Edexcel S3 - Wednesday 25th May AM 2016

So say if in the next part it involved 100 people we would times the probabiltes by 100.
So there is no scenario where we'd need to alter the estimated s.d?

Okay say if we were given a normal distribution (we don't need to calculate anything) at the start of q4 would we then divide by 50 I guess as it involves 50 people?

Indeed. You've already found the s.d for the sample, you'd only divide the s.f by the population number to get the sample variance, but you already have the sample variance so it doesn't need dividing.



If you're given the distribution of the population, yes.

But then the confusing thing is that, isn't those values already taking 50 into account since the data was calculated using those 50 people?

Hi, please could anyone help with exercise C question 5e https://771a1ec81340d97ae9ed29694f73dd633b1c7c70.googledrive.com/host/0B1ZiqBksUHNYV1BTbDkxWXZCVmc/CH3.pdf, how do we know that Var)Y) is the smallest?

Do you see how in the solutions the different variances all have the same form $\textup{something} *\frac{p(1-p)}{n}$ where for example in the second one (for $\displaystyle var(\frac{x_{2}}{2n})$), we have $\textup{something}=\frac{1}{2}$?

All you need to see is: $1>\frac{1}{2}>\frac{3}{8}>\frac{1}{3}$.

So the last one has the smallest.

Anyone else thinks calculating the test statistic in chapter 4 (chi-squared) is too messy and a big hassle? It's giving me a headache and takes too much time!!

Yes, I agree!! If it's any consolation, past papers tend to be slightly less messy/tedious/long than the textbook.

I haven't done any past paper question yet tbh. But I have yet to get one question right out of this chapter, I literally got every question I did wrong because of calculation mistakes.

Okay, I know precisely how you feel - I sat down for a good 3 hours and did all the mixed excercises, checked my answers and they were all wrong, every single one of them. (same thing for another topic but can't quite remember which) - the past papers are fine in comparison, so there's not much to worry about.

In spearmans rank if the observed value is greater than the calculated value then reject H0?

For all hypothesis testing questions in chapter 5 with linear regression and that is H0: p=0 for all questions?

Yeah.

Also (like in June 2015) whenever they ask to test the claim of a person it is always H1: p>0 regardless of the type of correlation right?

Yeah.

Also (like in June 2015) whenever they ask to test the claim of a person it is always H1: p>0 regardless of the type of correlation right?