S1 Jan 2012 Edexcel Post Exam Discussion Thread - Solutions in the first post

For some reason I can't see the solutions in the first post ?

Furthermore, the data given is not possible. I appreciate it states 'without further calculation' but if you were to remove the fake coin, reduce your data set to 9 coins, calculate new values of Stt, Sww, Stw.

Stw changes to a positive value, meaning that there is now evidence of positive correlation, so PMCC WOULD INCREASE Numerically (if not statistically)

Sww is negative, which is impossible.

Other than Q2 I thought this was quite an easy paper, marks seem quite generous too.

Q1. Fairly easy histogram question (7)

Q2. Pretty tricky Venn diagram (9)

Q3. Standard DRV question (11)

Q4. Pretty easy/standard measures of location/spread question (11)

Q5. Pretty easy correlation/regression (issues with part f) (15)

Q6. Very easy Venn

Q7. Pretty simple normal distribution

I would expect fairly high boundaries for this paper

My guess would be

100UMS = 75
90 UMS = 70
80 UMS = 65
70 UMS = 58
60 UMS = 51
50 UMS = 45
40 UMS = 38

EDIT - This has been corrected on V2 of the solutions

EDIT - error on 4e)

I used the calculated mean and s.d. from earlier in the qu. For some reason, new values of the mean and s.d were given for part e)

scaled mean = 45
scaled sd = 9

(if you have made the same error, it is almost certainly 2 marks gone).



Edit 5f)

This question is very naughty.

From the calculations it would appear that the fake 5 year old coin weighing 20g is an outlier. Statistically, removing an outlier from a data set should increase the correlation between the two sets of data (It should make any formula generated more reliable).

So we would say PMCC increases, hence my solution.

However, given that the value of PMCC was -0.908, it would get closer to -1. So the value of the negative number should decrease. That said, you can understand why someone would say it is getting bigger, the negative number is getting bigger....

I have no idea how they will mark this one.

Furthermore, the data given is not possible. I appreciate it states 'without further calculation' but if you were to remove the fake coin, reduce your data set to 9 coins, calculate new values of Stt, Sww, Stw.

Stw changes to a positive value, meaning that there is now evidence of positive correlation, so PMCC WOULD INCREASE Numerically (if not statistically)

Sww is negative, which is impossible.


With such a small data set, 1 extreme piece of data will have a massive impact, so having an original PMCC so close to -1 doesn't make any sense in the first place.

i one of the previous question they said area= 0.2 if within range
So prob not in range is 1-0.2=0.8

Then it was just
In range x Not in range + Not in Rangex In range Or one of those multiplied by 2

I did it but forgot the x2

[Disclaimer: I'm sure you're better than I am on the S1 syllabus, but my thoughts anyhow...]

To be honest, since the question is clear about asking whether the value increases or decreases, (rather than the correlation), it's fairly clear the correct answer is "decrease". I agree it's harsh on people dropping marks due to misinterpretation, though.

Just to confirm:

For 10 coins: $\sum w^2 = S_{ww} + (S_w)^2/10 = 0.16 + (111.75)^2/10 = 1248.99625$
So after removing a coin weighing 20 grams: $\sum w^2 = 848.99625, \sum w = 91.75$. Then $S_{ww} = 848.99625 - (91.75)^2 / 9 = -86.3740277...$

So yes, those figures are not mathematically possible.

I'm appalled at this. Why on earth you wouldn't use a "real" dataset for this (by which I mean, 10 values picked to have roughly the right amount of correlation), I have no idea.

if I didn't multiply by 2 in the last question how much will I lose ???


also in question of given he runs find p of swim and not cyle
if I miss read the value of p(R) insted of 65 i wrote 55 how much will I lose

can you look at this post plz

Well I'm certain at least two super posh boys sitting to the left and right of me weren't.. they finished paper so quick and looked really bored ..

And I'm sure you have done fine, you seem to be good at the ''harder'' stuff, if it doesn't go according to plan, then retake in June

if I didn't multiply by 2 in the last question how much will I lose ???


also in question of given he runs find p of swim and not cyle
if I miss read the value of p(R) insted of 65 i wrote 55 how much will I lose

What about S2 MS Arsey??

[Disclaimer: I'm sure you're better than I am on the S1 syllabus, but my thoughts anyhow...]

To be honest, since the question is clear about asking whether the value increases or decreases, (rather than the correlation), it's fairly clear the correct answer is "decrease". I agree it's harsh on people dropping marks due to misinterpretation, though.

Just to confirm:

For 10 coins: $\sum w^2 = S_{ww} + (\sum w)^2/10 = 0.16 + (111.75)^2/10 = 1248.99625$
So after removing a coin weighing 20 grams: $\sum w^2 = 848.99625, \sum w = 91.75$. Then $S_{ww} = 848.99625 - (91.75)^2 / 9 = -86.3740277...$

So yes, those figures are not mathematically possible.

I'm appalled at this. Why on earth you wouldn't use a "real" dataset for this (by which I mean, 10 values picked to have roughly the right amount of correlation), I have no idea.

if it 3sf or more it should be okay as long as you didn't round too early in the question

I'm appalled at this. Why on earth you wouldn't use a "real" dataset for this (by which I mean, 10 values picked to have roughly the right amount of correlation), I have no idea.