Edexcel S3 - Wednesday 25th May AM 2016

BTW, if anybody gets their hand on the S3 UK/IAL papers after the exam, send them over to me, I might do model solutions if there's demand.

Ayman probably cud?
I would love it if you did.


Posted from TSR Mobile

As above, see my post. In your question, the expected value of the difference is $0\,$ allowing us to use the symmetry and get to the expression, this is not true for the other question.

This paper, the first question the mean is 100 and so we couldn't use the 2P(...) thing but the mark scheme allows it as a valid method that gets the marks. Could you please explain ?

https://869d950bf149eefd5cd2652b400160ceb38ca5c7.googledrive.com/host/0B1ZiqBksUHNYVGVDMmVJYzVnRnc/Mock%20QP%20-%20S3%20Edexcel.pdf

Hi, I want to ask regarding to the accuracy required in the exam of Statistics 3, Edexcel. Because in the formula booklet, sometimes, in normal distribution, a certain value does not give you the probability or z in the required accuracy. So I wonder if it matters? (In the mark scheme, they always use values of 4 decimal places, but that value is not in the booklet). So where do I get those values?

Whatever page you're on for the normal tables, move one page down to a table called "percentage points" those give 4 d.p values for specific probabilities and these are the ones that are mainly used in S3.

1. It's $>$ which is fundamentally different to $<$. If you sketch the normal distribution and compare the two, you'll see how different they are.

2. They're not quite using the form that's being asked by the previous users, the other users were asking about 2P(...) - 1. Which is missing from this.

3. All they've done is just combined the two probabilities into one because even if the mean isn't 0, there's still symmetry about the point x=100.

4. Sketch the normal dist. graph, it'll make sense.

Ohh I see, I never payed attention to that table! That's why sometimes they used values of 4 decimal places and sometimes 2 decimal places(I guessed that if the value is not in that table, then I would use the 2 decimal places table)! Thanks! Zacken!

And one more thing, regarding to amalgamation of expected frequencies, you only do them when you want to show that certain distribution is suitable for something or you also do it when you are showing that there is association between 2 variables?

Is there anything else to do apart from edexcel/solomon/IAL

Got it now, thank you.

I'm unsure as to what you're asking here. You pool classes together when the expected frequency is less than 5. You seem to be talking about contingency tables and goodness of fit? I'm confused.

Yes I am talking about contingency tables and goodness of fit. I am asking if you ONLY pool classes together when you are doing goodness of fit or does it also apply to contingency tables? (Sorry my question was really unclear!)

Oh, uhm. You should pool classes together for contingency tables too if the need arises but I've never seen that happen before.

Yes, I was thinking about it, actually I have never seen it in a question. So I should pool classes together whenever the expected frequency of contingency values fall below 5.

Yep, but I doubt the case will arise.

If you are calculating spearmans rank and two units have the same rank, you take an average of that rank and the next and then use PMCC. How would you use PMCC if the data is ranked/ordinal?