S2 Edexcel Tuesday 17th January SOLUTIONS IN FIRST POST

We haven't had a definition question since June 2010. There has never been a gap of three papers with no definition papers, so we are due for a nice dosage of definition questions in Jan 2012.

I've memorised the definitions for these ones:

Sampling Unit
Sampling Frame
Sampling Distribution
Census
Sample
Hypothesis Test
Critical Region
Population
Statistic

The most popular seems to be 'statistic' followed closely by 'sampling distribution'. Would they ask us about ones like these, do you think:

One-tailed Test
Two-tailed Test
Discrete Variable
Continuous Variable
Statistical Model

No, I think the words in your first list are much more likely to come up. As you said, words such as 'statistic' are much more common; 'statistic' in particular has appeared in several past papers.

I'll learn them anyway just to be sure...

These are the definitions I'm using, can anyone tell me how to improve them or if any are wrong/unclear?

Sampling Unit - An individual member or element in a population or sampling frame.

Population - A complete collection of all items or individuals.

Statistic - A random variable that is a function of only known observations from a population, consisting of no unknown parameters.

Census - A census is where every member of the population is investigated.

Hypothesis Test - A hypothesis test is a mathematical procedure used to examine the value of a population parameter proposed by the null hypothesis H₀, compared with the alternative hypothesis H₁.

Sampling Frame - A list of all the members or elements in a population.

Sampling Distribution - All possible samples are chosen from the population; the values of a statistic and its associated probabilities form what is known as a sampling distribution.

Critical Region - The range of values of a test statistic that would lead you to reject the null hypothesis (H₀) in a hypothesis test. The boundary (or boundaries) of a critical region is/are called the critical value(s), denoted c₁ and c₂.

One-tailed test - Testing whether or not a parameter is greater (or less) than a given value.

Two-tailed test - Testing whether or not a parameter is different to a given value.

Discrete Variable - A variable that can only have certain values, usually integers.

Continuous Variable - A variable that can take any value, usually in a given range.

Statistical Model - A mathematical description with allocates a probability p to each value of a variable X reflecting the relative likelihood of that value occurring.

S2 Jan 2011 Q1a)

A disease occurs in 3% of a population.

State any assumptions that are required to model the number of people with the disease in a random sample of size n as a binomial distribution.

When doing PDF to CDF or vice versa, how do you know if the boundaries are for example just '<' or '<orequalto'?

I don't think it really matters because it's continuous, but in the mark schemes they appear to usually use the ≤ signs for the part that isn't just saying it is 0 or 1, and for a CDF you say 0 if x < a, and 1 if x > a. That's how I've seen it. Shouldn't really matter that much...

You can also use the {0, otherwise. instead of x &gt; a etc, I've seen that on the mark scheme.

Im stuck on Jan 09 paper with this question:
'It is suggested the PDF f(t) is not a good model for T, Sketch the graph of a more suitable PDF for T&quot;
Given that I have the mode, mean and variance of f(t), how do I go about doing this? I can draw f(t) fine but 'a more suitable PDF' I have no clue.

Is this the one with telephone calls?

Yeah, it's question 4e on January 09.

Dug this old thread up http://www.thestudentroom.co.uk/showthread.php?t=1857041
I still don't really get it. I mean, if they are looking for a graph that's non-negative and has a peak, then what is wrong with the current model? If you roughly sketch f(t) as is then it is both non-negative and has 1 peak.

Edit: Ahhh, I finally get it. Because the model assumes calls over 10 minutes have 0 probability of occuring (which seems unlikely given the mode is 10), they are expecting you to draw a graph with a peak followed by a long tail that gradually reaches 0 to account for longer calls.

Yeah, it's question 4e on January 09.

Dug this old thread up http://www.thestudentroom.co.uk/showthread.php?t=1857041
I still don't really get it. I mean, if they are looking for a graph that's non-negative and has a peak, then what is wrong with the current model? If you roughly sketch f(t) as is then it is both non-negative and has 1 peak.

Edit: Ahhh, I finally get it. Because the model assumes calls over 10 minutes have 0 probability of occuring (which seems unlikely given the mode is 10), they are expecting you to draw a graph with a peak followed by a long tail that gradually reaches 0 to account for longer calls.

X

Ive gota learn all these definitions, apart from that, the module isn't difficult at all.

Exact same as me. It's just so boring!

I quite like S2, very short module... hope the boundaries are not awful like in Jan 2011 though, 69/75 for an A, 72/75 for an A*... ugh.

I quite like S2, very short module... hope the boundaries are not awful like in Jan 2011 though, 69/75 for an A, 72/75 for an A*... ugh.

I saw the grade boundaries and went from to to in about 10 seconds! Why were they so high?!