Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
388 views

S2 Revision Notes

S2 Revision Notes

Uploaded by

imnul
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
388 views

S2 Revision Notes

S2 Revision Notes

Uploaded by

imnul
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Statistics 2 Revision Guide

Binomial and Poisson Distribution


Distribution
Binomial

P (X = r)

n r
p (1 p)nr
r

Mean ()

Variance ( 2 )

np

np(1 p) = npq

Poisson

1.1

r!

Approx.

Given that...

X B(n, p)

Y N (np, npq)

np, nq > 5

Yes

X B(n, p)

Y P o(np)

n large, p small

No

X P o()

Y N (, )

> 10

Yes

Continuous Random Variables

For p.d.f. f (x)


Z
f (x) dx = 1

Z
=
xf (x) dx
Z
2 =
x2 f (x) dx 2

Conditions for Binomial Distribution

2. Each trial is either a success or a failure (two outcomes)


3. Trials are independent
4. Probability of success, p, at each trial is constant

For c.d.f. F (x)

0 F (X) 1
Z x0
F (x0 ) = P (X x0 ) =
f (x) dx

Conditions for Poisson Distribution

Events occur...

continuity correction?

The probability density function of a continuous random variable provides a model of a


relative frequency histogram3 and so to calculate probablilities for a continuous random
variable you need to find an area under the probability density function! If X is a
continuous random variable:

1. Must be fixed number or trials, n

1.2

1
2

Distribution

1. Gradient of F (x) is never negative since

1. Singly in continuous space or time

2. Mode M satisfies f 0 (M ) = 0

2. Independently of each other

3. Median m satisfies F (m) = 0.5

3. At a constant rate in the sense that the mean number of occurrences in the
interval is proportional to the length of the interval

4. Quartile Q1 satisfies F (Q1 ) = 0.25

d
F (x) = f (x) 0 x
dx

5. Quartile Q3 satisfies F (Q3 ) = 0.75

1.3

Useful Calculations for Probabilities

For the continuous random variables X and Y , where c Z,


1. P (X c) = 1 P (X c 1)
2. Given that p > 0.5,
P (X c|X B(n, p)) = 1 P (Y n c 1|Y B(n, 1 p))

1.4

Continuous Distributions

A random variable X, having a continuous uniform distribution4 over the interval (, )


gives:
Distribution

Approximations

Remember, the binomial and Poisson distributions are used with discrete random variables and the normal distribution with continuous random variables. Hence we are
approximating a discrete random variable by a continuous one in a normal approximation and an allowance must be made by using the 12 continuity correction1 . It is
usually the case that np 10 when using a Poisson approximation.2

know, when P (X = 52) ' P 51 12 < Y < 52 12 .
2 This is so that the Poisson table can be used.
1 You

f (x) =

0
3 Remember,

<x<
otherwise.

Mean ()

Variance ( 2 )

+
2

1
( )2
12

in S1?
P (X = r) = 0 because when finding probabilities using continuous random
variables it is not possible to find exact probabilities.
4 Remember,

Sampling

1. Sample - selection of individual members/items from a population/sampling


frame; Generally cheaper than census-taking
2. Random sample - of size n - where every possible member has an equal chance
of being chosen
3. Population - collection of individual items
4. Finite population - each individual member can be given a number
5. Infinite population - impossible to number each member
6. Countably infinite population - one which is infinite in size, but each member
can be given an individual number.
7. Sampling unit - individual member of a population
8. Sampling frame - list of all sampling units used to represent a population
9. Sampling distribution - gives all the values of a statistic and the probabilty that
each would happen by chance alone.
10. Census - Information obtained from all members of the population - used when
population is known and easily accessed.
11. Simple random sample - of size n - taken with replacement; every member
in the population has an equal chance of being selected.5 the observations are
independent random variables and have the same distribution as population
12. Unrestricted random sample - taken without replacement; every member does
not have an equal change of being selected.
13. Statistic - quantityPcalculated solely from the observations in a sample, e.g. the
sample mean X = nXi . It can not involve any unknown parameters 6
14. (Population) parameter - A numerical property of a sample7
15. Sample survey - An investigation using a sample
Advantages of sampling

Disadvantages of sampling

Cheaper than taking census

Cost per sampling unit studied is


greater than for census

Data more readily available for


analysis

Uncertainty associated with sampling,


which takes the form of. . .

Testing every item to destruction


not needed

Natural variation among sampling units

Quality of information per sampling unit better than for census

Bias, via sampling from an incomplete sampling frame, personal subjective


choice, non-response, or substituting convenient sampling units

5 Usually

used with small populations, or with infinite populations - the probability change
from sampling w/out replacement is so small, it can be ignored).
6 This

P
(X)2

is why the function


is not a statistic since the parameter is usually
n
unknown.
7 Greek letters are usually used for these.

Hypothesis Tests

Used to test how well a mathematical model works.


1. Hypothesis - statement made about the value of a population parameter that we
wish to test by collecting evidence in the form of a sample
2. In a statistical hypothesis test the evidence comes from a sample which is summarised in the form of a test statistic
3. Hypothesis test - mathematical procedure to examine a value of a population
parameter proposed by the null hypothesis H0 compared with an alternative hypothesis H1 .
4. Null hypothesis (denoted by H0 ) - hypothesis that is assumed true unless proved
otherwise
5. Alternative hypothesis (denoted by H1 ) - tells us about the value of the population parameter if our assumption is shown to be wrong
6. Critical region - range of values that would lead you on to rejecting H0
7. Critical values - values on the boundary of the critical region
8. A one-tailed test looks either for an increase or a decrease in a parameter, and
has a single critical value, e.g.
H0 : = m H1 : > m Reject if P (X x)
H0 : = m H1 : < m Reject if P (X x)
9. A two-tailed test looks for both an increase and a decrease in a parameter and
has two critical values e.g.
H0 : = m H1 : 6= m Reject if P (X x) 21 or P (X x) 12
10. After conducting the hypothesis test, relax, have a cup of tea, and address the
following:
Is the result significant or not?
What are the implications in terms of the context of the original problem?
11. Significance level - the threshold probability that we call unlikely.
5% (0.05), 1% (0.01), or 10% (0.1). Denoted by the greek letter

Usually

References
[1] Greg Attwood, Gill Dyer, Gordon Skipworth Statistics 2 (Heinemann Modular
Mathematics for Edexcel AS and A Level 2000 : Heinemann
[2] Keith Pledger, Alan Clegg Edexcel AS and A Level Modular Mathematics - Statistics
2 2006 : Edexcel
c 2013 Sketch.
Copyright

You might also like