!! HL Maths 2015 - predictions, guesses, Q & A, discussion ...

numbers13

Kremin wrote: »

Could anyone tell me the difference between the two means? Some questions are throwing in two means, like x bar and mew or whatever.. Whats the difference??

I believe the difference between them is like the difference between p and p-hat. x bar is the mean of a sample of the population, where mew is the mean of the population as a whole.

Bah_Humbug

In trig do we just need to derive formulae for cos(A-B) and cos(A+B) or are there others? When people say 1-7 and 9, what are they? Because nothing in our book is numbered like that..

iang198

What proofs do we need to know?
I know Sine Rule and cosine rule but that's about it! Do we need to know more?

135083

Bah_Humbug wrote: »

In trig do we just need to derive formulae for cos(A-B) and cos(A+B) or are there others? When people say 1-7 and 9, what are they? Because nothing in our book is numbered like that..

I'm assuming they mean 1-7 and 9 on page 45 of this document (warning, pdf) www(dot)ncca(dot)ie/en/Curriculum_and_Assessment/Post-Primary_Education/Project_Maths/Syllabuses_and_Assessment/LC_Maths_English_2013.pdf

skippy1977

Bah_Humbug wrote: »

In trig do we just need to derive formulae for cos(A-B) and cos(A+B) or are there others? When people say 1-7 and 9, what are they? Because nothing in our book is numbered like that..

Syllabus numbers them.

Here they are:

Kremin

skippy1977 wrote: »

Syllabus numbers them.

Here they are:

I actually hope one or more of these come up, so handy and quite easy to remember and even deduce (the later ones anyway)

Troxck

Kremin wrote: »

I actually hope one or more of these come up, so handy and quite easy to remember and even deduce (the later ones anyway)

What's the best way to learn these? I find the Theorems fine because all you have to do is remember a diagram and you can do it yourself

fin709

How likely are proving those trig identities like sin rule etc. to come up? Haven't they only been on last year since 2010?

Fiona G

Dreading geometry questions where you have to prove something without any values given... just give me a length or angle to find please don't make me prove something...

Kremin

Fiona G wrote: »

Dreading geometry questions where you have to prove something without any values given... just give me a length or angle to find please don't make me prove something...

This will be the only thing to trip me up, honestly.
I can do ANYTHING but prove that |od|=1/2OFOSOMFSOFM or whatever, its all gibberish to me. And everyones like GEOMETRY IZ EZ.

Kremin

Anyone have the solutions to Q8 on the debs mock? (the new inferential question)

Dkey

Kremin wrote: »

Anyone have the solutions to Q8 on the debs mock? (the new inferential question)

Is it the question about light bulbs?

RoRo979

ah now lads, what is this feckin mew thing?? i always thought this is all we need

Margin of error = E = 1/root(n) or 1.96root(p-hat(1- p-hat)/n)
central limit everything stays but standard deviation = standard deviation/root(n)
z score = x - mean/stan dev
mean hypothesis testing = mean -1.96(stan dev/root(n)) and vise versa
P values = xbar - mean/stan dev//root(n)

is that all?

Fiona G

RoRo979 wrote: »

ah now lads, what is this feckin mew thing?? i always thought this is all we need

Margin of error = E = 1/root(n) or 1.96root(p-hat(1- p-hat)/n)
central limit everything stays but standard deviation = standard deviation/root(n)
z score = x - mean/stan dev
mean hypothesis testing = mean -1.96(stan dev/root(n)) and vise versa
P values = xbar - mean/stan dev//root(n)

is that all?

Mew is just a symbol for the population mean. It looks like a 'u' with a longer tail, you'll see it in the tables.
Yes I think that's all the formulae we need. The margin of error for a mean and proportion are the only things not in the tables I think.

RoRo979

Kremin wrote: »

Anyone have the solutions to Q8 on the debs mock? (the new inferential question)

I got 24/25 on the B part, silly mistake you will see it. But didnt do the A part. However just did the A part there so ill stick it up and tell me if it looks right aswell

DarraghF197

Kremin wrote: »

Anyone have the solutions to Q8 on the debs mock? (the new inferential question)

I'm just doing it now!

I have the marking scheme on a sheet so I can give you the answers, or I can take a photo of them if you want.

8. a €1048.79<u<€1,251.21

ii we can be 95% certain that the mean monthly rent for a one-bedrrom apartment on Dublin lies between (above)

iii 2.905

Conclusion: reject null and accept alternative.

iv) p value. 0038. Since 38 is less than. 05 the result is significant and we reject the null hypothesis

I)0.21<P<0.31

Ii) N>1848


This is really confusing me. I have a couple of sheets photocopied from idk what book. I've already found one place in which it's wrong, and looking at it again, it appears that the confidence interval formula appears to be wrong as well. I think it might be the reason why I'm finding this section confusing.

In here, it says the confidence interval is 1.96[ sd (of population)/root n]. However looking at the marking schemes for this mock, they used 1.96[ sd(sample)/root n]. Can someone clarify which one is right?

Kremin

RoRo979 wrote: »

ah now lads, what is this feckin mew thing?? i always thought this is all we need

Margin of error = E = 1/root(n) or 1.96root(p-hat(1- p-hat)/n)
central limit everything stays but standard deviation = standard deviation/root(n)
z score = x - mean/stan dev
mean hypothesis testing = mean -1.96(stan dev/root(n)) and vise versa
P values = xbar - mean/stan dev//root(n)

is that all?

As far as I know, you never use 1/rootn anymore, unless told to use it....

"In fact 0.98 ( ___1
√
__
n
) < ___1
√
__
n
. We can say that at the 95% confidence level, the margin of error for a population
proportion will not exceed ±___1
√
__
n
, where n is the sample size. While this approximation gives wider confidence
intervals, it is easy to construct. On our course we will use the margin of error E = ±1.96√
________
^p(1 – ^p) ________
n unless
specifically asked to use the ±___1
√
__
n
margin of error."

dalta5billion

Kremin wrote: »

This will be the only thing to trip me up, honestly.
I can do ANYTHING but prove that |od|=1/2OFOSOMFSOFM or whatever, its all gibberish to me. And everyones like GEOMETRY IZ EZ.

I know, I hate that notation so much too. What I do is rename the sides x, y, z, a, b, c and the angles opposite them X, Y, Z, A, B, C.

I find when you can represent them in algebraic form it's a good bit easier to form equations.

RoRo979

1/√N can still be used but next year will not be allowed.
Its easier and quicker to use than the 1.97√p-hat(.....)/n one

Kremin

Ok. Could someone just sum this up for me real quick.
When do we use the normal z code formula x-u/s and when do we have to use x-u/s/rootn?
When should we use standard error of the proportion and when do we use the sampling formulas in general?

Kremin

RoRo979 wrote: »

1/√N can still be used but next year will not be allowed.
Its easier and quicker to use than the 1.97√p-hat(.....)/n one

Are you sure it can still be used? I got 0 for it in the mocks when I used it.

RoRo979

attached pictures of q8 there however cant remember now if it is debs or examcraft, could some just check the first question and ensure I am right thanks

Kremin wrote: »

Are you sure it can still be used? I got 0 for it in the mocks when I used it.

I went to DublinSchool of grinds over Easter and they said it is fine this year. I also got full marks in my mock using it.

When do we use the normal z code formula x-u/s and when do we have to use x-u/s/rootn?
When should we use standard error of the proportion and when do we use the sampling formulas in general?

x-u/s = just general use for calculating z-score etc.
x-u/s/rootn = called central limit theorem. When multiple large sample groups are chosen. So question will say many groups of 20/30/40/50 people where questioned. The mean is the same for both, but the standard deviation = s/√n
The shape of the distribution is also normal for central limit.
x-u/s/rootn is also used for Test Statistic in calculating P values

Lucky Leprechaun

Would anyone know if there is a Sample Paper for stand 2? I can only find strand 1. Thanks

OMGeary

What are the formulae that need to be learnt off?

dazzadazza

DarraghF197 wrote: »

I'm just doing it now!

I have the marking scheme on a sheet so I can give you the answers, or I can take a photo of them if you want.

8. a €1048.79<u<€1,251.21

ii we can be 95% certain that the mean monthly rent for a one-bedrrom apartment on Dublin lies between (above)

iii 2.905

Conclusion: reject null and accept alternative.

iv) p value. 0038. Since 38 is less than. 05 the result is significant and we reject the null hypothesis

I)0.21<P<0.31

Ii) N>1848


This is really confusing me. I have a couple of sheets photocopied from idk what book. I've already found one place in which it's wrong, and looking at it again, it appears that the confidence interval formula appears to be wrong as well. I think it might be the reason why I'm finding this section confusing.

In here, it says the confidence interval is 1.96[ sd (of population)/root n]. However looking at the marking schemes for this mock, they used 1.96[ sd(sample)/root n]. Can someone clarify which one is right?

Anyone know how the p value is gotten? In this case 0.0038 in part (iv) of (a).

DarraghF197

dazzadazza wrote: »

Anyone know how the p value is gotten? In this case 0.0038 in part (iv) of (a).

It's P= 2[1-p(z<2.905)]

That exact z-score isn't in the log tables, but you just half the sum of z<2.9 and 2.91.

dazzadazza

DarraghF197 wrote: »

It's P= 2[1-p(z<2.905)]

That exact z-score isn't in the log tables, but you just half the sum of z<2.9 and 2.91.

Oh I see. Thank you. So the p value is just that table in the log tables?

BlueWolf16

so wait, we can still use the normal equation for conf. intervals? like x - mean / 1/root N ?? We have a choice between the two ( x - mean / sd/root N) in whatever question we want, or is there some kind of specific one that you use each for? They will give you dif. answers, so will both be marked correct?

DarraghF197

BlueWolf16 wrote: »

so wait, we can still use the normal equation for conf. intervals? like x - mean / 1/root N ?? We have a choice between the two ( x - mean / sd/root N) in whatever question we want, or is there some kind of specific one that you use each for? They will give you dif. answers, so will both be marked correct?

X bar is used on its own, multiplied by 1.96. This is the standard error, and you do NOT divide it by root n.

When you are given the standard deviation, you divide this by the root of number of samples taken. And, of course, multiply by 1.96

Edit: I just read over your question lol. Ahm, I wouldn't recommend using 1 over root n, you can only use that when approximating the confidence interval.

dazzadazza wrote: »

Oh I see. Thank you. So the p value is just that table in the log tables?

The p-value is the probability of getting a value as extreme or more extreme than the test statistic (z score of the claimed mean against the sample mean) obtained. If it's greater than five percent, the result is not significant and we do not reject the H0, and vice versa.

dazzadazza

BlueWolf16 wrote: »

so wait, we can still use the normal equation for conf. intervals? like x - mean / 1/root N ?? We have a choice between the two ( x - mean / sd/root N) in whatever question we want, or is there some kind of specific one that you use each for? They will give you dif. answers, so will both be marked correct?

Tbh I don't think your allowed use one over root n. Just to be safe I'd always go with the other one.