[Arsey]

Revision thread for the S1 Jan paper.

This is taken from the June 11 thread but has all the attachments you will find useful. I will put back up the June paper and model answers in the wee k of the Jan exam. I have taken them down as they may be used for mock exams.

I DO NOT HAVE ACCESS TO THE EXAMINATION PAPERS BEFORE THE EXAM, I ONLY HAVE ACCESS TO THEM ONCE THE PAPER IS COMPLETED

I will be happy to remove any content I have uploaded if requested by site admin

[Arsey]

Grade Boundaries and Past Papers

Session	100'	90'	80'	70'	60'	50'	40'
Jan-01	72	65	58	51	44	37	31
Jun-01	67	61	55	49	44	39	34
Jan-02	64	58	52	46	40	35	30
Jun-02	65	58	51	44	37	31	25
Nov-02	70	64	58	52	46	40	34
Jan-03	63	57	51	45	39	33	27
Jun-03	75	69	62	55	48	41	35
Nov-03	67	61	55	49	44	39	34
Jan-04	72	66	60	54	48	42	36
Jun-04	71	64	57	50	44	38	32
Nov-04	69	63	57	51	45	40	35
Jan-05	68	62	56	50	44	39	34
Jun-05	67	60	53	46	39	32	25
Jan-06	75	69	62	55	48	41	35
Jun-06	71	65	59	53	47	41	35
Jan-07	75	68	61	54	48	42	36
Jun-07	72	65	58	51	44	37	30
Jan-08	75	69	63	56	50	44	38
Jun-08	72	66	60	54	48	42	37
Jan-09	71	64	57	50	43	36	30
Jun-09	75	68	60	52	44	36	29
Jan-10	75	70	65	58	51	45	39
Jun-10	69	62	55	48	41	35	29
Jan-11	73	66	59	52	45	38	31

[Arsey]

Model Answers

[Arsey]

Revision Notes

Another good tip is that

Var = Sxx / n

The reason this is useful is that Sxx is given in the formulae book

[fudgesundae]

thanks for the revision notes, really useful

[Arsey]

More Model answers


Tip - An obvious one. Learn how to use your Stats mode on your calc to check your answers.

[fudgesundae]

(Original post by Arsey)
Tip - An obvious one. Learn how to use your Stats mode on your calc to check your answers.

I've never actually used that before. What sort of stuff can that mode do?

[Arsey]

More model answers


It can work out measures of location and dispersion.

If you enter the data for a frequency table, it will calculate all the sums (x, x^2); it will also calculate the mean and standard deviation (but not median or quartiles etc).

You can enter a bivariate table and it will calculate all of the sums (x, X^2, y, y^2, xy); it will calculate PMCC and the values of a and b in the least squares regression line.

The Casio 991ES can also do some normal distribution calculations.

[,,,,]

Arsey, cheers bro

[Extricated]

(Original post by Arsey)
It can work out measures of location and dispersion.

If you enter the data for a frequency table, it will calculate all the sums (x, x^2); it will also calculate the mean and standard deviation (but not median or quartiles etc).

You can enter a bivariate table and it will calculate all of the sums (x, X^2, y, y^2, xy); it will calculate PMCC and the values of a and b in the least squares regression line.

The Casio 991ES can also do some normal distribution calculations.

Hi arsey,

if i click stat mode i get a load of options.

If, for instance, i choose option (2) = a+bx then is this the regression line mode? when i click it i just get a table of x with a table of y next to it. How do i get it to be of use.

thanks again

[Groat]

(Original post by Arsey)
Model Answers

I see on the Jan 2011 paper, you've written:

then time, continuous.

If it weren't continuous, what would you have done differently?

[Arsey]

you are quite unlikely to see a discrete grouped data question. However, if the data were discrete the values in the tables would be the boundary points, not, as in this case, the midpoint between the classes.

So, if the data were discrete the median would be 0.5 higher.

[H94]

(Original post by Extricated)
Hi arsey,

if i click stat mode i get a load of options.

If, for instance, i choose option (2) = a+bx then is this the regression line mode? when i click it i just get a table of x with a table of y next to it. How do i get it to be of use.

thanks again

Activate state mode for a list of options on your calculator (doing this from memory, so please excuse any minor mistakes)

The list of options contain things such as 1-Var (the variance setting) and A+Bx (regression line)

Selecting the 1-Var setting gives you a table with X values - input them and then press AC. Then press Shift-1 to enter stat functions using your previously tabulated information.

Shift-1 gives you a few options, and heres what they do -

Type-change type of date
Data-edit table
Sum-Sigma X squared and Sigma X
Var - Have a guess.

If you had chosen the regression setting in the orignial list, you would get two tables X values and Y values, because it is regression, and then if you had chosen Shift-1, you would get the following options

Type-Same as before
Data-Same as before
Sum-All your Sigma Xs and Ys
Var - All your Var, means and Standard Deviation stuff
Reg - Regression stuff, you want option 3, the 'r' value

[Extricated]

(Original post by H94)
Activate state mode for a list of options on your calculator (doing this from memory, so please excuse any minor mistakes)

The list of options contain things such as 1-Var (the variance setting) and A+Bx (regression line)

Selecting the 1-Var setting gives you a table with X values - input them and then press AC. Then press Shift-1 to enter stat functions using your previously tabulated information.

Shift-1 gives you a few options, and heres what they do -

Type-change type of date
Data-edit table
Sum-Sigma X squared and Sigma X
Var - Have a guess.

If you had chosen the regression setting in the orignial list, you would get two tables X values and Y values, because it is regression, and then if you had chosen Shift-1, you would get the following options

Type-Same as before
Data-Same as before
Sum-All your Sigma Xs and Ys
Var - All your Var, means and Standard Deviation stuff
Reg - Regression stuff, you want option 3, the 'r' value

You are a legend, thanks! I don't get the last point though (probably because i'm thick), what do you mean by the 'r' value...i see it on the calculator but for regression don't you just need a and b? what does r tell u?

also can your calculator also find variance from grouped data
thanks

[Rahul.S]

(Original post by extricated)
you are a legend, thanks! I don't get the last point though (probably because i'm thick), what do you mean by the 'r' value...i see it on the calculator but for regression don't you just need a and b? What does r tell u?

Also can your calculator also find variance from grouped data
thanks

nt probably...ur are thick! Rip shu

[Arsey]

(Original post by Extricated)
You are a legend, thanks! I don't get the last point though (probably because i'm thick), what do you mean by the 'r' value...i see it on the calculator but for regression don't you just need a and b? what does r tell u?

also can your calculator also find variance from grouped data
thanks

most scientific calcs can find standard deviation in the stats mode, square that to give variance.

on the casio calcs you need to turn on freq

mode - down arrow - stats - frequency on

shift 1 - stats mode - data

when you put the data in you must use the mid-points as the first column of data and then the freq in the second column.

the only thing it will not do is interpolate to find the median.

[pHneutral]

Question:
The random variable X ~ N(mean, sigma^2)
The lower quartile of X is 25 and the upper quartile of X is 45.
Find the value of mean , and the value of sigma

I dont understand how one would do this question systematically?
thanks!

[sohail.s]

(Original post by pHneutral)
Question:
The random variable X ~ N(mean, sigma^2)
The lower quartile of X is 25 and the upper quartile of X is 45.
Find the value of mean , and the value of sigma

I dont understand how one would do this question systematically?
thanks!

Think about it, the lower quartile is a point below which there is 25% of data
Upper quartile is a point below which 75% data is.

We're told X ~ N (mu, sigma^2)

Forming equations now

P(Z<(25-mu)/sigma) = 0.25 equation 1 (25% of data is less than lower quartile)
P(Z<(45-mu)/sigma) = 0.75 equation 2 (75% of data is less than upper quartile)

What you'd do next is form two simultaneous equations, eliminate mu or sigma, solve for one and substitute to find the other.
Be careful with signs and phi value from tables when forming the two equations

Can't really do the rest as i don't have tables with me atm

That should help, though i have a feeling it might be wrong since i haven't touched stats for a few weeks now.

[sohail.s]

(Original post by Extricated)
You are a legend, thanks! I don't get the last point though (probably because i'm thick), what do you mean by the 'r' value...i see it on the calculator but for regression don't you just need a and b? what does r tell u?

also can your calculator also find variance from grouped data
thanks

r tells you the PMMC correlation co-efficient - dont just copy this and write it down since you could lose marks for accuracy and no working etc.

It's a useful check to use your calculator at the end to check your answers.


For grouped data like arsey said just plug in midpoints.

[No.1 Loner]

I dont know if i'll want to check, if I realise i failed it'll kill the rest of my holiday.