pun intended) ensured to be dead-on accurate.Posted from TSR Mobile


I must have missed a chapter with this n>50 condition. But the central limit theorem describes the tending of the sum of multiple variables to a normal distribution, when n is relatively quite large the central limit theorem acts as the basis for us to say X bar (in this case) is approximately normally distributed but the judgment of n being 'sufficiently large' is an assumption in itself is it not?

Original post by Zacken
Second, it's again not an assumption that s and mu are unbiased estimates of population standard deviation. They are unbiased estimates, I can prove it if you'd like. The assumption is that we assume they are equal to the population standard deviation and mean which isn't true in general.


The question threw me as they never specified that they were actually unbiased estimates in the question. Although looking at it now your assumption makes a lot more sense as an answer :doh:

Edit: I didn't catch your edit when I posted this :tongue:
(edited 7 years ago)
Original post by Euclidean
This question was 2 marks. I wrote two assumptions, one that

"the sample size was sufficiently large for the central limit theorem to render the sample mean approximately normally distributed"

and two

"s and mu are unbiased estimates of the population standard deviation and mean respectively"

Is this not correct?


I agree with zacin that sadly the second is unlikely to attract credit, though I may hold out some hope for the former (I essentially put it too and I think it's justifiable as an assumption).

Posted from TSR Mobile
Original post by Krollo
I agree with zacin that sadly the second is unlikely to attract credit, though I may hold out some hope for the former (I essentially put it too and I think it's justifiable as an assumption).

Posted from TSR Mobile


Thanks both of you :smile:
Reply 1427
Avatar for Inges
Inges
3
For all those saying CLT is an assumption.

From what I remmember, The question said "You may assume they are normally distributed"

Therefore the two assumptions are
1) S2 = σ2 .
2 ) They are independent

The second one is because the question did not state the samples were independent and for you to use the big formula involving both X & Y. They have to be independent.

Correct me if i'm wrong please :tongue:
Original post by Inges
For all those saying CLT is an assumption.

From what I remmember, The question said "You may assume they are normally distributed"

Therefore the two assumptions are
1) S2 = σ2 .
2 ) They are independent

The second one is because the question did not state the samples were independent and for you to use the big formula involving both X & Y. They have to be independent.

Correct me if i'm wrong please :tongue:


You are correct.


Posted from TSR Mobile
Reply 1429
Avatar for Inges
Inges
3
Original post by physicsmaths
You are correct.


Posted from TSR Mobile


Yay, no more doubts!
Anyone remember how many marks the very last part was worth, where n=44? Thanks.
Original post by paradoxequation
Anyone remember how many marks the very last part was worth, where n=44? Thanks.


Believe it was 4 marks.

With only 1 or 2 dedicated to actually getting it right I wager.
Original post by Krollo
Since the normal condition (no pun intended) is that n>50 for the central limit theorem to hold, I think it's reasonable to say that the CLT is a bit of an assumption in this instance (n=36 and 42 iirc) since it's by no means (no pun intended) ensured to be dead-on accurate.

Posted from TSR Mobile

But in the questions it already said you may assume they are bormally distributed.
Original post by physicsmaths
But in the questions it already said you may assume they are bormally distributed.


Oh, fair enough.

Posted from TSR Mobile
Original post by Krollo
Oh, fair enough.

Posted from TSR Mobile


How did it go for you? 100?


Posted from TSR Mobile
Original post by physicsmaths
How did it go for you? 100?


Posted from TSR Mobile


Nah, I'll have lost one there, but everything else was fine I think.

Posted from TSR Mobile
Original post by Krollo
Nah, I'll have lost one there, but everything else was fine I think.

Posted from TSR Mobile


Hopefully still 100 UMS tho, im happy i think i got a B atleats haha. You sitting S4 M4-5 aswell?


Posted from TSR Mobile
Original post by physicsmaths
Hopefully still 100 UMS tho, im happy i think i got a B atleats haha. You sitting S4 M4-5 aswell?


Posted from TSR Mobile


Nah doubt it, it was a fairly easy paper. My combo has been broken, cri :-(

In all seriousness I don't particularly care, it's only stats. Not doing S4 but I am going for M4 and M5... mechanics is much better tbh

Posted from TSR Mobile
Original post by Krollo
Nah doubt it, it was a fairly easy paper. My combo has been broken, cri :-(

In all seriousness I don't particularly care, it's only stats. Not doing S4 but I am going for M4 and M5... mechanics is much better tbh

Posted from TSR Mobile


If the A boundary is above 60 and an even number 100ums will be 74 :smile:.


Posted from TSR Mobile
Original post by physicsmaths
If the A boundary is above 60 and an even number 100ums will be 74 :smile:.


Posted from TSR Mobile


Weird thing is, I would probably be a lot less stressed about exams if I knew I'd dropped marks already... at the moment it seems teachers / friends / folk online are kind of expecting me to get full ums again when I know it isn't realistically possible. Knowing that I've already dropped marks can make things much easier, since beyond that A Levels serve little real purpose for me provided I don't completely **** them up.

My brain is a strange place, sorry for pouring it out onto the s3 thread guys :tongue:

Posted from TSR Mobile

Quick Reply

Related discussions

Latest

Trending

Trending

Articles for you