A-Level

Mathematics

eek noe
MMErlee

~~ = = The Complete Course for Edexcel

Complete AQA and Edexcel modules...

Edexcel

Edexcel

Edexcel i

Edexcel | 4

A-Level
MELE Cece

...from the UK’s most popular

educational publisher
 INANE
ft ketve ww

| earning Kesource Centre

Park Road, Uxbridge, Middlesex UB8 INQ
Renewals: 01895 853344

Please returi this item to the LRC on or before the last

Date stamped below:

UXBRIDGE COLLEGE
LEARNING CENTRE

The Complete Course for Edexcel S1

Introduction — : ;| a
About this book Probability
1. Elementary Probability
The basics of probability 53
Introduction The sample space 54
MENA
2. Venn Diagrams
1. Mathematical Models in Using Venn diagrams 56
Probability and Statistics
3. Laws of Probability
The benefits of mathematical models
The addition law 61
The modelling process
Mutually exclusive events 63
The product law
— conditional probability
Ore CTet Independent events 68
Tree diagrams 71
Data Review Exercise — Chapter 2 77h
1. Representing Data Exam-Style Questions — Chapter 2 78
Data basics
Histograms
Stem and leaf diagrams foliryeyic] em)
2. Location: Mean, Median and Mode
Discrete Random Variables
The mean 14
The mode and the median 16 1. Probability Distributions
Averages of grouped data 19 Discrete random variables 80
Comparing measures of location 22 Probability distributions and functions 80
The cumulative distribution function 86
3. Dispersion and Skewness
Range, interquartile range 2. Expected Values, Mean and Variance
and interpercentile range 23 The expected value 90
Outliers and box plots 28 Variance 93
Variance and standard deviation 31 Expected value and variance of a
Coding oA function of X 96
Skewness 4] The discrete uniform distribution 98
Comparing distributions 45 Review Exercise — Chapter 3 101
Review Exercise — Chapter 1 49 Exam-Style Questions — Chapter 3 102
Exam-Style Questions — Chapter 1 51
eleva. Saya y
Correlation and Regression S1 Statistical Tables 150
1. Correlation Answers 152
Scatter diagrams and correlation 104 ary 200
106 Index 202
Sixx, Syy and Sey

‘Product moment correlation coefficient 108 Boru acc 204

Coded data 4
Limitations of the PMCC 113

2. Linear Regression
Explanatory and response variables 114
Regression lines 115
Interpolation and extrapolation 118
Linear regression with coded data 121
Review Exercise — Chapter 4 123
Exam-Style Questions — Chapter 4 124

eye ay
The Normal Distribution
1. The Normal Distribution
The normal distribution 126

2. The Standard Normal

Distribution, Z
The standard normal distribution 128
Using the tables in reverse 131

3. Normal Distributions and

Z-Tables
Converting to the Z distribution 135
The normal distribution in
real-life situations 139
Finding the mean and
standard deviation of a
normal distribution 143
Review Exercise — Chapter 5 147
Exam-Style Questions — Chapter 5 148
 About this book
In this book you'll find...

3. Laws of Probability Learning Objectives

ae i 5 : .
There are two main probability laws you need to know — the addition law Learning Objectives: S h OWIN 2 wh |ch b |ts of th e
and the product law (or multiplication law). You'll see how to use these laws ® Be able to use the ae Ee
to find probabilities, and how you can adapt them in different circumstances. addition law to find E d exc el S p eci fic at ion are [
eae probabilities. [
The addition law MISE
ie recnance R : I
For two events, A and B, there’s a nice formula linking the union of A and B mutually exclusive cove red in each sect on. I
and the intersection of A and B — the addition law: events. |
* Be able to use the
a oductlawtofind | a aaa
PANIES EO) RB) BAO) probailies i
® Be able to recognise |
You can see why this is true, using Venn diagrams. independent events. |
" Be able to understand |

Tip: If you didn’t f F |

P(A) + P(B) = P(A U B) cs P(A B) ce Clear explanations for every
topic, with lots of helpful tips.
counting ittwice —
So to get P(A U B) on its own, you need to subtract P(A A B) from P(A) + P(B). Spesontandlonee
The addition law is really useful for finding missing probabilities — as long as from B.
you know three of the values in the formula, you can rearrange the formula to eee
find the remaining probability. : e
Tip: You might also see |
the addition law called i
ee E the ‘sum law’. It'll be on
the exam formula sheet,
For two events A and B, P(A U B) = 0.75, P(A) = 0.45 and P(B’) = 0.4. but you should learn it
a) Find P(A 1 B). anyway.
* To use the formula, you need to know P(A), P(B) and P(A U B). ee eee
You're missing P(B), so start by finding that.
P(B) = 1—P(B) =1-0.4=0.6. E |
* Now rearrange the addition law formula to make P(A 9 B) the subject. xam p es
P(A UB) = P(A) + P(B)— P(A 9 B) = P(A MB) = P(A) + P(B) — P(A UB)
* And substituting in the probabilities gives:
PAM B) = 0.45 + 0.6-0.75 = 03 Pl enty of step-by-step :
b) Find P(A’ 9B). CO Tip: Watch out for when worked examp |ese I
= A’ B’ is the complement of A U B. rou anise Mevormula
= So P(A’ 1B’) =1—P(A UB) =1 —-0.75 = (0:25)
A MB! AUB
c) Find P(A m B’).
* Event A is made up of areas AB and A 9 B’ — see Tip on the right.
* So P(A NB’) = P(A) - P(A NB) = 0.45 - 0.3 = 0.15
d) Find P(A’ U B).
* It's easiest to do this by drawing a Venn diagram.
* Use the probabilities you've worked out above. 0.25)
= So P(A’ U B)= 0.3 + 0.25 + 0.

Chapter 2 Probability

In the next example, you're asked to find the probability that Z takes a value
between two limits. You can do this by subtracting one area from another.

Find the probability that 0.12 < Z< 0.82.

= Shade the area you want to find. ——————» i
* To find this area, you need to find the area to
the left of z= 0.82, then subtract the area to
the left of z= 0.12. That will leave you with
the area between the two values.
= So, P(0.12 <Z< 0.82) = P(Z< 0.82) — P(Z< 0.12)
7939 — 0.5478

Exercises (with worked answers)

Exercise
Lots of practice for every topic, with fully Use the table of the normal cumulative distribution function, ®(2), to
worked answers at the back of the book. find the following probabilities.
a) P(Z<1.87) b) P(Z<0,39) c) P(Z<0.99) d) P(Z<3.15)

Use the table of the normal cumulative distribution function, ®(z), to

find the following probabilities,
a) P(Z>2.48) b) P(Z>0.85) c) P(Z>1.23) d) P(Z>0.14)

By using sketches and the table of the normal cumulative distribution

function, &(@), find the following probabilities.
a) P(Z>-3.35) b) P(Z>-0.24) c) P(Z>-1.21)
d) P(Z<-0.62) e) P(Z<-1.14) f) P(Z<-2.06)
Question Hints Work out the following probabilities,
Occasional
% fe
hints to QaHine Forparse) a) P(1.34 <Z<2.18) b) P(0.76<Z<1.92)
: ee
and h), use symmetry
| ©) PU1.45 <Z<0.17) d) P(-2.14<Z<1.65)
point you In the right conieioel ear ' €) PH-1.66 <Z< 1.66) f) PC0.34<Z<0.34)
direction. escent g) P(-3.25 <Z<-2.48) h) P11 <Z<-0.17)

Chapter 5 The Normal Distribution

Review Exercise — Chapter 1 Review Exercises
QI The stem and leaf diagram on the right represents the lengths 12
Mixed questions covering the whole
(in cm) of 15 bananas. Write down the original data as a list. 13
chapter, with fully worked answers.
15
Key 12|8 means 12.8cm| 16 DB ML a ea RT
17 C=NUNne@
Ros
OwH

Q2 Twenty phone calls were made by a householder one evening.

The lengths of the calls (in minutes to the nearest minute) are recorded below.
Draw a histogram of the data.
Length of calls Q-2 | 3-5
Number of calls | 10

Calculate the mean, median and mode

of the data in the table on the right.

The speeds of 60 cars travelling in a 40 mph speed limit area were measured to the nearest mph.
The data is summarised in this table.
Calculate estimates of the mean and median, pecan pul
and state the modal class. Frequenc

Find the mean and standard deviation of the following numbers: 11, 12, 14, 17, 21, 23, 27.

The scores in an IQ test for 50 people are recorded in the table below.
Score [100
-106 | 107-113
|114-120 | 121
-127 | 128
-134
Frequency 6 11 I 22 9 2
Estimate the mean and variance of the distribution,

Exam-Style Questions — Chapter 1

Fora set of data, n = 100, (x — 20) = 125, and (x — 20) = 221.
Find the mean and standard deviation of x, 1 The profits of 100 businesses are given in the table.
Number ofbusinesses
24
26
Chapter 1 Data 21

a) Represent the data in a histogram. (3 marks)

b) Comment on the distribution ofthe profits of the businesses, (2 marks)

Exam-Style Questions A group of 19 people played a game. The scores, x, that the people achieved are
summarised by:
D(x — 30) = 228 and (x — 307 = 3040

Questions in the same style as the a) Calculate the mean and the standard deviation of the 19 scores. (3 marks)

ones you'll get in the exam, with b) Show that }).x = 798 and >)x’ = 33820. (3 marks)

worked solutions and mark schemes.

c) Another student played the game. Her score was 32.
Find the new mean and standard deviation of all 20 scores. (4 marks)

In a supermarket two types of chocolate drops were compared.

The masses, a grams, of 20 chocolate drops of brand A are summarised by:
Da = 60.3 Da
= 219
The mean mass of 30 chocolate drops of brand B was 2.95 g,
Formula Sheet and Statistical Tables and the standard deviation was | g.
a) Find the mean mass ofa brand A chocolate drop. (1 mark)

Containing all the formulas and statistical b) Find the standard deviation ofthe masses ofthe brand A chocolate drops. (3 marks)

tables you'll be given in the S1 exam. c) Compare brands A and B. (2 marks)

d) Find the standard deviation of the masses ofall 50 chocolate drops. (4 marks)

Chapter 1 Data

Glossary
All the definitions you need to know for
the exam, plus other useful words.
A-Level
“Mathematics
for Edexcel

Practice Exam Papers (on CD-ROM)

Two printable exam papers, with fully A eas
ag
worked answers and mark schemes.
Published by CGP
Editors:
Ali Palin, Andy Park, Charlotte Whiteley.
Contributors:
Katharine Brown, George Davies, Alastair Duncombe, Anna Gainey,
Claire Jackson, James Nicholson, Gary Rowlands.

ISBN: 978 1 84762 807 7

With thanks to Mona Allen and Glenn Rogers for the proofreading.
With thanks to Alastair Duncombe for the reviewing.

Groovy website: www.cgpbooks.co.uk

Printed by Elanders Ltd, Newcastle upon Tyne.

Jolly bits of clipart from CorelDRAW®

_Photocopying — it's dull, grey and sometimes a bit naughty. Luckily, it's dead cheap, easy and
quick to order more copies of this book from CGP — just call us on 0870 750 1242. Phew!
Text, design, layout and original illustrations © Coordination Group Publications Ltd. (CGP) 2012
All rights reserved.
 Introduction to Mathematical NOR ente

1. Mathematical Models in
Probability and Statistics
Mathematical models are basically descriptions of things in the real world, Learning Objectives:
and are usually written to help solve some kind of problem. They’re usually Be able to explain
much simpler than the real-world situations they describe, but are still useful. the mathematical
modelling process
The benefits of mathematical models Be able to describe
some of the reasons
Understanding real-world problems for carrying out
= A technology company is going to launch a new product mathematical
— the USB-Penguin. modelling
The company’s managers would like to maximise profits from
USB-Penguin sales. (At the very least, they need to sell enough
USB-Penguins to recover the money spent on development.)
= They are trying to decide what would be the best price to charge
for the USB-Penguin. They realise that the variables ‘price’, ‘sales’,
and ‘profit’ are linked.
For example:
= If the price is low, then sales should be higher.
(i) However, if sales aren’t as high as expected, the low price may mean
that the company will not be able to recover the development costs.
(ii) But if the low price encourages lots of people to buy the
USB-Penguin, then profits could be high.
= Ifthe price is high, then they can expect to sell fewer USB-Penguins.
(i) However, if sales turn out to be higher than expected, then the
company will make a very big profit.
(ii) But if the high price puts people off buying, then the company won't
recover the development costs.
= The managers need to answer the following question:
What is the best price to charge for the USB-Penguin to maximise profits
(or to have the best chance of recovering the development costs)?

Saving money and reducing risk

They decide to try and answer the question using a mathematical model.
= This is a mathematical description of the situation using equations.
= Because real-life situations can be so complicated, some assumptions and
simplifications are usually made — basically, to make the maths easier.
= The model allows possible prices to be tried out (in the model) and the
sales and profits to be predicted.
there
a Using a model to test what might happen in various scenarios means
is less risk of something going expensively wrong in real life (such as the
‘company losing so much money it goes out of business).
Introduction to Mathematical Modelling
 The modelling process
There are seven important steps in creating a mathematical model.

(1) Recognising a real-world problem

The managers’ pricing dilemma, for example.

(2) Making the mathematical model

= This will involve writing down one or more equations.
= For example, there might be an equation describing how the
Tip: Linear regression is retail price (7) affects sales (s). There might also be another
ts ee Pratheaatical equation describing how the price and sales affect profit (P).
modelling — see = Writing these equations might involve making a number of
Chapter 4. assumptions and simplifications.
— For example, the managers might have to make their
‘best guess’ at the effect on sales of reducing the price.
— And they might not take into account factors such as how
increased sales might lead to more competition (as other
firms make products similar to the USB-Penguin).

(3) Use the model to make some predictions

The equations can then be used to predict how
many USB-Penguin sales are expected at various
prices, and the effect this would have on profits.

(4) Collect some real-world data

When the USB-Penguin goes on sale, the company will
be able to collect some real-life data — by recording the
actual sales of the USB-Penguin at the price they choose. Saeva

(5) Compare the predictions with the real-world data

The predicted number of sales and the actual number of sales
can then be compared.

Tipetvouswon’t be (6) Test whether the model ‘works’

expected to carry out ‘Statistical tests can be carried out to test how well
any statistical tests in $1. the model describes the real-world situation.

@) Improve the model

Tip: When you update If necessary, the model can be improved by changing
the model, you need to some of the equations (or adding some new ones) so that
go through steps 3 to 6 it makes predictions that are closer to what is observed in
again. real life. This might allow the company to adapt its prices
in the future so that sales and profits improve.

Introduction to Mathematical Modelling

oe
 Chapter 1 Data

1. Representing Data
SEES GSS SUS

Data is to statistics what fuel is to a car — without data, all the statistics Learning Objectives:
knowledge in the world won’t be much use. This chapter covers the essentials
" Be able to recognise
about data — from graphs, through to location, dispersion and skewness. different types of
variables.
Data basics " Be able to interpret :
A lot of the subject of statistics involves analysing data. frequency tables and
grouped frequency
Data consists of a number of observations (or measurements). tables.
Each of these observations records the value of a particular variable. Be able to draw and
There are different kinds of variables. interpret histograms,
= Variables that take non-numerical values (i.e. they’re not numbers) and stem and leaf
diagrams. '
— these are called qualitative variables. espeene gn sr orem gr ;

# Variables that take numerical values (i.e. they’re numbers)

— these are called quantitative variables.
There are then two different types of quantitative variables.
= A discrete variable can only take certain values within a particular
range (e.g. shoe sizes) — this means there are ‘gaps’ between possible Tip: Or you can think
values (you can’t take size 9.664 shoes, for example). of a discrete variable
= A continuous variable can take any value within a particular range uausiues in steps’.
SE
(e.g. lengths or masses) — there are no gaps between possible values.

elem a es
An employer collects information about the computers in his office.
He gathers observations of the 5 variables shown in this table.
1. Manufacturer Bell _|Banana | Deucer | Deucer
2. Processor speed (in GHz) 2.6 Dive) oeies
3. Year of purchase 2009 | 2010 | 2011
4. Memory (in MB) . 2 3 Jel
5. Colour Grey Grey
Which of the five variables are:
a) qualitative?
® The variables ‘Manufacturer’ and ‘Colour’ take values that are not
numbers.
= So there are two qualitative variables: ‘Manufacturer’ and ‘Colour’ .
b) quantitative?
= The variables ‘Processor speed’, ‘Year of purchase’ and ‘Memory’
take values that are numbers.
= So there are three quantitative variables:
, ‘Processor speed’, ‘Year of as * and ‘Memory’ .
Chapter1 Data =
 The variables below are all quantitative.
(i) length, (ii) weight, (iii) number of brothers, (iv) time,
(v) total value of 6 coins from down the back of my sofa
Tip: ‘Number of a) Which of these 5 quantitative variables are continuous?
brothers’ can only take = ‘Length’, ‘weight’ and ‘time’ can all take any value in a range.
whole-number values.
= So the continuous variables are: ‘length’, ‘weight’ and ‘time’.
‘Total value of 6 coins’
can only take particular b) Which of these 5 quantitative variables are discrete?
values. For example, = ‘Number of brothers’ and ‘total value of 6 coins’ can
they could be worth 12p only take certain values.
or eanbut not 12.
iei
= So there are two discrete variables — these are: ‘number of brothers’
and ‘total value of 6 coins’ .

Data is often shown in the form of a table.

There are two types you need to be really familiar with.
Tip: Frequency just = Frequency tables show the number of observations of various values.
means ‘the number
of times something For example, this frequency Number of bananas 18 |9 [10/11]12,
table shows the number of
bananas in thirty 1.5 kg bags. Frequency |3|7|10]6|4|
happens’.

Tip: Frequency tables = Grouped frequency tables show the number of observations whose
and grouped frequency values fall within certain classes (i.e. ranges or groups of values).
tables can also be They're often used when there is a large range of possible values.
drawn ‘vertically’, like For example, this grouped frequency table shows
this:
the number of potatoes in thirty 25 kg sacks.
Number | Frequency
of Muriberof
pote 50 55 260| 61-65 | 66-70| 71-75
bananas
Freguenc 12
= Notice how grouped See tables don’t tell you the exact value
of the observations — just the most and the least they could be. ,
» And notice how the different classes don’t overlap. In fact, there are
‘gaps’ between the classes because this is discrete data.

Grouped frequency tables are also used for continuous data. Since there are
no ‘gaps’ between possible data values for continuous variables, there can be
no gaps between classes in their grouped frequency tables either.
For example, this grouped frequency table
shows the masses of 50 potatoes.
= Inequalities have been used to define
Mass of
Tip: You don’t always the class boundaries (the upper and
potato (m, in g)

a
need to leave the top lower limits of each class). There are no
and bottom classes ‘gaps’ and no overlaps between classes.
without a lower and
= The smallest class doesn’t have a lower 100 nara |
< m < 200
upper limit — e.g. if you
know for a fact that very
small or very large data
limit — so very small potatoes can
still be put into one of the classes.
[200sm<300[16 |
isse «00a a
values are impossible. Similarly, the largest class doesn’t have
an upper limit.

Chapter 1 Data
This grouped frequency table shows the
lengths (to the nearest cm) of the same
Length of Frequency
potato (/, in cm)
50 potatoes.
= The shortest potato that could go in the aoe 3 Tip: Even though a
6-7 class would actually have a length of 6-7 u potato of length 7.5 cm
5.5 cm (since 5.5 cm would be rounded up 8-9 15 would go in the 8-9
to 6 cm when measuring to the nearest cm). 10-11 class, this is still the
| 16
upper class boundary of
So the lower class boundary of the 12-13 | 3 the 6-7 class.
6-7-Class-is-5.5-cm.
= The upper class boundary of the 6-7 class
is the same as the lower class boundary of Tip: The upper class
boundary of the 12-13
the 8-9 class — this is 7.5 cm. This means
class will be 13.5 cm.
there are never any gaps between classes.
= For each class, you can find the class width using this formula:
class width = upper class boundary — lower class boundary Tip: A class with a
lower class boundary
of 50 g and upper class
= And you can find the mid-point of a class using this formula:
boundary of 250 g can
lower class boundary + upper class boundary be written in different
mid-point = 5 ways.
So you might see:
e ‘100 — 200, to the
nearest 100 g’
Example 1 or eee
e ‘50 < mass < 250’

A researcher measures the length (to the nearest 10 cm) of 40 cars. e ‘50’, followed by
Her results are shown in the table. reneinn on ‘250— for the next
Frequency class, and so on.
Add four columns to the table to show: &
(i) the lower class boundaries 250-350 They all mean the same.
360-410
(ii) the upper class boundaries
(iii) the class widths 420-450
(iv) the class mid-points 460-500

=» The shortest car that measures 250 cm (to the nearest 10 cm) is 245 cm
long. So the lower class boundary of the 250-350 class is 245 cm.
= The upper class boundary of the 250-350 class is 355 cm (even though Tip: 355 cm must be the
a car measuring 355 cm would actually go into the 360-410 class). upper class boundary
for the 250-350 class,
= Once you have the class boundaries, use the above formulas to find the because no number less
class widths and the mid-points. than this would work.
Length |Frequency| lower class | upper class | class width |mid-point For example, the upper
(cm) boundary (cm)|boundary (cm) (cm) (cm) class boundary can’t
250-350 5 245 355 110 300 be 354.99 cm, because
360-410] 11 355 415 60 385 then a car of length
435 354.999 cm wouldn't fit
ACO ee 415 455 40 into any of the classes.
ACOs OOMe oem 455 505 50 480 And the upper class
boundary can’t be
354.999 cm, because
then a car of length
354.9999 cm wouldn’t
fit into any of the
classes. And so on.

Chapter 1 Data
 A researcher measures the (nose-to-tip-of-tail) length of 40 cats.
His results are shown in the table. Length, /(mm)_| Frequency
Add two columns to the table to show:
: 3
(i) the lower class boundaries 250 = = a 7.
(ii) the upper class boundaries 350 s/<
410 <1< 450 iv
450 <1<500 5

= The classes are written using inequalities this time.

That makes it easy to tell where the class boundaries are.
Tip: Even though a = The shortest cat that can go in the 250 </ < 350 class measures
cat of length 350 mm 250 mm. The length of a cat any shorter than 250 mm wouldn't satisfy
wouldn’t go into the the inequality.
250 <1< 350 class, the
upper class boundary of = Similarly, the upper class boundary of the 250 </< 350 class is |
this class is still 350 mm 350 mm.
(since no number
smaller than 350 mm is Length, 7 (mm) Frequency lower class upper class
suitable — see page 5 boundary (mm)
for more information).
350 <1< 410
7
450 <1< 500 450 500

eee
Q1 Amechanic collects the following information about cars he services:
Make, Mileage, Colour, Number of doors, Cost of service
Write down all the variables from this list that are:
a) qualitative
b) quantitative
»

Q2 Amy is an athletics coach. She records the following information

about each of the athletes she trains:
Number of medals won last season, Height, Mass, Shoe size
Write down all the variables from this list that are examples of:
a) discrete quantitative variables
b) continuous quantitative variables

Q3 The heights of the members of a

history society are shown in the table. Height, h (cm) Number of
a) Explain why ‘height’ is a
continuous variable. 140:<h
<150
b) For each class, write down the: 150<h< 160
(i) lower class boundary 160s
2 = 170
(ii) upper class boundary 1703) 1860
(ili) class width 180 <h< 190
(iv) class mid-point 190 <h< 200

Chapter 1 Data
Histograms
Histograms look like bar charts. However, because they're used to show Tip: The formula for
frequencies of continuous variables, there are no gaps between the bars. frequency density (f.d.)
can actually be written:
To plot a histogram, you plot the frequency density rather than the frequency
(as you would in a bar chart). Use this formula to find frequency density: _ 1. frequency
oe k ~* Class width’
frequency where k can be any
Frequency density Eee
= wiih number.

However, it usually
= Here’s some data showing the heights of 24 people. makes sense to use k = 1
when drawing your own
|Height (cm) lower class | upper class | class | Frequency | Frequency histograms.
boundary | boundary | width density
(cm) (cm ) You'll see on the next
130<h<150| 130 DO 11h Cees 0.15 page how to interpret
150<h<160| 150 roo ars histograms where
different values for k
160<h<165] 160 5 5
have been used.
.
170 (20
Tip: If you just plotted
= Here’s the same data
the frequency (rather
plotted as a histogram. than the frequency
1.0
density), then your graph
would look like this:

density
Frequency
noon
© Oo+nNme
oO
130 140 150 160 170 180 190 — = -
130 140 150 160 170 180 190
Height (cm) Height(c

Notice how: It looks like there are

lots of tall people but
(i) The vertical axis shows frequency density. this is just an illusion
(ii) The horizontal axis has a continuous scale like an ordinary graph, created by the width
and there are no gaps between the columns. of the final class. If
this data was split into
(iii) A bar’s left-hand edge corresponds to the lower class boundary. classes all the same
A bar's right-hand edge corresponds to the upper class boundary. width, then the graph
would look a lot more
like the histogram on
eT om the left.
Draw ahistogram to represent the data below, showing the
masses of parcels (given to the nearest 100 g).
Mass of parcel (to nearest 100 g) | 100 - 200 | 300 - 400 | 500 - 700 | 800- 100)
Number of parcels 100 250 600 50

= First draw a table showing the upper and lower class boundaries,
plus the frequency density:
Mass of | Lower class | Upper class | Class Frequency
parcel boundary aay oa ee density
100 - 200 50 250 | 200 100 0.5
300 - 400 250 |. 450 ~—'|~-200 250 ies
~ [500-700 | 450 - 750 a 600 2
|800-1100| 750 =o 1150 7 400 50 0.125
Chapter 1 Data
 = Now you can draw the histogram.

Density
Frequency

- — tt ee
0 200 250 400 450 600
800
Mass of parcel (g)

Tip: In a bar chart, it’s On a histogram, the frequency ina class is proportional to the area of its bar.
the height of a bar that In other words, frequency = k x area of bar (where k is a number).
represents the frequency.
Example 2
This histogram shows
the heights of a group
of people.
There were 6 people between
155 and 160 cm tall.
Tip: You always need
to go through an initial
stage of working out
how area and frequency
are related — basically, Height (cm)
you're working out the
value of k that’s been a) How many people in the group are between 130 cm and 155 cm tall?
used. (See the ‘Tip’ Before looking at the 130-155 class, you need to use
about the frequency the information you have about the 155-160 class.
density formula on the
previous page.) # Work out the area of the bar for Width of bar = 160-155 =5
155 cm to 160 cm to find how Height of bar = 0.6 :
area and frequency are related. So area. of bar=5 x0.6=3
Tip: Or you can say that = An area of 3 represents a frequency of 6 — so ‘frequency = 2 x area’.
1 square unit represents
2 people. * Now you need to find the area of | Width of bar = 155 — 130 =25
the bar from 130 cm to 155 cm. Height of bar = 0.2
So area of bar=25x0.2=5

= Remember... So 10 people are between

frequency = 2 x area. 130 cm and 155 cm tall.

b) How many people in the group are over 165 cm tall?

« There are two bars representing 165-170 cm:
people over 165 cm tall. Area ot bar = 50.6 =4
You need to find the This represents 8 people. |
frequencies represented 170-190 cm:
by both of them. Area Gf Dal = 20>q0:4 = 6
This represents 16 people.
= Add the individual frequencies So 8+ 16 = 24 people
to find the total frequency. are over 165 cm tall.

Chapter 1 Data
Example 3 eee
The histogram below shows the speeds of cars along astretch of road.
There were 26 cars travelling between 50 and 60 mph.
10
©

density
Frequency

Tip: You can only

estimate the number of
cars travelling at 25 mph
or less because you
OR omen Oslo 20. 25 30s 30° 40) 45.50, 55.60 don’t know exactly what
Speed (miles per hour)
the speeds in the group
Estimate the number of cars travelling at 25 mph or less. 20-30 pr Were:
=* Work out the area of the bar from Width of bar = 60 —50 = 10
50 mph to 60 mph to find how Height of bar = 2.6
area and frequency are related. So area of bar = 10 x 2.6 = 26 Tip: Or you can say that
1 square unit represents
« An area of 26 represents a frequency of 26 — so ‘frequency = area’. Vperson,

You now need to find the area of the bars to the left of 25 mph
on the horizontal axis (because you need the
frequency of cars travelling at 25 mph or less).
« First, find the area of Width of bar = 20 —0 = 20
the 0-20 mph bar. Height of bar = 0.4
So area Onpah=20pe0.4.—= 8
This represents 8 cars. Tip: Finding half of the
area under the 20-30 bar
= Now find half the area of ‘ Width of bar = 30—20= 10 kes aasecmpucn
the 20-30 mph bar — this Height of bar = 1.8 eaviaate nomen
represents speeds from Total area of bar = 10 x 1.8= 18 travelling at 20-30 mph
20 mph up to 25 mph. So half the bar’s area = 18 = 2 =9 were going 25 mph or
This represents 9 cars. less while the other half
were travelling faster
, than 25 mph.
= Add these figures together Total travelling at 25 mph or less
to estimate the total number —~3+9=- 8
travelling at 25 mph or less.

Chapter 1 Data
 Exercise 1.2

Ql A pet shop records the mass (m, in g) Mass (m, in g)

of their adult hamsters.
|100m
<110)| ea
The results are shown in the table.
Draw a histogram to show the
Ria ae
masses of the hamsters.

Q2 The histogram below shows the audition times (in seconds) for
contestants applying for a place on a television talent show.
The auditions for 54 contestants lasted between 30 and 45 seconds.

ioe)On
Qo
N Oo
on

NO °

ee w
density
Frequency
1.0

OS

(0) 1) AY 380 40 SO 6 70 80 90
Time (seconds)

a) Work out the number of contestants whose auditions lasted less

than 30 seconds.
b) Work out the total number of contestants who auditioned.

Q3 A butterfly enthusiast measures the wingspans (w, in mm), to the

nearest millimetre, of a sample of tortoiseshell butterflies. She groups
her measurements and displays the data in a histogram.
The group containing butterflies with 44.5 <w < 47.5 has a
frequency of 12. This group is represented on the histogram
by a bar of width 1.5 cm and height 9 cm.
a) Show that each butterfly is represented on the histogram by an
area of 1.125 cm’.
b) The bar representing the butterflies with 51.5 < w < 53.5
has an area of 22.5 cm?. Work out the frequency for this group.
c) The frequency for butterflies with 53.5 < w < 58.5 is 14.
Work out the width and the height of the bar used
to represent this group.

Chapter 1 Data
Stem and leaf diagrams
Tip: The stem represents
Stem and leaf diagrams are another way to represent data. a much bigger part of
In fact, the data values themselves appear in a stem and leaf diagram. the value than the leaf.
= Each data value is split into a ‘stem’ and a ‘leaf’. i ea OAs
= A complete stem and leaf diagram looks something like this: stem might be 2 (which
represents 20), while the
These are the stems. => 1| 1 3 7 +—sJlhese are the lea leaf might be 8 (which
we 23 | Ae oo 758 a oh represents 8).
Bo | COON - Or with the data value
a) | ioe Key: 2 | 8 means 28 3.7, the stem might be
3 (representing 3), while
5 the leaf might be 7
= A stem and leaf diagram always needs a key to tell you how to read it. (representing 0.7).
So in the above stem and leaf diagram, the first row represents the values
11, 13 and 17, while the second row represents the values 24, 25, 25, 27
; a Tip: The leaves in each
and 28. You can read the other two rows in a similar way. rowi are usually sorted
into the correct order,
from lowest to highest.
Example 1
The lengths in metres of cars in a car park were measured to the nearest
10 cm. The results were: 2.9, 3.5, 4.0, 2.8, 4.1, 2.7, 3.1, 3.6, 3.8 and 3.7.
Draw a stem and leaf diagram to show the data.
y
» Use the numbers before the decimal point 3
as the stems. So the stems will look like this: ——— 4
# Use the numbers after the decimal point as the leaves. Tip: It’s best to do a
Add them to the correct row of the diagram one at a : : i é oo rough version first, and
time. It’s a good idea to cross out the numbers in the ie then put the leaves in
list (in pencil) as you add them to your diagram. order afterwards.
= Then put the digits after the decimal point in order. —,» 2|7 8 9
;
= And finally, add a key. Key 2|9 means2.9m] Dia els Daler
4|o 4

Example 2
The results of a survey of people’s heights are shown below in the form of
a stem and leaf diagram.
a) How many people’s heights were measured? 14/1 3779 (5) Paes on
: :
Count the leaves — thiswill be the number 15 22.255 (4) ee
Ip:
ee
e numbers
In

of people who had their height recorded. 1610-0 J lea (5) Het aes
This is 22 people . 17}5 6 89 (4) row. These aren’t always
all ener ae (4)
b) What was the:
[Key 15|2 means 152 cm| etme
(i) minimum height recorded?
The minimum height recorded was 141 cm.
(ii) maximum height recorded?
The maximum height recorded was 183 cm.

Chapter 1 Data
 A back-to-back stem and leaf diagram is basically two normal stem and leaf
diagrams drawn either side of the same stems.
For example, this back-to-back stem and leaf diagram shows the ages of
two sets of newspaper readers — for the Daily Rant and the Daily Pry.
The stems are the numbers in the middle, between the two vertical lines.

Daily Rant Daily Pry

QA EDD 3 Ae
828 77 343: 3a 050 2256
(5302 2 led Oa ea
6-6-2 bore eo 9
Key: 1 |2 |3 means 21 for Daily Rant and 23 for Daily Pry
—

Tip: Also, since smaller

leaves are shown nearer The stem and leaf diagram for the Daily Rant is read ‘backwards’ — because
the stem, the leaves for the stems are on the right of the leaves. Make sure you understand the key.
the Daily Rant decrease
in size as you read from
left to right.
The heights of boys and girls in a year 11 class
; : 151-9
are given to the nearest cm in the back-to-back 3/16/15 9
stem and leaf diagram on the right. 981/17/0235
Write out the data in full. 5 2 |18] 0
iets
Key 1|17|0 means 171 for Boys and 170 for Girls
= Remember, the boys’ heights are read backwards.
So the first boy, 8]16|,.has height 168 cm.
* So the heights in cm of the boys are: 168, 171, 178, 179, 182, 185, 191
" The first girl, |15|9, has height 159 cm.
So the heights in cm of the girls are: 159, 161, 165, 169, 170,
V/2) V732175: 4180

Example 2
Construct a back-to-back stem and leaf diagram to represent >
the following data:
Boys’ test marks: 34, 27, 15, 39, 20, 26, 32, 37, 19, 22
Girls’ test marks: 21, 38, 37, 12, 27, 28, 39, 29, 25,5, 24, 31, 36
= Use the tens digits as the stems. Boys G irls
So the stems will look like this: 1
2
3
Tip: The smallest leaf = Add the units digits one Boys Girls
should be the one ata time as the leaves. OS eetwie
nearest the stem. So 2,6 0 74); 25) leet Seo. 5 24
the leaves for the boys
Fs Deets Avanl. SBinal aBtighes Ogle:
are in descending order
(getting smaller from = Then put the leaves Boys Girls
left to right), while the in order, and add a key. Husa) Tine
leaves for the girls are in EF iGyay: beledWi leutine ius Zp
ascending order (getting
bigger from left to right). ee Se ah ee ke te
Key 0|2|1 means 20 for Boys and 21 for Girls

Chapter 1 Data
Exercise 1.3
Q1__ The stem and leaf diagram 1 246 79 (4)
shows the number of members Palle De Je eb Gok 8. (8)
in each of the choirs taking SSS a7 20 (5)
par
ina t
choir competition. 4/145578 (6)
Deine Ome eo (5)
[Key: 1|2=12 members] 6 (0)
Fa Le (2)
a) How many choirs were competing in this competition altogether?
b) How many choirs had 45 members?
c) How many choirs had more than 52 members?
d)) What was the size of the largest choir?

Q2__ The stem and leaf diagram shows the Cole (2)
minimum recorded temperature (to 9147899 (5)
the nearest 0.1 °C) in 24 towns one POG tari a7 OPO)

a
June night. | Hl 0.4 5a (4)

Wo 2 ey # (3)

a) How many towns had a minimum temperature below 10 °C?

b) Which temperature was recorded most frequently?
c) What was the difference between the highest and lowest
temperatures recorded in these towns?

Q3 Sixteen children belong to an orchestra. The distances (in kilometres)

that the children live from the practice hall are shown below.
Se Ta bsa2
a) Draw a stem and leafae to show this information.
b) Use your stem and leaf diagram to find:
(i) the furthest distance that a child lives from the practice hall.
(ii) how many children live less than 2 km from the practice hall.

Q4 Freddie has a collection of films on DVD. He has 24 comedies and

20 dramas. This back-to-back stem and leaf diagram shows the
running times (to the nearest minute) of these films.
Comedies Dramas
(5) Je der 0 en (0)
(Groote 76 (1)
(6) acCdb Ou tate ts tO Foo (2)
(4) Beemer: Gull Went ae7 (4)
(2) Ae je Oued a A a he S787)
(1) Aa Wl We eed 5 ele)
[Key: 1|8| 7 means 81 minutes for comedies and 87 minutes for dramas|
a) What is the running time of Freddie's shortest comedy?
b) How many of his dramas last more than 90 minutes?
c) How many of his comedies last between 70 and 80 minutes?
Chapter 1 Data
 Learning Objectives: 2. Location: Mean,
Be able to calculate
the mean, median and Median and Mode SO

mode for a data set.

are
Be able to estimate The mean, median and mode are measures of location — this means they
values for the mean a summary of where the ‘centre’ of the data lies. Measures of location are
and median for a sometimes called averages. You'll have met the mean before, for sure.
grouped data set, and
to find a modal class. The mean
Be able to recognise The most common measure of location is called the mean (and is often just
when each measure called ‘the average’... though not while you're doing A-level Maths).
of location is most
Suitable: The formula for the mean (X, pronounced ‘x-bar’) is:

Tip: The & (sigma) just

means you add things where each x is a data value, fis the frequency of each x (the number
together — so Ux means of times it occurs), and n is the total number of data values.
you add up all the
values of x.

Find the mean of the following list of data: 2, 3, 6, 2, 5, 9, 3, 8, 7, 2

Tip: n=), f
cu nae ~. ee = First, find 5 v= 2 43672 454 9493 S64 72 47
= Then divide by n = 10 (since there are 10 values).
Tip: Instead of 2x
(which means ‘add up This gives: x = it
=
az
= 10 ~=&
all the values of x’), you
sometimes see Xi.
Here, you have to
imagine there are
A scientist counts the number of
x-values x,, X,, X5...
— then x means ‘add
eggs in some song thrush nests. [Number ofeggs,x|2 |3 |4 |5 |6
up the values of x,, for His data is shown in this table. [Number ofnests, f|4 |9 |16]8 |3
all the different values of Calculate the mean number of eggs in these nests.
i’ (which is just another
* This time you have frequencies. It’s best to add to the table:
way of telling someone
(i) a row showing the values offx, .
to add up all the
different values of x). (ii) a column showing the totals >)f and >) fx.

So }x and }x: mean |Numberofeggs,x|2 |3 |4 |5 |6 [Total]

exactly the same thing. |Number ofnests,f|4 |9 |16|8 |3 | 40
[fe 8 27[64]40)18]157)
:7 Now use the formula for the mean: =x De = 749
= ye ae — 3925 eg8s7

If you have two sets of data and then combine them, you can
find the mean of the combined set of data in the following way.

Tip: Here, x, is the

mean of one data set, a) Find the mean (x,) of the data values 2, 3, 4, 5, 6.
and x, is the mean of
another data set. Themeanof thisdata
is = 4% —2+3+4+5+6 _ 20 _g

= Chapter 1 Data
 b) Find the mean (x,) of the data values 10, 12, 14.
The mean of this data is x, — Dx
: wey lLOsrl 2c 1A.= 36
c) Find the mean (x) of the combined data set 2, 3,4, 5, 6, 10, 12, 14. Tipsevoulcanitinediida
The Se of the combined data set is X, and ¥, together.
=_ 2* _ 24+34+4+4+546+10+1
nonin = tpt tt * =2 And you can’t just find
the mean of xX, and x).

If you know adata set of size n, has mean x, and another data set
of size n, has mean X,, then the combined mean is ¥, where:

Example 4
A data set consisting of 5 values has mean X, = 4.
A second data set consisting of 3 values has mean x, = 12. Tip: This is the same
problem as in
Find the mean (x) of the combined data set.
Example 3, only it’s
= Here, n, = 5 and-n, = 3. been solved using the

» Using the formula: Nae te salon

x = Pagan (Oe ot (3yee
12) aay5G ise 7. belies

Exercise 2.1
Q1_ Katia visits 12 shops and records the price of a loaf of bread.
Her results are shown in the table below.
£1.08 | £1.15 | £1.25 | £1.19 | £1.26
|£1.24 |
£1.15 | £1.09 | £1.16 | £1.20 | £1.05
| £1.10 |
Work out the mean price of a loaf of bread in these shops.

Q2 Twenty students sit a maths exam. The teacher records their marks
{x,,.-+; X94 and calculates that Die 07
Work out the mean mark for these students.

Q3 The numbers of goals scored by 20 football teams in their most

recent match are shown in the table.
Numberof goals |0 1 i
Frequency, f 5 eZ ae oe

Calculate the mean number of goals scored by these teams in their

most recent match.

Q4 Adrama group has 15 members.

The mean age of the members is 47.4 years.
a) Work out the total of the ages of all members of the drama group.
b) A person aged 17 joins the drama group. Find the new mean age.

Chapter 1 Data
 The mode and the median
There are two other important measures of location you need to know about
— the mode and the median.
Tip: The mode is often Mode = most frequently occurring data value. } _
called the modal value.

eT a ————-

Find the modes of the following data sets.

a) 23; 6820589 As oT
The most frequent data value is 2, appearing three times.
So the mode = 2.
b) 4, 3, 6, 4,5, 9, 2, 8, 7,5
Tip: If a data set has two This time there are two modes — the values 4 and 5 both appear twice.
modes, then it is called
bimodal. So the modes = 4 and 5.
©)44, 3).6,.1185, 972 18075 12
This time there are no modes — each value appears just once.

The median is slightly trickier to find than the mode.

Median = value in the middle of the data set when all
the data values are placed in order of size.

First put your n data values in order, then find the position of the median in
the ordered list. There are two possibilities:
i) sur S is a whole number (i.e. 1 is even), then the median is halfway
between the values in this position and the position above.
(ii) if 5 is not a whole number (i.e. 7 is odd), round it up to find the
position of the median. :

Examples
Find the medians of the following data sets.
a). 2;3;6, 276905 87
= Put the values in order first: 2, 2, 3, 3, 6, 6, 7, 8,9
« There are 9 values, and 5 = : = 4.5. Rounding this up to 5 means
that the median is the 5th value in the ordered list — median=6.
b) 4,3, 11,
4,10, 9, 3,8, 7,8
" Put the values in order first: 3, 3, 4, 4, 7, 8, 8,9, 10, 11
" There are 10 values, and 5 = u = 5,
Tip: The value halfway * So the median is halfway between the 5th value in the ordered list fe—7 }
between two numbers is
their mean. and the 6th value (= 8). So the median=7.5.

Chapter 1 Data
If your data is in a frequency table, then the mode and the median are still
easy to find as long as the data isn’t grouped.

The number of letters received

Number of letters| Number of houses
; 0 11 Tip: Frequency means
one day in a sample of houses 1 an 25 the number of times a
is shown in this table. 2 | 27 data value occurs.
3 | 21
| 4 | 9 Here, the data values
Wis 5 ul tie 7 are the ‘numbers of
, letters’ they received.
a) Find the modal number of letters. So the frequencies are
= The modal number of letters just means the mode. the ‘numbers of houses’
(that received that many
« The highest frequency is for 2 letters — so the mode = 2 letters . letters).
b) Find the median number of letters.
* It’s useful to add acolumn — [No, of letters No. of houses] Cumulative
to show the cumulative (frequency)
frequency — this is just 2} 1 _
a running total of the
frequency column.

= The total number of houses is the final cumulative frequency, so n = 100.

= Since 5 = ue = 50, the median is halfway between the 50th and 51st
data values.
« Using the cumulative frequency, you can see that the data values in
positions 37 to 63 all equal 2. This means the data values at positions
50 and 51 are both 2. So the median = 2 letters .

Finding the median and mode from a stem and leaf diagram also shouldn't
cause too many problems.

ren : 4st 3° 78729 (5)

I 1 t
meme sage ott ee fem ae 15/2225 4) Tip: It’s easy to find the
in this stem and leaf diagram. 16100114 (5) ESOT Wee
17/5 689 (4) from a stem and leaf
181112 3 (4) diagram, because the
[Key 15 |2 means 152 cm data values ae in order.
a) Find the modal height.
Look for the number that appears consecutively most often
in one of the rows of leaves. Here, there are three 2’s in the
second row, representing 152 cm. So the mode = 152 cm.
b) Find the median height.
Find the total frequency — here you've been given the number of values
in each row, so you can just add them up to find that n = 22.
Since 54 Sts— 11, the median is halfway between the
11th data value (= 160 cm) and the 12th data value (= 161 cm).
So the median = 160.5 cm.

Chapter 1 Data
 Exercise 2.2 f
a

Seventeen friends took part in a charity fun run.

-
Q1
The amount of money that each friend raised is shown below.
£500 | £123 | £185 HO
G48 | £6" F194 7 sf08) tis
£187 | £216 | £84 | £96 | £14 0
a) Find the median amount of money raised by these friends.
b) Explain why it is not possible to find the mode for this data.

Q2_ A financial adviser records the interest rates charged by 12 different

banks to customers taking out a loan. His findings are below.
6.2% 6.9% 6.9% 8.8% 6.3% 7 A%
6.5% 64% 9.9% 6.2% 6.4% 6.9%
a) Write down the mode of these interest rates.
b) Find the median interest rate charged by these banks.

Q3 Anonline seller has received Number otcustomers

the ratings shown in this table.
a) Write down the modal
customer rating.
b) Work out the median
customer rating.

Q4_ A theatre stages 35 performances of its

pantomime one year. O10 3°5' 5°56" 8°99 clo)
This stem and leaf diagram sli fies ae ate aly cera (7)
shows the number of empty 9] 9 23445 (6)
rai cat Ee loadba 311456889 (7)
a) Write down the value
of the mode. Bliceee ee 5
Salers (1)
b) Work out the median _
number of empty seats. [Key: 3 | 1 = 31 empty seats |

Q5__ Kwasi and Ben each check the download speeds on their computers
on a number of different occasions.
Their results are shown in this back-to-back stem and leaf diagram.
Kwasi Ben
3 5 gate (3)
(6) 7 APD OF 0 a ow eee (5)
(5) G28 4.3.0 |esis oO (4)
(6) OOS 7 61 G1 ee ae
(2) G7 A eee (3)
Key: 3 |5|9 represents a speed of 5.3 Mbit/s for Kwasi
and a speed of 5.9 Mbit/s for Ben.
a) Write down the modal download speed(s) for: (i) Kwasi (ii) Ben
b) Find the median download speed for: i
(i) Kwasi ii) Ben
Chapter 1 Data
Averages of grouped data
If you have grouped data, you can only estimate the mean and median.
This is because the grouping means you no longer have the exact data values.
And instead of a mode, you can only find a modal class.

Modal class
= To find a modal class, you need to find the class with Tip: See p7 for more
the highest frequency density. about frequency density.
= If all the classes are the same width, then this will just
be the class with the highest frequency.

Find the modal class for this data showing the heights of various shrubs.
Height of shrub
to nearest cm
i

# In this example, all the classes are the same width (= 10 cm).
= So the modal class is the class with the highest frequency.
» This means the modal class is 31-40 cm.

Mean
= To find an estimate of the mean, you assume that every reading in a class
takes the value of the class mid-point.
= Then you can use the formula ¥ = x a Tip: This is the formula
from page 14.

Example ppl tele eel sr

The heights of a number of trees were recorded.
The data collected is shown in this table.

Heightoftree |os 610 MTT |s6.2

to nearest
m

Number of trees| 26 17 Tf) | to

Tip: The frequency (/) is
Find an estimate of the mean height of the trees. the ‘Number of trees’.
= It's best to make another table. Include extra rows showing:
(i) the class mid-points (x),
(ii) the values offx, where fis the frequency.
= And add an extra column for the totals >)f and >)fx. Tip: For the first class:
Lower class
Height of a 11-15 | 16-20 Total boundary = 0
to nearest m Upper class
boundary = 5.5
Class mid-point, x| =TS
= Df) So the class mid-point
|Number of trees, f 17
= (065.5) = 2 = 2975
| fc 108 ereSfx)|
(See p4-5 for more
= Now you can use the formula for the mean given above. information.)

Mean = ¥ = : = 498.9 — 7.64 m (to 2 d.p.)

Chapter 1 Data
 Median
Tip: ‘Linear
interpolation’ is To find an estimate for the median, use linear interpolation.
sometimes just called This table shows the ‘tree data’ from Height of tree .
‘interpolation’. the second example on the previous to nearest m
page. I’ve also added a row
showing the cumulative frequency. Curmulthe
Tip: The red crosses in
the diagram aren’t the frequency
actual data values —
= First, find which class the median is in.
they’re interpolated data
values (we've made an 4 == — 30, there are 30 values less than or equal to the median.
Since 5
assumption about them
This means the median must be in the 6-10 class.
being at these points).
= Now, the idea behind linear interpolation is to assume that all the readings
Tip: In this method, you
in this class are evenly spread. So divide the class (whose width is 5) into
17 equally wide intervals (one interval oan ~- 105
only need to find 5 for each of the data values in the Os ele ee
— you don’t then need ee ee
to follow the rules
class), and assume there’s a reading in
the middle of each interval, like this. 26 values 5 (= 30) 43
described on page 16.
However, the two ways = The numbers on the top of the scale are heights (measured in metres).
of using 5 turn out to be The upper and lower class boundaries are shown, and the median (m)
equivalent. is also marked (but you don’t know its value yet).
For example... here, = The numbers underneath the scale are cumulative frequencies.
using the rules on p16, The cumulative frequency at the lower class boundary is 26,
you'd expect the median while the cumulative frequency at the upper class boundary is 43.
to be between the 30th
and 31st values. And if b
<<
you look at the picture, = In fact, you don’t, need to draw the —_—_—>
that’s where it is. small intervals and the data points 5.5 median, m 10.5
The difference is that every time — a simplified version
we're now using - to like the one on the right is enough n
decide how many of to find a median. : 2 6. wa z 39 43
those ‘mini-intervals’ to
count along, whereas on
p16 we were using 4 to
decide the position of ® To find the median, m, you need to solve: | — = 7
a term in a sequence of
numbers.

Tip: This is the Estimate the median Height of tree

same data as in the height for the trees to nearest m
explanation above — recorded in this table.
so you already know
which class contains the
= First draw the picture of the
median.
class containing the median (m) 5.5 a, - 10.5
— just show the important
Tip: Work out the values numbers you're going to need.PE | -
of a,, b,, a,, b, from the b
diagram: « Then solve the equation 4 = @
a,=m—5.5
i by”
b, = 10.5-5.5 » Substituting
uting iinn th the numbers gives:
i _M=so.9% 2) 8026
Tp =— Er = 7396
a, = 30-26
b,= 43— 26 (5,is just 2
the class frequency from And so fil
“===OL sera ww
= 77, 0rm=5 4
X75 +5.5 == 6.68m (to 2d.p.)
the original table).

Chapter 1 Data
 Estimate the median length of the newts described in this table. Tip: You don’t always
need to add a row
Length (to nearest cm) °-
2 Yo
7 to the table to show
Number of newts cumulative frequency.
Here, the numbers are
« There aren =3+18+12 + 4 = 37 data values in total. small enough that you
So 5 = 18.5 — meaning the median will be in the 3-5 class. can quickly tell which
class the median lies in.
= Now draw the picture of the eS
class containing the median. —.2.5 & m 55 You can see that 5
is greater than the
= So you need to solve: 3 a, 18.5 21 cumulative frequency
m=2.5) 2 le5
—3 of the first class, but
Sen LSD °, less than the cumulative
frequency of the second
This gives moe. — 15) Ong ==) x D2 +25 = 5.1cm(to 1d.p.) — so the median is in
the second class.

Q1 Hint: Don’t give

your answers to too
many decimal places
when you're estimating
something.

a) Copy and complete the table.

b) Use the table to work out an estimate of the mean time it took
a
iat
these children to change.

Q2 A postman records the [Number of letters |Number of houses

number of letters delivered 0-2 00 O
to each of 50 houses one day. 3.5 16
The results are shown in 6-8 7
this table. 9-11 =

a) State the modal class. Peet l2eb4 2

b) Estimate the mean number of letters delivered to these houses.
c) Write down the interval containing the median.

Q3__ The table shows the amount of Rainfall (, mm)_| Frequency

rainfall (r, in mm) recorded at Oe
Poa ial 5
hy come ee
some weather stations one March.
Use linear interpolation to | 50<r<60
estimate the median rainfall | 60<r<80 .
at the weather stations. 80 <r< 100 8
F100 <r < 120 2
Chapter 1 Data
 Q4 The table below shows the times that a random sample of 60 runners
took to complete a marathon.
a) Estimate the mean time of these Time (¢,mins) | Frequency,f
runners. (You may use 180 <¢< 240 8
> fe = 16740, where xis 240 < t< 270 19
the mid-point of a class.) 270 <¢< 300 21
b) Calculate an estimate for the Be <t< 360 9
median time it took these runners [360 < t< 480 3
to complete the marathon.

Comparing measures of location

You've seen three different measures of location — the mean, the median and
the mode. Each of them is useful for different kinds of data.
Tip: There’s more about
outliers on page 28. Mean
= The mean’s a good average because you use all your data in working it out.
= But it can be heavily affected by extreme values / outliers and by
Tip: A symmetric data
set is one where the distributions of data values that are not symmetric.
distribution of data ® And it can only be used with quantitative data (i.e. numbers).
values above the mean
is the mirror image of Median
the distribution of values
= The median is not affected by extreme values
below the mean.
— so this is a good average to use when you have outliers.
See page 41 for more # This also makes it a good average to use when then data set is
about symmetry of data not symmetric.
sets.
Mode
Tip: You saw a bimodal = The mode can be used with qualitative (non-numerical) data.
data set (with two =™ But some data sets can have more than one mode (and if every value in a
modes) on p16. data set occurs just once, then the mode isn’t very helpful at all).
Other data sets may
have more than two
modes. yey 2 : -
Q1__ Explain whether the mean, median or mode would be most suitable
as a summary of each of the following data sets.
Q1 Hint: Think about
the shape of the a) Salaries of each employee at a company.
histogram you might b)) Length of adult female adder snakes.
expect for each data set. C ) Make of cars parked in a car park.
For example, in part a),
think about how many d)) Weight of all newborn full-term babies born one year at a hospital.
people you might expect e) ) Distance a firm’s employees travel to work each morning.
to earn quite a low
salary, and how many Q2__ Hosi records the number of bedrooms in the houses lived in by a
you might expect to earn sample of 10 adults. His results are shown in the table.
a very high salary.
|Number of bedrooms Late [eees Tea|Silesian lau
|Frequency | eae reo eon ee
Explain why the mean may not be the most suitable measure of
location for the data.

Chapter 1 Data
3. Dispersion and Skewness
SESS

A measure of location tells you roughly where the centre of the data lies. Learning Objectives:
Dispersion, on the other hand, tells you how spread out the data values are.
And skewness tells you how symmetrically distributed they are. " Be able to calculate
the range, interquartile
range, and
Range, interquartile range interpercentile range.
and interpercentile range Be able to determine
whether a reading is
Range an outlier.

The range is about the simplest measure of dispersion you can imagine. Be able to draw and
interpret box plots.
Range = highest value — lowest value = Be able to calculate
and interpret variance
But the range is heavily affected by extreme values, so it isn’t really the most
and standard
useful way to measure dispersion. deviation (including
with the use of
interquartile range coding). _
A more useful way to measure dispersion,is to use the interquartile range Be able to calculate
— but first you have to find the quartiles. You’ve already seen how the and interpret
median divides a data set into two halves. The quartiles are similar — there coefficients of
are three quartiles altogether (usually labelled Q,, Q, and Q,) and they divide skewness.
the data into four parts.
= Q, is the lower quartile — 25% of the data is less than or
equal to the lower quartile.
= Q, is the median — 50% of the data is less than or equal to the median.
= Q, is the upper quartile — 75% of the data is less than
or equal to the upper quartile.
For example, the values in the data set below have been sorted so that they’re
Tip: There are various
in numerical order, starting with the smallest. The three quartiles are shown.
ways you can find the
quartiles, and they
Paes 444 5[5 5 6 § 7 7.9 Be sometimes give different
Oe results. But if you use
Q, Q, Q,
the methods below,
you'll be fine.
The quartiles are worked out in a similar way to the median
— by first finding their position in the ordered list of data values.
Tip: This is the same
To find the position of the lower quartile (Q,), first work out r as the method used on
page 16 for finding the
= if a is a whole number, then the lower quartile is halfway median, only with na5
between the values in this position and the position above. replaced by 7.
= if a is not a whole number, round it up to find the position of
the lower quartile.
Tip: This is the same
as the method used on
To find the position of the upper quartile (Q,), first work out an
page 16 for finding the
a if = is a whole number, then the upper quartile is halfway median, only with 4
between the values in this position and the position above. replaced by on,

= if“H is not a whole number, round it up to find the position

_ of the upper quartile.
Chapter 1 Data
 the
Once you've found the upper and lower quartiles, you can find
Tip: You don’t need interquartile range (IQR).
to know the median
(Q,)
(Q,) to calculate the
interquartile range.
Interquartile range (IQR) = upper quartile (Q,) — lower quartile

= The interquartile range is a measure of dispersion.

= It actually shows the range of the ‘middle 50%’ of the data.
= This means it’s not affected by extreme values, but it still tells you
something about how spread out the data values are.

a) Find the median and quartiles of the following data set:

>5, 3/11) 678, G87 6; 2a 2379 1a B lore,
= First put the list in order:
1,2, 2, 33) 5; 6, 6; 7, 6, 8) 9) Velo

= You need Q,, Q, and Q,, so find 7 =2 and 37fnwhere n = 18.

18 _ 45. This is nota ha ay so round up to 5.

4° = 4
n

This means the lower quartile is equal to the 5th term: Q, =3

. 5== a = 9 isa whole number. The median is halfway

between the 9th term (= 7) and the 10th term (= 8). So Q, =7.5

. a = 2 = 13.5 is not a whole number, so round up to 14.

This means the upper quartile is equal to the 14th term: Q, = 11
b) Find the interquartile range for the above data.
= Interquartile range = Q,- Q,=11-3=8
/
Tip: You don’t have the
exact data values with
grouped data, so like When your data is grouped, you'll need to use linear interpolation to find
with the median you can an estimate for the lower and upper quartiles.
only estimate the lower
and upper quartiles.

a) Estimate the lower and upper quartiles for the tree heights in this table.
H eight of tree 0-5 11-15
Tip: This is the same to nearest m
data as on p19-20.
Number of trees Lindl il

« First add a row to the table showing cumulative frequency.

Height of tree
to nearest m 11-15} 16-20
LS Nuniber ortees 7 2eeylel eee
Tip: The method is the [Cumulative frequency| 26 | 43 | 54 | 60_|
same as the method on Now to find the quartiles, starting with the lower quartile (Or):
page 20 for estimating
the median of grouped a First
i calculate iia
ike 60 ie
— 15:
data. = Using the cumulative frequency, you can see that Q, is in the class 0-5.

Chapter 1 Data
 " Now draw a picture of the class containing Q, — just show
the important numbers you're going to need.
Put heights one side of the ‘ b, >
line (I’ve put them on top) 0 a, Q
Tip: Remember... 5,
5.5 should match the class
and cumulative frequencies
[+ ________
on the other side. frequency from the
0 —_— a, 15 (=3) 2° original frequency table.
+ ——->
b,
= Then solve the equation 4 = &.,
b, by

= Substituting in the numbers gives: eee

ba = 0 26220

» And so a — i OF-Oy=-5.5 & 42 = 3.2m

(to1d.p.)

Find the upper quartile (Q,) in the same way.

# First, calculate = = fe = A.
aoc OMisin ihe class 11-15.
= Now draw your picture of rl
the class containing Q,. 10.5 a, Oo. 155
{2
__—_-}—____________
43 a, 45 a4
Se ee es
b,
2 eubeathe eG eee
ubstituting numbers into Re. oe a
ee a5
UL 45 43
ess

- Or 105° 62 i Ds
And so ~—— = 74,0 Qa = 10.5 +5 x & = 11.4m(to 1d.p.)

b) Estimate the interquartile range for this data.

« Interquartile range (IQR) = Q, -Q,
= 11.4-—3.2 =(82 mie 1d).

Interpercentile range
You've seen how the quartiles divide the data into four parts, where each part Tip: The median is the
50th percentile and Q,
contains the same number of data values. Percentiles are similar, but they
is the 25th percentile,
divide the data into 100 parts. and so on.
The position of the xth percentile (P_) is700 x total frequency (7).

For example, to find the 11th percentile in a data set containing a total
of 200 values:
= 11
Calculate 700 200 = 22.
Tip: When you're
= Use linear interpolation to estimate the value in this position in the finding percentiles, the
ordered list of data values. data set is usually large,
and will probably be
You can find interpercentile ranges by subtracting two percentiles. grouped.
The a% to b% interpercentile range is P, — P..
= For example, the 20% to 80% interpercentile range is P,, — P,o.
Chapter 1 Data
 A reptile specialist records the mass
erred
(m, in kilograms) of 150 tortoises.
0.67 ma 110)eSTa
Her results are shown in the table.
ALOEme 4[7 GEE |
14<m<18Se
i
ee
a) Estimate the 10th percentile for this data.
» It'll help to add a column showing
cumulative frequency. Mass (kg) | Frequency| frequenc
. AO! Ol - 02 kin50:6 | 27
Now calculate T00
0.6sm<10| 43 |
1.0<m<1.4 a
* Using the cumulative frequency,
ei - 136
you can see that this will be in
the ‘0.2 <m< 0.6’ class.
* So draw a picture of this class b,
Tip: As always with showing the important masses ——> Ae
linear ee and cumulative frequencies. OP 2 ae oe
Ao
solve2 = 2, tae
«>
15 7
FE b;
>
b,
= The equation45 =n
= gives 02 eS
5— which means:
1 . a, .

Pro = 0.2 + 0.4 x 15

5 = 0.42kg
27
(to2d.p.)

a— Estimate the 90th percentile for this data.

Seer
Teale
* So the 90th percentile will be in the class ‘1.4 < m < 1.8’.

by
142 & P,, “18 R
105, 5, 135 136
b,
= Th e€ equation
tion if
2 =&s gi
gives
Po— 1.4 _ Le
13
ea,—5
136= 105
105°
AS ae1

This means: Py = 1.4 + 0.4 x 30 = 1.79kg

(to 2d.p.)

c) Find the 10% to 90% a range for this data.

" 10% to 90% interpercentile range= P,. — Po
= i 79 - 0.42 =4 a7 PER

Chapter 1 Data
Exercise 3.1
Ql The diameters (in miles) of the eight planets in the Solar System are
given below:
3032, 7521, 7926, 4222, 88846, 74898, 31763, 30778
For this data set, calculate:
a) the range
b) (i) the lower quartile (Q))
(ii) the upper quartile (Q,)
(iii) the interquartile range (IQR)

Q2 Each of the three data sets below shows the speeds (in mph) of
18 different cars observed at a certain time and place.
In town at 8:45 am: ;
Peon la loool, 10,4 oy tor tonto la, Sala, 17, 18, 15
In town at 10:45 am:
BO 3 1 29,3025, 29, 33e4e son al
On the motorway at 1 pm:
CO OM ae OOa) 4, (O97 1 5, Oy 99 0) ly OF nO, / 3.7 Opel Dy 4

For each set of data, calculate:

a) the range
b) the interquartile range (IQR)

Q3 The weights (w, in kg) of the items of baggage checked in on aflight

from Manchester to New York are shown in the table.
For this data, estimate: Weight of | Number of
a) the lower quartile (Q,) baggage (w) items
b) the upper quartile (Q,)
c) the interquartile range (IQR).
14<w<18 102
18<w<25 DS

Q4 The lengths of a zoo’s beetles [

Length (/) Number of beetles
measured to the nearest mm (/)
are shown in this table. 0-5 82
6-10 28
For this data, estimate: 11215 44
a) the 20th percentile 16305 | 30
b) the 80th percentile Se s0n Mul 16
c) the 20% to 80% interpercentile range.

UXBRIDGE COLLEGE
LEARNING CENTRE

Chapter 1 Data
 Outliers and box plots
Outliers
Tip: There are various An outlier is a freak piece of data that lies a long way from the
ways to calculate the majority of the readings in a data set.
values of the fences, To decide whether a reading is an outlier, you have to test whether it
but in an exam you will falls outside certain limits, called fences.
always be told what
method to use. A very common way to test for outliers is to use the following values
for the fences:
= for the lower fence, use the value Q, — (1.5 x IQR)
= for the upper fence, use the value Q, + (1.5 x IQR)

So using these fences, x is an outlier if:

either: x <Q, —(1.5 x IQR)
or: x>Q,+(1.5 x IQR)

The lower and upper quartiles of a data set are 70 and 100.
Use the fences Q, - (1.5 x IQR) and Q, + (1.5 x IQR) to
decide whether the data values 30 and 210 are outliers.
= First work out the IQR: Q, - Q, = 100 — 70 = 30
= Then you can find where your fences are.
Lower fence: Q>— (1.5 x IQR) = 70 =(1.5 x 30) =25
Upper fence: Q, + (1.5 x IQR) = 100 + (1.5 x 30) = 145
# 30 is inside the lower fence, so it is not an outlier.
210 is outside the upper fence, so it is an outlier.

Box plots
Tip: Box plots are
sometimes called ‘box A box plot is a kind of ‘visual summary’ of a set of data. Box plots show the
and whisker diagrams’ median, quartiles and outliers in an easy-to-look-at kind of Way.
(where the whiskers are They look like this:
the horizontal lines at
either end of the box).
Lowest value Lower Median Upper quartile
(that isn’t an outlier) quartile Highest value

a
Outliers marked \ (that isn’t an outlier)
with crosses

O25 4 Se? 8 10 327 4e Abe 00 i a

The scale is really important — always include one.

The box extends from the lower quartile to the upper quartile. —
(So the box shows the ‘middle 50%’ of the data.)
A vertical line drawn on the box marks the median.
® A horizontal line is drawn from each end of the box.
These lines extend in
each direction as far as the last data value that isn’t an outlier
.
Outliers are marked with crosses.

Chapter 1 Data
The back-to-back stem and leaf diagram below shows the IQs of year 11
students at two schools, Vale Academy and Cherry Comprehensive.
Vale Academy Cherry Comprehensive
(1) BAO 3 (1)
(8) 9 8 7 665 3.1/10/5 8 9 (3)
(5) SO 4 21 tl OM 2° 3"5' 76'S *9"(7)
(3) ime AS 22.) 0. 0 153 4 678 (7)
(2) 61/13 |2 4 (2)
(1) 0|14/ 4 (1)

Key: 1 | 10 | 5 represents an IQ of 101 at Vale Academy and

an IQ of 105 at Cherry Comprehensive
a) Draw a box plot to represent the data from Cherry Comprehensive.
Use Q, - 1.5 x IQR and Q, + 1.5 x IQR as your fences for
identifying outliers. '
= First work out the quartiles (Q,, Q, and Q,) and the fences.
« Find the number of data values (n) for Cherry Comprehensive.
Here, n = 21.

» Since 5 = 10.5, the median is the 11th data value.

So the median (Q,) = 119.
= Since a = 5.25, the lower quartile is the 6th data value.
So Q, = 112.

» Since = = 15.75, the upper quartile is the 16th data value.

So Q, = 124.
= The interquartile range is IQR = Q, - Q, = 124-112 = 12.
= This gives a lower fence of Q, — (1.5 x IQR) = 112 — (1.5 x 12) = 94.

= And an upper fence of Q, + (1.5 x IQR) = 124 + (1.5 x 12) = 142.

= Now you can decide if you have any outliers.
» The value 93 is outside the lower fence (94), so 93 is an outlier.
« And the value 144 is outside the upper fence (142), so 144
Tip: Remember... only
is also an outlier, but there are no other outliers. extend a line as far as
the biggest or smallest
= Now you can draw data value that isn’t an
the box plot itself. outlier.

90 100 110 120 130 140 150

Tip: Remember... when
b) Draw a box plot to represent the data from Vale Academy. you're counting along
= For Vale Academy: the values for the left-
» n=20. hand part of a back-
to-back stem and leaf
» “4 _ 10, so the median is halfway between diagram, you have to
2D
count along the leaves
the 10th and 11th data values — median (Q,) = 111.5.
from right to left.

Chapter 1 Data
 . i== 5, so the lower quartile is halfway between
the 5th and 6th data values — Q, = 106.
= 15, so the upper quartile is halfway between
“B's
the 15th and 16th data values — Q,= 123.
=» The interquartile range, IQR is Q, — Q, = 123 — 106 = 17.
= This gives a lower fence of Q, — (1.5 x IQR) = 106 — (1.5 x 17) = 80.5.
And an upper fence of Q, + (1.5 x IQR) = 123 + (1.5 x 17) = 148.5.
This means there are no outliers at Vale Academy.
® So the box plot PLDT

mel
ee Fae
PEEP CE EEE: PAPE ERP
SEER
EEE
Re
EEEEEE
ae
OOKS IIKe MIs:

90 100 iO. 120 130 190

Exercise 3.2
Q1__The lower and upper quartiles of a data set are 19 and 31.
Use the fences Q, — (1.5 x IQR) and Q, + (1.5 x IQR) to decide
whether the data values 4 and 52 are outliers.

Q2__ A set of data was analysed and the following values were found.
minimum value = 4, maximum value = 49
Q, = 16, median = 24, Q, = 37
a) Find the interquartile range.
b) Are there any outliers in this data set?
Use the values Q, — (1.5 x IQR) and Q, + (1.5 x IQR) as fences.
c) Draw a box plot to illustrate the data set.

Q3_ Atraffic officer counted the On aa (1)

number of cars usingatown 11/0 2° 7 i (3)
centre car park on different Oe Me es (3)
occasions. His results are in 3) TAO SU TANGe BITS BO 4 (o)
the stem and leaf diagram. Avlly 3256.6 EAS a O19)
D1)O old BSG 5, Goes
Keys 1 [7 = 17.cars a ae 5
7 NE 225 (2)
S10 (1)
Dial 9 (2)
a) Use the stem and leaf diagram to calculate:
(i) the median (ii) the lower quartile (Q.)
(iii) the upper quartile (Q,) (iv) the interquartile range (IQR)
(v) the lower fence, Q,=1-5 x1OR)
(vi) the upper fence, Q, + (1.5 x IQR)
b) Are any of the data values outliers? If so, which ones?
c) Draw a box plot to illustrate the data.

I Chapter 1 Data
 Q4 The numbers of items of junk mail received in a day by people living
in the towns of
Goossea and Pigham Goossea Pigham
are shown in this (3) Oiee 0-0
back-to-back stem Cea 7 (2)
andleaf diagram. Se rae) 2) 00. 2.6 G (4)
(3) Dee em so) 5 OS)
(2) AWA) 1 2 6 8 (4)
(1) O52 455 (3)
6
(1) BZ
Key: 4|2|0 represents 24 items of junk mail in Goossea
and 20 items of junk mail in Pigham

a) For the data from Pigham, calculate the following:

(i) the median (ii) the lower quartile
(ili) the upper quartile (iv) the interquartile range
(v) the lower fence, Q, — 1.5 x IQR
(vi) the upper fence, Q, + 1.5 x IQR
b) Are any of the data values from Pigham outliers?
Explain your answer.
c) Draw a box plot to illustrate the data from Pigham.
d) Draw a box plot to illustrate the data from Goossea.
(Use the same rules to find the fences as you used for Pigham.)

Variance and standard deviation

Variance and standard deviation are two (very closely related) measures of
dispersion — they give you an idea of how spread out the data values are Tip: You can think of the
from the mean. The bigger the variance (or standard deviation), the more formula on the right as
spread out your readings are. saying:
‘The variance is equal to
Variance the mean of the squares
= There are two ways to write the formula for the variance. ee ) minus the square
The second one in the box below is usually much easier to use.
of the mean (X°). /

variance =
De — ¥F Or
,
Varlance =
ee
Tip: Although the
formula on the right is
= Here, the x-values are the data, ¥ is the mean, and 7 is the total number easier to use, the one on
of data values. the left makes it easier
= The two formulas above are equivalent to each other — you can rearrange to understand what the
one to get the other (although you won't be asked to do this in the exam). variance actually is —
it’s ‘the average of the
squared deviations from
Dees = 1 (x? — 2+ F) (multiplying out brackets)
the mean’.
=15=2: ye Ly;x (writing as 3 summations)

=15 7-2 . “nx + ie (since >)x = nx) Tip: Remember... ¥ is

just a number, so you
ayy —2-FRio
34
emt
slo
+Fee (since })x°= nx’) can ‘take it outside the
summation’
aoe = ee (combining the 2nd and Le, ) Xe = ek
3rd terms)
Chapter 1 Data
 Standard deviation
variance.
Tip: The variance is The standard deviation is equal to the square root of the
measured in squared
units. For example, Standard deviation = vvariance
if the data values are
the data values
measured in metres The standard deviation is measured in the same units as
of dispersion
(m), then the variance is themselves — this sometimes makes it a more useful measure
measured in m?. than the variance.

BIE cscs
Example
Find the variance and standard deviation of the following data set:
Tip: The alternative 2,3, 4,4, 6911712

+ Fndthesumel ea 243 444 HOE EM=a

formula for the variance
is variance =
ee
But it’s more fiddly ¥ = S— => = 6
= Then finding the mean is easy:
to use this formula
than the one used in
* Next find the :
the example. This is
sum of Vir = 4 oF 16 tes 36 1214 144 aan
because once you've
worked out the mean, the squares:
you then have to * Now finding the ‘mean ye Mile 4G
subtract it from each
individual data value, of the squares’ is easy: ens)
before squaring and
= The variance is the ‘mean of the squares minus the square of the mean’:
adding the results, and 2
then dividing by n. Variance = pis — x = 330 _ 6 = 13.428... = 13.4 (to 3s)
The formula used on
the right involves only a = Take the square root of the variance to find the standard deviation:
single subtraction. Standard deviation = 713.428... = 3.66 (to 3s.f.)

If your data is given in a frequency table, then the variance formula can be
written like this, where fis the frequency of each x.

Tip: Here, n= >)f. variance = Qui —x*, where x =

ye
Tip: Remember...
fx? meansf x (x)
——Notnyx): Find the variance and standard x 213.415 7
deviation of the data in this table. | frequency, f| 2 | 5 | 5 | 4 1

Tip: You could also

write this formula as = It’s best to start by adding
an extra row to the table
ear ae Pa oseater
[4 lisfoopzo6
[7]
variance =
showing the values of fx.
But it’s trickier to use
than the formula in
the blue box — again, = The number of values is: })f=2+5+5+4+1+1=18
because of all the extra
subtracting you need = The sum ofthe values is: }) fe =44+15+20+20+6+7=72
to do.
= So the mean of the values is: ¥ = dF en
teal:
Chapter 1 Data
 . alge oo more beak [ l x2 4|9 |161251 361] 49
y € showingx* and fx’. | fre 8 | 45 | 80 |100] 36 | 49
= Now you can work out >>fa’.
> oe 45 + 80+ 100 + 36 49 = 318
* And now you can find the variance.
Ngee
Variance = SP —x 5|
= a — 4’ = 1.666... =Wh67 Gover)

= Then take the square root of the variance to find the standard deviation:
Standard deviation = /1.666... = 1.29 (to 3s.f.)

If your data is grouped, then you can only estimate the variance and standard Tip: This is the same
assumption you used
deviation (because you don’t know the actual data values — see p19).
when you found the
In this case, assume that each data value is equal to the class mid-point. mean of grouped data
Then go through the same steps as in the example above. on p19.

The heights of sunflowers in a garden were measured,

and are recorded in the table below. Tip: The variance is
Height of sometimes labelled o7
sunflower, / (cm) (where a is the Greek
Frequency, f letter ‘sigma’), and
sometimes s?. There is
Estimate the variance and the standard deviation of the heights. a difference, although
you don’t need to know
# Start by adding extra rows for the class mid-points x, about it for S1.
as well asfx, x? and fx?:
But if you're curious...
enter 150 <h<170|170<h<190|190<h<
210] 210 <h< 230 The symbol o? is
sunflower, / (cm) ; used when you have
12 information about an
200 entire population, and
the symbol s? is used
fi g00 . | 1800 2400 660 when you only have
x2 25600 | 32400 | 40000 48 400 information about a
480000 145 200 sample.
eg ly 128000 324 000
o and s’ are actually
ofvalues is: }) f =5 + 10+12 +3 =30
= The number worked out slightly
differently too.
= The sum of the values is: >)fr = 800 + 1800 + 2400 + 660 = 5660 (The method in this book
Ac is the one for working
= So the mean of the values is: ¥ = as = 2 out o”.)

= 2000 4000 +. 480000 + 145200 = 1077 200

37
2 2,

® So variance = ee = =dene — (298°) =—=(3.11.5555...

Tip: Remember...
—(312 cm? {to 3 sis. fig.) variance takes ‘squared’
units.
= And standard deviation = /311.5555... = 17.7 cm(to 3 sig. fig.)

Chapter 1 Data
 sets to
You can use the variances (or standard deviations) of two small data
ed set
find the overall variance (or standard deviation) of the larger, combin
of data. But it’s a bit fiddly, and so you need to do this very carefully.

The mean of 10 boys’ heights is 180 cm, with a standard deviation

of 10 cm. The mean height of 9 girls is 165 cm, with a standard
deviation of 8 cm.
Find the variance and standard deviation of the combined group of
19 heights.
You need to know the mean of the combined set of heights to be able to
work out the variance. So find that first.

Tip: You could use the = Call the boys’ heights x and the girls’ heights y.
formula from p15 | = First, write down the formula for the
to find the combined mean, and substitute in the mean for . yx
Meanie Gens ls | the boys (x) and the total number of Beg eed a
| oA et h m = — 1800
=c=_ 1X + h% values (n). This gives you the su
a of all the boys’ heights, x. 2
& See y= 22 5165 = 4% = py = 1485
Tip: Round the fraction of the girls’ heights, D/y.
one pase our = So the sum of the heights
NSS). URSIN, of the combined group pete Dy = 01800 451485 285
you use the mean in
more calculations, use
of boys and girls is:
the fraction (or your : The total number of boys and girls is10+9=19. 3985
calculator’s memory) so “T9 = 172.9cm
So the mean height of the combined group is:
you don't lose accuracy.

Now you need to go through a very similar process to find the sum of
the squares of the combined set of heights.
* The standard deviation for the boys is 10, so the variance = 10? = 100.
The standard deviation for the girls is 8, so the variance = 8? = 64.
Tip: Here, variance. is : * Substitute the boys’ variance into the formula for the variance.
the Warianeetontne boys’ This will give you the sum of the squares of their heights, >)x’.
heights, and variance, is ee “s ye
the variance of the girls’ Vallance t= 1 Oa 130100
heights.
i So )ix?= 10 x (180?+ 100)= 325000
= Now do the same for the girls to find >)y’.
variance, = Dow
fi 7 ee
2

64
So >}y = 9 x (165’+ 64) = 245601
= So the sum of the squares for the combined group of boys and girls is:
Dx + yy’ = 325000 + 245601 = 570601

Chapter 1 Data
 * You now have all the information you need to find the variance for the
combined group. Tip: Don't use the
{ 570601 7 rounded mean (172.9)
variance = “a — (2282) = 139.0415... =(139°0emr (to 1d.p.) — you'll lose accuracy.

= And finally the standard deviation of the boys and the girls is:
standard deviation = /139.0415... = 11.8 cm (to 1d.p.)

You've seen that different measures of location are useful in different ways
(see page 22). The same is true of measures of dispersion — the range,
interquartile range, variance and standard deviation all have pros and cons.

Range
= The range is the easiest measure of dispersion to calculate.
= But it’s heavily affected by even a single extreme value / outlier.
And it depends on only two data values — it doesn’t tell you anything
about how spread out the rest of the values are.

interquartile range
= |t’s not affected by extreme values — so if your data contains outliers, then
the interquartile range is a good measure of dispersion to use.
® It’s fairly tricky to work out.

Variance
= The variance depends on all the data values — so no values are ‘ignored’.
= But it’s tricky to work out, and is affected by extreme values / outliers
(meaning that ‘freak’ values have more influence than they deserve).
= It’s also expressed in different units from the actual data values, so it can
be difficult to interpret.

Standard deviation
= Like the variance, the standard deviation depends on all the data values —
so no values are ‘ignored’.
= |t has the same units as the data values themselves, and so is easier to
understand and interpret. »
= But it is also tricky to work out, and affected by extreme values / outliers.

Exercise 3.3

Q1_ The attendance figures (x) for Wessex Football Club’s first six matches
of the season were: 756, 755, 764, 778, 754, 759.
a) Find the mean (x) of these attendance figures.
b) Calculate the sum of the squares of the attendance figures, Die
c) Use your answers to find the variance of the attendance figures.
d) Hence find the standard deviation of the attendance figures.
e) ) Explain why the standard deviation is a reasonable measure of
dispersion to use with this data.
Chapter 1 Data
 Q2 The number of runners in the first eight Broughton marathons were:
322107 5722).88, 998R7Age
a) Calculate the variance for these figures.
b) Find the standard deviation of the number of runners.

Q3 The figures for the number of TVs (x) in the households of 20 students
are shown in the table below.

frequency,
f]7|8[4 |1)
a) Find the mean number of TVs (x) in the 20 households.
b) By adding rows showing x? and fx? to the table, find >) fi’.
c) Calculate the variance for the data above.
d ) Hence find the standard deviation.

Find the variance of the data [ = x | 7] 8|9 [10]11|12|

inthis equeney table.
Q4

Q5 The pulse rates while resting of a number of students were measured.

They are shown in the grouped frequency table below.
(You can assume that a pulse rate is always a whole number.)

frequency]
1 [627 sear a ee
a) Add four extra rows to the table showing:
(i) the class mid-points (x)
(ii) fc (ili) x? (iv) fx?
b) Use your table to find: (i) }}f (ii) peyitibe. (iii) ope
c) Use your answers to estimate the variance of the pulse rates.

Q6 The yields (w, in kg) of potatoes

from a number of allotments is
shown in the grouped frequency
50 <w<60
table on the right.

[s0sw<90
|6
a) Estimate the variance for this data.
b) Estimate the standard deviation.

Q7 Su and Ellen are collecting data on the durations of the eruptions of

the volcano in their garden. Between them, they have recorded the
duration of the last 60 eruptions.
° Su has timed 23 eruptions, with an average duration of 3.42 minutes
and a standard deviation of 1.07 minutes.
° Ellen has timed 37 eruptions, with an average duration of
3.92 minutes and a standard deviation of 0.97 minutes.
They decide to combine their observations into one large data set.
a) Calculate the mean duration of all the observed eruptions.
b) Find the variance of the set of 60 durations.
c) Find the standard deviation of the set of 60 durations.
Chapter 1 Data
Coding
Coding means doing something to all the readings in your data set to make
the numbers easier to work with. That could mean: Tip: Choose what
= adding a number to (or subtracting a number from) all your readings, to add/subtract and
multiply/divide by based
= multiplying (or dividing) all your readings by a number, on what makes your
= both of the above. data easiest to work
with.
For example, finding the mean of 1831, 1832 and 1836 looks complicated.
But if you subtract 1830 from each number, then finding the mean
of what's left (1, 2 and 6) is much easier — it’s 3. So the mean of
the original numbers must be 1833 (once you've ‘undone’ the coding).
That’s coding in a nutshell.

® You have to change your original variable, x, to a different one, such as

y (so in the example above, ifx = 1831, then y = 1).
= An original data value x will be related to a
coded data valuey by an equation of this form: y= 5
where a and b are numbers you choose.
= The mean and standard deviation of the original data values will then
be related to the mean and standard deviation of the coded data values
by the following equations:
Tip: Note that if you
don’t multiply or
a Ls — @ where x and y are the means of variables x andy divide your readings by
anything (i.e. if b = 1),
" standard deviation of y = standard deviation
5 of x
then the dispersion isn’t
changed.

Find the mean and standard deviation of: 1862 020, 1862040, 1862010
and 1862050.
« All the original data values (call them x) start with the same four digits
(1862) — so start by subtracting 1 862 000 from every reading to leave
20, 40, 10 and 50.
= You can then make life even simpler by dividing by 10 — giving 2, 4, 1
and 5. These are the coded data values (call them y).
= So putting those steps together, each x-value is ey = 1862000 Tip: This means
related to a corresponding y-value by the equation: y = “——zy5———— a = 1862000 and
bieallO:
= Now work out the mean and
standard deviation of the y= 2+ 44145512
4 ad
(easy-to-use) coded values.
2 2 2 2
Tip: Remember...
standard deviation of y = j2teet te — 3? standard deviation

= /46
— 8fig.
to3sig.
9 = /25 = 1.5

= Then find the mean and standard deviation of the original values
using the formulas above.
y= oe a + by. This means: x = 1862000+ 10y
— 1862000
+(10 x 3)
= 1862030

Chapter 1 Data
 ae f
» And standard deviation ofy = standard deviationofx
So standard deviation ofx = b x standard deviation of y
= 10x 1.58 = 15.8 (to3 sig. fig.)

Carry out the method in exactly the same way with grouped data. Remember,
with grouped data, you assume that all the readings equal the class mid-point,
and so this is the x-value that you use with the coding equation y = ~ :
. . . — a

Example oe oes

Estimate the mean and standard deviation of this eee

data concerning job interviews using the coding y = San
Tip: So here, a= 15.5 Length of interview, to nearest minute
and b = 10. Frequency,f

* Make a new table showing the class mid-points (x) of the original data,
and the coded class mid-points (vy). Also include rows for fy, ? and fy”.
Length of interview, to nearest minute

[| ___.Coded value, 7 aaa nc ae ae

[i ER ag fama in
y 4 9
pe

Now use your table to find the mean of the coded values (7).
« The number of coded values is: })f = 17 + 21+27+15 =80
= The sum of the coded values is: }\fy =0 + 21+54+45 = 120
= So the mean of the coded values is: y = dS = fe = 1.5
ae
Now for the standard deviation...
2S fy =0+ 21 = 108+
135 = 264
2

" So variance =LSY

ening:
sy & 264 deg 5 ee IOS

= This gives a standard deviation for the coded data of:

standard deviation of y= 71.05 = 1.02 (to 3 sig. fig.).
_And now you can use these figures to find the mean and
standard deviation of the original data.
2 y= ee = a+ by = 15.5 + 10x 1.5 =30.5 minutes

"= And standard deviation ofy = Standard deviation ofx

So standard deviation ofx = b x standard deviation of y
= 10 x 1.02 = 10.2 minutes (to 3 sig. fig.)
= This means the standard deviation of
the interview lengths is 10.2 minutes (to 3 sig. fig.) .

Chapter 1 Data
Sometimes, you won't have the data itself — just some summations.

A travel guide employee collects some data on the cost (c, in £) of a

night’s stay in 10 hotels in a particular town. He codes his data using
d = 10(c — 93.5), and calculates the summations below. Tip: The variables aren't
x andy this time, but
>4d= 0 and 91d’ = 998 250 you can still go through
Calculate the mean and standard deviation of the original costs. exactly the same steps
First find the mean of the coded values: as before.

=» The number of values (v) is 10.

» So the mean of the coded values is: d = Bee = 5= 0

Now for the standard deviation of the coded values:

i ae 95
=" variance = Due a wn oe = 0’ 9— 993825

» This gives a standard deviation for the coded data of:

standard deviation = /99 825 = 316.0 (to4 sig. fig.).

Now you can find the mean and standard deviation of the original data.
= d= 10(¢— 93.5), sot = 93.5 + 4 = 93.5, ie. & = £93.50
= standard deviation of d = 10 x standard deviation of c,
so standard deviation of c =
standard deviation
10 ofd _ 316.0
ia _ 31 69

That is, standard deviation of c = £31.60 (to the nearest penny) .

There are lots of different ways to ask questions about this topic.

Example ee

A set of 10 numbers (x-values) can be summarised as shown below:

di@ — 10) = 15 and >)— 10)’ = 100
Find the mean and standard deviation of the numbers.
= You can simplify those summations if you writey = x — 10.
This is the same as coding the x-values.
= Rewriting the summations means: })y = 15 and iy 1,00
= Now you can work out the mean and standard deviation of the coded
data in the normal way.
Tip: The dispersion (i.e.
. eee ilies the standard deviation)
ofx is the same as the
» Variance of coded data = 2) -y = ch 5 = Oe 2 5 27/5 dispersion ofy since
you've only subtracted
So standard deviation of coded data = V7.75 = 2.78 to 3sig. fig. 10 from every number
Now finding the mean and standard deviation of the x-values Is easy:
— you've not done
= anything to change how
7=x— |0sox=y+10= 1.5 +10 = ee spread out or tightly
packed they are.
Standard deviation of x = standard deviation of y
= 2.78 (to3 sig. fig.)

Chapter 1 Data
 Exercise Cy
tion of the
Ql In each case below, find the mean and standard devia
original data sets.
a) A set of data values (x) are coded usingy = x — 500.
ard deviation
The mean of the coded data (7) is 12, and the stand
of the coded data is 4.22.
Q1 Hint: In part b), b) A set of data values (x) are coded usingy = 4x.
tion
=| The mean of the coded data (7) is 6, and the standard devia
of the coded data is 2.14.
x — 20 000
c) A set of data values (x) are coded using y = TS
The mean of the coded data (7) is 12.4, and the standard
deviation of the coded data is 1.34.

Q2 In each case use the given coding to find the mean and standard
deviation of the given data.
a) 2003, 2007, 2008" Use the coding: y = x — 2000

b) 20:02) 0217,-0.03 20, 1ir-0:07, Use the coding: y = 100x

C1935 305.235 14360; 357.5 Use the coding: y = 2(x — 350)

d) -7900, -7930, -7960, —8000, —7940
8000
Use the coding: y = x + 10

Q3 The widths (in cm) of 10 sunflower seeds in a packet are given below.
0.61, 0.67, 0.63, 0.63, 0.66, 0.65, 0.64, 0.68, 0.64, 0.62
a) Code the data values above (x) to form a new data set consisting
of integer values (vy) between 1 and 10.
b) Find the mean and standard deviation of your coded values (y).
c) Use your answers to find the mean and standard deviation of the
original values (x).
»

Q4 The table below displays the weight (x, to the nearest gram) of 12
items on a production line.
Weight (to nearest g)
Frequenc

Use the coding y = x — 102 to estimate the mean and standard

deviation of the items’ weights.

Q5 Twenty pieces of data (x) have been summarised as follows:

(e+ 2)=7 and \i(x + 2 = 80.
Calculate the mean and standard deviation of the data.

Chapter 1 Data
Skewness
Skewness tells you how symmetrical or asymmetrical (lopsided) the
distribution of your data is.

= This is a typical symmetrical

distribution.
= Notice: mean = median = mode

= A negatively skewed distribution

has a ‘tail’ on the left.
= This means most data values are
relatively high, but there are some Tip: It’s these ‘extreme’
‘extreme’ low values. low (or high) values that
reduce (or increase) the
= Notice: mean < median < mode mean and make it an
unsuitable measure of
® A positively skewed distribution location for a data set
has a ‘tail’ on the right. that isn’t symmetrical —
: oD
= This means most data values are a aa
relatively low, but there are some
‘extreme’ high values.
= Notice: mode < median < mean

For nearly all distributions, the following is approximately true:

mean — mode = 3 x (mean — median)

A primate keeper at a zoo records the mass (in kilograms) of fruit eaten by
50 monkeys and apes in one morning. Her results are shown in the table.
Mass (kg) | 0.2-0.6|0.6-1.0| 1.0-1.4 |1.4-1.8 |1.8-2.2 |2.2-2.6 |2.6-3.0
[Frequency| 10 14 11 8 4 2 1
Draw a histogram of the data, and state the type of skewness present.
14 Tip: All the classes are
# Draw the histogram. 12 the same width, so I’ve
plotted the frequency,
# This has a ‘tail’ on the =°
rather than the
right-hand side. © frequency density.
Frequency
= Most of the values are a

4 Plotting the frequency

relatively low.
erates density wouldn’t change
® So this distribution is 2 the ee of the graph
positively skewed . 0
ONO O60 eA Bi 202 E26: 3.0)
at all.
Mass of fruit (kg)

Chapter 1 Data
 You don’t always have to judge skewness by eye, though.
p-one-end’
You can use a coefficient of skewness to measure how ‘all-u
Tip: These two your data is. There are a couple of important coefficients of skewness
definitions should give — the first of these is Pearson’s coefficient of skewness.
approximatel y the same
3(me an — median)
answer, because for
nearly all distributions,
icient t of
Pearson’s coefficien o skewness = —— 7_
standard 4devia
_7 tion
‘mean — mode’ and
_ mean — mode
‘3 x (mean — median)’
standard deviation
are approximately equal
(see p41).

(In the exam, you'll be Pearson’s coefficient of skewness usually lies between —3 and +3.
told what formula to = Positive values mean that your data is positively skewed.
use.)
= Negative values mean that your data is negatively skewed.
= Asymmetrical distribution would give a skewness coefficient of zero.
= The more highly skewed your data is, the closer to 3 or -3 Pearson’s
coefficient of skewness will be.

So a coefficient of 2.7 would mean your data is very heavily positively

skewed — with a lot of low values but a long tail on the right.
But if Pearson’s coefficient of skewness is —0.1, then the distribution is
slightly negatively skewed.
Tip: You have to use the
formula involving the
median, because you
don’t have afigure for
the mode.
For the ‘primate’ data in the previous example, the mean = 1.136,
the median = 1.036 and the standard deviation = 0.60.
Find Pearson’s coefficient of skewness for the data.
Tip: Pearson’s
coefficient of skewness 3(mean — median)
# Pearson’s coefficient of skewness = ——
is positive and not too standard deviation
large — just as you’d
expect it to be from _ 3 Xx (1.136— 1.036)
the histogram on the 0.60 =
previous page.

The quartile coefficient of skewness is another coefficient of skewness.

Tip: Remember thatQ,
is the lower quartile, Quartile coefficient of skewness = (Qs — Q2) = (Qa = Qi) = Q3 = 2Q:+Q)
Q, is the upper quartile, Q;—Q OF
and Q, is the median.

Although it’s calculated using a different formula, you interpret it in much the
same way as Pearson’s coefficient of skewness.
= A positive quartile coefficient of skewness means your data is positively
skewed, while a negative quartile coefficient of skewness means your data
is negatively skewed.
= The larger the quartile coefficient of skewness, the more highly skewed
your data is.
= A value of zero means your data is symmetrically distributed.

Chapter 1 Data
A box plot can quickly tell you the skew of a data set. Each of these three
histograms has its corresponding box plot shown underneath.

= IfQ,-Q,=Q,-Q, then the quartile

coefficient of skewness will be zero.
= This corresponds to a symmetrical
distribution of data and a
symmetrical box plot.

# IfQ,-Q,>Q,-Q then the quartile

coefficient of skewness will be positive,
showing that the distribution is
positively skewed.
= This corresponds to a long tail on the ,
right-hand side of the histogram,
and to Q, (the median) being closer
to Q, than to Q, on the box plot.

# IfQ,-Q,<Q,-Q then the quartile

coefficient of skewness will be negative,
showing that the distribution is
negatively skewed.
= This corresponds to a long tail on the
left-hand side of the histogram,
and to Q, (the median) being closer
to Q, than to Q, on the box plot.

Tip: These figures

The masses (in kilograms) of fruit eaten by 50 monkeys and apes in a zoo describe the ‘primate’
one morning have Q, = 0.671, Q, = 1.036, and Q,= 1.525. data in the example on
page 41.
Find the quartile coefficient of skewness for this data.
Or 2Or1 1
# Quartile coefficient of skewness = Oai@y
=e
ted 25 9290036 520:671 Tip: Again, this is
ia 1525) 0671 positive but not too
large (which is what
= 0.145 (to 3 d.p.) you'd expect from the
histogram on page 41).

Chapter 1 Data
 1 |9 e
The ages of 15 lottery es
winners are given in this : ;:,
t
7a Semen
d leaf diagram.
4|1479 (4)
Sil ez ese/, (4)
Key 3|8 3 means 38 years of age

ee. _ 3(mean — median)

e
Calculate Pearson’s coefficient of skewness using <eqntiard deviation
I vaou e
a—

= The sum of the ages is })x = 600. So the mean age is a = 40.
= The sum of the squares is })x° = 26102.
So the standard deviation is oa — 40? = 11.84 (to 2 d.p.).

= There are 15 values, so the median is the 8th value in the ordered list.
This means the median = 41.
; — 4]
= So Pearson’s coefficient of skewness isSGU = -0.253 (to 3 dip.)

Tip: These two answers

are similar, but not : b) Calculate the quartile coefficient of skewness.
oa. = As there are 15 values, the lower quartile is the 4th value in the list.
So Q, = 27.

E = The upper quartile is the 12th value in the list. So Q, = 51.

Tip: This small, negative
coefficient of skewness * The quartile coefficient of skewness is
52 eA le7
corresponds to the . Bit 27
slight negative skew = -0.167 (to 3 d.p.)
you can see on the
stem and leaf diagram i
(though you have to c) Use your answers to describe the skewness of the data.
imagine it rotated 90° * These skewness coefficients are both small and negative,
anticlockwise). so the data is slightly negatively skewed. ‘

Exercise 3.5

QI Match each diagram below to one of the descriptions in the box:

positively skewed negatively skewed — symmetrical
a) b) C) 48
agit etat eats 8
ues
AAS ST7 9
©
—
BON
NONN

Chapter 1 Data
 Q2__ Find Pearson’s coefficient of skewness for each distribution below,
and use your answer to say whether the distribution is positively
skewed, negatively skewed or symmetrical.
a) mean=7 mode = 5 standard deviation = 1.6
b) mean = 12 median = 13.5 standard deviation = 1.8
c). mean = 14 mode = 14 standard deviation = 2.37
d)) mean =—-9 median = —8 standard deviation = 2.5
Q3__ Find the quartile coefficient of skewness for each distribution below,
and use your answer to say whether the distribution is positively
skewed, negatively skewed or symmetrical.
Se) Ag MC 15
b) Q,=35 Q.=119 Q,=194
c) lower quartile =—28 median=-23 — upper quartile =-10
d) lower quartile= 2.4 median =5.3 upper quartile = 8.2
Q4_ The following data shows the winnings (in pounds)
of a bingo player one summer: 1, 6, 9, 10, 10, 10, 11, 11
a) Find the mean, median and standard deviation of the data.
3(mean — median)
b) Find Pearson’s coefficient of skewness using
standard deviation ©
c) Use your answer to describe the skewness of the distribution.
Q5__ The salaries (in thousands of pounds) of a small firm’s ten
employees are shown below.
2A Bil G2 24 22,722, D4, 24127;,60
mean — mode
a) Calculate Pearson’s coefficient of skewness using standard deviation’
b) Use your answer to describe the skewness of the distribution.

Comparing distributions
It’s really common in the exam to be asked to compare two distributions.
To do this, there are different kinds of things you can say, depending on what
information you have about the distributions. You can:
= Compare measures of location, such as the mean, median or mode.
— You'll need to say which distribution has the higher mean/median/mode,
Tip: ‘...in the context of
and by how much. the question’ means you
— Then say what this means in the context of the question. need to use the same
= Compare measures of dispersion, such as variance, standard deviation, ‘setting’ in your answer
as the question uses.
range, interquartile range, or interpercentile range.
— You'll need to say which distribution’s data values are more ‘tightly For example, if the
packed’, or which distribution’s values are more spread out. question is all about the
weights of tigers in a
— Then say what this means in the context of the question. zoo, then you need to
= Compare measures of skewness — such as Pearson's coefficient of talk about the weights of
skewness or the quartile coefficient of skewness. tigers in the zoo in your
answer as well.
— For each distribution, say whether it is symmetrical, positively skewed,
or negatively skewed, and whether it is highly skewed, or just slightly.
— ‘Then say what this means in the context of the question.
Chapter 1 Data
 This table summarises the marks obtained by a group of students in
Maths ‘Calculator’ and ‘Non-calculator’ papers.
a) Comment on the location and dispersion of the distributions.
Calculator paper oo, Sen Non-calculator paper
Sais hecasesdl Lower quartile, Q, | 35
35

ri nae
Location:
« The mean and the median are both higher for the Calculator paper
(the mean is approximately 9 marks higher, while the median is 16
marks higher).
= This means that scores were generally higher on the Calculator paper.
Dispersion:
= The interquartile range (IQR) for the Calculator paper is
Q,
-Q, = 70
-40 = 30.
= The interquartile range (IQR) for the Non-calculator paper is
Q,=Q, 256-35 =21.
= So the IQR and the standard deviation are both higher for the
Calculator paper.
« This means the scores on the Calculator paper are more spread out
than those for the Non-calculator paper.

b) Calculate Pearson’s coefficient of skewness for each paper,

3(mean — median)
using
standard deviation *
Comment on your answers.
For the Calculator paper:
" Pearson’s coefficient of skewness =eens) = 70.425 (0 3 dp)

For the Non-calculator paper:

* 3 XAAG ded)
Pearson’s coefficient of skewness = 178 = 0.691 (to 3 dp.)
Tip: Don’t forget to
= The Calculator paper scores are negatively skewed.
give your answer in This means that most scores were relatively high,
the context of the but there was a‘tail’ of low scores.
question, so here you = The Non-calculator paper scores are positively skewed.
need to talk about This means that most scores were relatively low,
scores on Calculator and | but there was a ‘tail’ of high scores.
Non-calculator papers. |

Chapter 1 Data
You may have to base your comparison on a graph instead of numbers.

The box plots on the right Harvest 1: 7 }—

———_{____
show how the masses (in g) Harvest 2: lee
of the tomatoes in two Ua
etagei el
harvests were distributed. Os 20 208 60s 60 100
Mass (g)
Compare the distributions of the two harvests.

Location:
= The median is more than 25 g higher for Harvest 1.
« This shows that the tomatoes in Harvest 1 were generally heavier.
Dispersion:
« The interquartile range (IQR) and the range for Harvest 1 are higher
than those for Harvest 2.
# This shows that the masses of the tomatoes in Harvest 1 were more
varied than the masses of the tomatoes in Harvest 2. ‘ nes
Tip: This question is
Skewness: to do with harvests of
=" The distribution of the tomatoes’ masses in Harvest 1 is negatively tomatoes, so you need
skewed, showing that there isa tail on the left of this distribution. ee INES Se2
= This means the masses of the tomatoes were generally relatively high,
but there were some much lighter tomatoes as well.
» Harvest 2 is very different — for this harvest, the distribution of the
tomatoes’ masses is positively skewed, showing that there is a tail on
the right of this distribution.
# This means the masses of the tomatoes were generally relatively low,
but there were some much heavier tomatoes as well.

Exercise 3.6

Q1__The box plots below represent the prices of shoes (in £) from two
different shops.

A et
oe eee ee
OS eta le. ee
Ome Ome On eoC 1405950) 460" 70) 80
Price (£)

Use the box plots to compare:

a) the location of the two shops’ prices
b) the dispersion of the two shops’ prices
c) the skewness of the two shops’ prices

Chapter 1 Data
 Q2 This back-to-back stem and leaf diagram shows the daily sales of two
competing small cafés over a 14-day period.
Café A CaféB
(0) 5 Ss 4.9 (3)
(1) BoiNGeH ht" 2 (2)
(6) Oc Fre | 7 Ow (2)
(7) Qe Zeats 4 4 Aug de le Be 86. (6°%7 (3)
(0) 9/05 (2)
(0) 10 | 8 (1)
(0) 1S (1)
Key: 8|6]|1 represents £68 of sales in CaféA
and £61 of sales in Café B

a) )( i) Calculate the mean sales and the median sales for both cafés.
(ii) Use your answers to compare the locations of the two data sets.
b) (i) Calculate the standard deviation for both cafés’ sales.
(ii) Use your answers to compare the dispersion of the two data sets.
c) )( (i) Find the quartiles for both cafés.
Hence find the quartile coefficient of skewness of each data set.
(ii) Use your answers to compare the skewness of the two data sets.

Q3 10 men and 10 women were asked how many hours of sleep they got
on a typical night. The results are shown below.
Men: 0.9, OO, Oe OneD
Wroirnen?s 950 67825 eld 2).0,nGal ar
a) (i) Calculate the mean and median of the individual data sets.
(i!) Use your answers to compare the locations of the two data sets.
b) (i) Calculate the standard deviation of the individual data sets.
(ii) Use your answers to compare the dispersion of the two data sets.
) i) Find Pearson’s coefficient of skewness
cM
for each data set, using 3(mean — median)
standard deviation *
(ii) Use your answers to compare the skewness of the two data sets.

Chapter 1 Data
 Review Exercise — Chapter 1

_ QI The stem and leaf diagram on the right represents the lengths Taee
(in cm) of 15 bananas. Write down the original data asalist. S| F295
TA} 3 (3256988
Tea \e2e29
Key 12|8 means 12.8cm| 16] 1 1 2 3
TS Or 2

Q2_ Twenty phone calls were made by a householder one evening.

The lengths of the calls (in minutes to the nearest minute) are recorded below.
Draw a histogram of the data.
Length of calls Oi2 | 3sor 6 Bia 9 15
Number of calls ,| 10 6 3

Q3 Calculate the mean, median and mode eae ee eo eal aa

of the data in the table on the right.

Q4 The speeds of 60 cars travelling in a 40 mph speed limit area were measured to the nearest mph.
The data is summarised in this table.
Calculate estimates of the mean and median, po ea
Een 34 |35- IOs =|
and state the modal class. Frequency 2

Q5__ Find the mean and standard deviation of the following mumbers> it, F214, 79 21R ose.

Q6 The scores in an IQ test for 50 people are recorded in the table below.
Score cee vats Pie 20 |121 27 L281
Frequenc 22 9 Z

Estimate the mean and variance of the distribution.

Q7 For aset of data, n = 100, }(x — 20) = 125, and > (x — 207 = 221.
Find the mean and standard deviation of x.

Chapter 1 Data
Q8_ The time taken (to the nearest minute) for a commuter to travel to work on
20 consecutive work days is recorded in the table.
Use the codingy = x — 35.5 to estimate the mean and standard deviation of
the times, where x is the class mid-point.
Time to nearest minute | 30-33 | 34-37
Frequencyi Liat |. @ Be Foun eee pa

Q9 A data value is considered an outlier if it’s more than 1.5 times the IQR above the upper quartile
or more than 1.5 times the IQR below the lower quartile. If the lower and upper quartiles of a data
set are 62 and 88, decide which of the following data values are outliers:
a) 124 b) 131 c) 28

Q10 a) Find the median and quartiles of the data below.

Amount of pocket money (in £) received per week by twenty 15-year-olds:
1025,-20, 50; 5,.1,°6,5) 158205 507757
10) 12. 4 Bon 7 a0:
b) Using the values Q, - 1.5 x IQR and Q, + 1.5 x IQR as fences, determine whether any of
these values are outliers.
c) Draw a box plot to illustrate the data and use it to comment on any skewness.

Q11 A set of data has a median of 132, a lower quartile of 86 and an upper quartile of 150.
Calculate the quartile coefficient of skewness, and draw a possible sketch of the distribution.

Q12 Two workers iron clothes. Each irons 10 items, and records the time it takes for each,
to the nearest minute:
WorkerA: 35271045
5 4 12
WorkerBey 494<8 677.8.9.10 14 9
a) For worker A’s times. Find:
(i) the median
(ii) the lower and upper quartiles
(iii) whether any of these values are outliers, using the values
Q, — 1.5 x IQR and Q, + 1.5 x IQR as fences.
b) On graph paper draw, using the same scale, two box plots
to represent the times of each worker.
c) Make one statement comparing the two sets of data.
d) Which worker would be better to employ? Give a reason for your
answer.

Chapter 1 Data
 Exam-Style Questions — Chapter 1
The profits of 100 businesses are given in the table.
Profit, £x million _| Number of businesses
4.0<x<5.0 24
5.0<x< 5.5
5.5<x< 6.0 21
6.0<x<6.5
65<x<8.0 |
a) Represent the data in a histogram. (3 marks)

b) Comment on the distribution’of the profits of the businesses.

(2 marks)

A group of 19 people played a game. The scores, x, that the people achieved are
summarised by:
>\(x — 30) = 228 and +(x — 30 = 3040
a) Calculate the mean and the standard deviation of the 19 scores. (3 marks)

by Show that > x = 798 and > x* — 33820. aman)

c) Another student played the game. Her score was 32.

Find the new mean and standard deviation of all 20 scores. (mene)

In a supermarket two types of chocolate drops were compared.

The masses, a grams, of 20 chocolate drops of brand A are summarised by:
60.3 Se =—219
The mean mass of 30 chocolate drops of brand B was 2.95 g,
and the standard deviation was | g.
a) Find the mean mass of a brand A chocolate drop. ome

b) Find the standard deviation of the masses of the brand A chocolate drops. rrens)

c) Compare brands A and B. | (2 marks)

d) Find the standard deviation of the masses of all 50 chocolate drops. (4 marks)

Chapter 1 Data
 4 The stem and leaf diagram shows the test marks for 30 male students
and 16 female students.
Male students Female students
eee
5
OFT. 6 6 l 3 4x5. 60729
99 1 9
Key 5/6|2 means Male student test mark 65
and Female student test mark 62

a) Find the median test mark of the male students. (1 mark)

b) Compare the distribution of the male and female marks. (2 marks)

The table shows the number of hits received by people at a paint ball party.

[Prequendy.)2
1/74] 65k78) 6 |Be aR ae | oO Oe
Find the median and mode number of hits. (3 marks)

An outlier is a data value which is more than 1.5 x (Q, — Q,) above Q, or
more than 1.5 x (Q,—Q,) below Q.. Is 25 an outlier? Show your working. peri

Sketch a box plot of the distribution and comment on any skewness. Oman

How would the shape of the distribution be affected if the value of 25 was
removed? (Priel)
2

The histogram shows the nose-to-tail lengths of 50 lions in a game reserve.

density
Frequency

150 160 170 180 190 200 210 220 230 240 250 260 270
Length (cm)
Use the histogram to find the number of lions who measured
over 220 cm from nose to tail.
(5 marks)

Chapter 1 Data
 Chapter 2 Probability

1. Elementary Probability
Probability is a measure of how likely events are to happen. We’re errs Learning Objectives:
with a reminder of the basics, which you'll have seen before. But it’s all
econenpeanae!

" Understand the

important stuff as you'll be using it throughout the rest of the chapter. meanings of the terms
The basics of probability Peo
" Be able to calculate
In a trial (or experiment), the things that can happen are called outcomes. probabilities of events
For example, if you roll a six-sided dice, the numbers 1-6 are the outcomes. when all the outcomes
= Events are ‘groups’ of one or more outcomes. Soa possible event for the are equally likely.
dice roll is that ‘you roll an odd number’ (corresponding to the outcomes Be able to identify
1,3 and 5). If any outcome corresponding to an event happens, then you sample spaces.
can say that the event has also happened.
= When all the possible outcomes are equally likely, you can work out the
probability of an event using this formula:
Tip: ‘P(event)’ is short
Number of outcomes where event happens for ‘the probability of an |
P(event) = 2
Total number of possible outcomes event’.

= Remember, the probability of any event has to be between 0 (the event is

Tip: Remember, you |
impossible) and 1 (the event is certain to happen). re oreeines
as fractions, decimals or
Examples Sl cela percentages. :

A bag contains 15 balls — 5 are red, 6 are blue and 4 are green.
If one ball is selected from the bag at random, find the probability that:
a) the ball is red
= The event is ‘a red ball is selected’.
# There are 5 red balls, so there are 5 outcomes where the event happens
and 15 possible outcomes altogether.
» So: P(red ball) = ae= 3 ‘ It’s usually best to simplify your
answer as much as possible.
b) the ball is blue
= The event is ‘a blue ball is selected’.
= There are 6 blue balls, so there are 6 outcomes where the event happens
and 15 possible outcomes altogether.
: Tip: Always check
» So: P(blue ball)= Be= z that your probability is
on between 0 and 1. |
c) the ball is red or green
= The event is ‘a red ball or a green ball is selected’.
= There are 5 red balls and 4 green balls, so there are 9 outcomes where
the event happens and 15 possible outcomes altogether.
= So: P(red or green ball) = z:= 2

Chapter 2 Probability
 The sample space
The sample space (called S) is the set of all possible outcomes ofatrial.
Drawing a diagram of the sample space can help you to count the outcomes
you're interested in. Then it’s an easy task to find probabilities using the
formula on the previous page. |
lf atrial consists of two separate activities, then a good way to draw your
sample space is as a grid.

Tip: Because the dice A six-sided dice is rolled twice.

is being rolled twice,
the outcomes here are a) Draw a sample-space diagram to show all the possible outcomes.
a combination of the = Draw apair of axes, with the outcomes for 64e ee ee
score on the first roll and | the first roll on one axis and the outcomes Soto vee lou
ee ae or) for the second roll on the other axis. At Ose ne eomae
roll. 4 ; , S3t50 ee eo
s = Mark the intersection of each pair of & ldo skcllcoy soa
numbers to show every possible §(208
i yp e.g. 1 then 1 oFeneg
Tip: When there are outcome for the two rolls combined.
NY SE MEG
two completely separate | 6 x 6 = 36 outcomes altogether First roll
parts to an experiment,
like the two dice rolls
here, the total number b Find the probability of rolling an odd number, followed by a 1.
—

of outcomes equals the * Circle the outcomes corresponding to the

number’of eamunanet
eee outcomes for = OT|eeeee pcan
wets event ‘ ‘odd number, then 1’.,
of outcomes for the 2 | Sit teingens * All of the outcomes are equally likely.
other part. eS . " , . , 5 There are 3 outcomes where the event
ian Ve Oe oe happens and 36 outcomes in total.

So, P(odd number, then - Sa

he 36 TD

Sometimes, the outcomes you’re interested in aren’t the numbers themselves,

but are calculated from them... :

Tip: A sample-space Two bags each contain five cards. Bag A contains cards numbered
diagram can also be rt 1, 3, 3, 4 and 5, and bag B contains cards numbered 1, 2, 4, 4 and 5.
drawn as a table — with | A card is selected at random from each bag and the numbers on the
INS SNCOESa VO: two cards are added together to givea total score.
activity along the top
and the outcomes for the — _Use a sample-space diagram to find the probability that the total
other activity down the | score is no more than 6.
left-hand side. E.g:
y Bona = Start by drawing a sample-space diagram
sPiltineseee Se ho 0G showing all the possible total scores.
. : ; ‘ : ; : * This time you need to show the total
lew score for each pair of numbers at
Sl ae ee each intersection. e.g.14+2=3
516 8 8 9 10 * Circle all the scores of 6 or less.
5 x 5 = 25 outcomes altogether

5A
Chapter 2 Probability
_" So now you can use the probability formula. There are 12 outcomes
where the event ‘total score is no more than 6’ happens and
25 outcomes altogether.
So, P(total score is no more than 6) = 1N
i) o

Exercise 1.1

QI One card is selected at random from a standard pack of 52 playing

cards. Find the probability of selecting each of the following: QT tees
cards is split into 4 suits
a) the 7 of diamonds b) the queen of spades — hearts, diamonds,
c) a9 of any suit d) a heart or a diamond spades and clubs.

Q2 The following sample-space diagram represents a dice game where

two dice are rolled and the product of the two scores is calculated:

a) Find the probability that the product is a prime number.

b) Find the probability that the product is less than 7.
c) Find the probability that the product is a multiple of 10.

Q3 A game involves picking a card at random from 10 cards, numbered

1 to 10, and tossing a coin.
a) Draw a sample-space diagram to show all the possible outcomes.
b) Find the probability that the card selected shows an even number
and the coin shows ‘tails’.

Q4 Martha rolls two fair six-sided dice and calculates a score by

subtracting the smaller result from the larger.
a) Draw a sample-space diagram to show all the possible outcomes.
b) Find P(the score is zero).
c) Find P(the score is greater than 5).
d) What is the most likely score? And what is its probability?

Q5 Spinner 1 has five equal sections, labelled 2,3, 5, 7 and 11, and
spinner 2 has five equal sections, labelled 2, 4, 6, 8 and 10.
If each spinner is spun once, find the probability that the number on
spinner 2 is greater than the number on spinner 1.

Chapter 2 Probability
 2. Venn Diagrams ENURESIS SESS

Learning Objectives: Venn diagrams are a really useful tool for solving probability problems.
= Be able to draw Venn Here you'll see how to draw them to represent two or three events,
diagrams for two or and how to use them to find probabilities.
three events.
Use the correct Using Venn diagrams
notation to describe
A Venn diagram shows howa collection of objects is split up into different
the different areas of a
Venn diagram. groups, where everything in a group has something in common.
Use Venn diagrams to = Here, for example, the objects are outcomes and the groups are events.
find probabilities. So the collection of objects, represented by the rectangle, is the
Understand and sample space (S). Inside the rectangle are two circles representing
use complementary two events, A and B.
events.
= The circle A represents all the outcomes
corresponding to event A, and the circle B
represents all the outcomes
corresponding to event B.
s The diagram is usually labelled with the CO
number of outcomes (or the probabilities) represented by each area.
= Since Sis the set of all possible outcomes,
the total probability in S equals 1.

So that’s the basic idea of Venn diagrams. To describe the different areas
of the diagram, you can use set notation.

Tip: If events don’t = The area where the circles overlap

have any outcomes represents all the outcomes corresponding
in common, then the to both event A and event B happening.
circles won't overlap. = This area is called the intersection of A and B
These events are called
— written AM B.
mutually exclusive —
and they're covered in
detail on pages 63-64. »

The shaded area represents all the

outcomes corresponding to either
event A or event B or both happening.
= This area is called the union of A and B
— written A U B.

= The shaded area represents all the outcomes

Tip: This is because if corresponding to event A not happening. S
there are n outcomes = These outcomes make up an event called
altogether in S, andp of
the complement of A — written A’
these are in event A,
(and sometimes read as ‘not A’).
then 1 — p must be in A’.
= Since an event A must either happen or
P(A) = ©,
n
P(A) = 7 not happen, and since P(S) = 1:
which means that
BCE Ap alle P(A)
+ P(A’) =1 = P(A’) =1-—P(A)

Chapter 2 Probability
 The shaded area represents all the outcomes
corresponding to event B happening and event
A not happening. This area is written A’ MB.

S
The shaded area represents all the
outcomes corresponding to event A not
happening and event B not happening.
This area is written A’ M B’.

Here’s an example where the objects are people, and they're divided into
groups based on whether they have certain characteristics in common.

Example 1 ee
There are 30 pupils in a class. 14 of the pupils are girls and 11 of the Tip: Each of the groups
pupils have brown hair. Of the pupils with brown hair, 6 are boys. needs to be based on a
a) Show this information on a Venn diagram. Se aa
person either has or
= OK, so first you need to identify the groups. Let G be the group of doesn’t have. Then if a
girls and BH be the group of pupils with brown hair. person has that
* Draw aVenn diagram to represent the groups G and BH. characteristic, they're
ne included inside the
= Now you need to label it with the numbers circle. Ifthey don’t have
of pupils in each part of the diagram. the characteristic, they're
= You're told that there are 6 boys with brown outside the circle. E.g.
hair, which means there are 6 pupils who all the girls are included
in circle G, and all the
have brown hair and aren't girls.
boys (‘not girls’) are
So label the area BH M G’ with 6 pupils. outside G.
« There are 11 pupils in total with brown hair,
so there are 11 — 6 = 5 girls with brown hair.
So label the area G M BH with 5 pupils. Tip: Remember, ’N
means ‘and’, so the
= Now, there are 14 girls in total, so the number who don’t have
area GM BH is the area
brown hair = 14 —5 = 9. So label the area G N BH’ with 9 pupils. ‘girls and brown hair’.
= Finally, don’t forget the pupils that aren’t in G or BH. There are 30
pupils altogether, so the number of boys who don't have brown hair
= 30-(9 +5 + 6) = 10. So label the area G’ M BH’ with 10 pupils.

b) A pupil is selected at random from the class. Find the probability that
the pupil is a girl who doesn't have brown hair. Tip: All the outcomes
Using the Venn diagram, there are 9 girls who don’t have brown hair, are equally likely, so you
out of the 30 pupils in the class. can use the probability
eo. formula on page 53.
So, P(girl who doesn't have brown hair) = 30 — 10

c) A girl is selected at random from the class. Find the probability

that she has brown hair.
Using the Venn diagram, there are 5 girls You already know that
who have brown hair, out of the 14 girls the pupil is a girl, so
aE you're only looking at
ST ee 5 the outcomes in circle G,
So, P(the girl selected has brown hair) = 14 not the whole class.

Chapter 2 Probability
 rs
The Venn diagram below represents two events, A and B. The numbe
two events .
show how many outcomes correspond to each of the

Use the diagram to find the following probabilities:

a) PAN B)
# This is the probability of event A and event B both happening.
= Use the Venn diagram to find the number of outcomes in A B
Tip: Remember... 2.»
A1)B includes all and the number of outcomes in total.
the outcomes in There are 7 outcomes inAM B,
both A and B. and 15 + 7 + 12 + 6 = 40 outcomes in total. ——— juct add all
settee la BIA Ce) = 7 the numbers.
* So using the formula, P(A B) = 40

b) P(B)
= Use the Venn diagram to work out the number of outcomes in B.
There are 7 + 12 = 19 outcomes in B, and 40 outcomes in total.
19
" So using the formula, P(B) = 7q

ot c) P(A’)
* Use the Venn diagram to work out the number of outcomes in A’.
Tip: Remember...
A’ includes all the There are 12 + 6 = 18 outcomes in A’, and 40 outcomes in total.
outcomes that are
= So using the formula, P(A’) = AO ge
not inA. 40 2 .

d) P(A U B’)
# P(A U B’) is the probability of either event A happening or event B not
happening or both of the above.
Tip: A U B’ includes all
the outcomes that are « Use the Venn diagram to work out the number of outcomes in A U B’.
either in A, There are 15 + 7 = 22 outcomes in A, plus another 6 outcomes that
or not in B, aren't in B. So that’s 22 + 6 = 28 outcomes in A U B’,
or both of the above (i.e. and 40 outcomes in total. ;
in A but not in B).
* So using the formula, P(A U B’) = = os a

You also need to be able to draw and use Venn diagrams for three overlapping
groups (or events). There’s an example on the next page. Instead of showing
the ‘number of objects’ (or outcomes), the numbers show proportions, but the
ideas are exactly the same.

Chapter 2 Probability
A survey was carried out to find out what pets people like.
The proportion who like dogs is 0.6, the proportion who like cats is 0.5,
and the proportion who like gerbils is 0.4. The proportion who like dogs
and cats is 0.4, the proportion who like cats and gerbils is 0.1, and the
proportion who like gerbils and dogs is 0.2. Finally, the proportion who
like all three kinds of animal is 0.1.
a) Draw a Venn diagram to represent this information.

First, identify the groups. Let C be the group ‘likes cats’, D be the group
‘likes dogs’ and G be the group ‘likes gerbils’.
Draw a Venn diagram to represent the groups C, D and G. Tip: Label the diagram
Now you need to label it — this time with the with proportions this
time, because that’s the
proportions for each area. It’s best to start in the information you're given

CR
middle and work outwards... _ in the question.
The proportion who like all 3 animals is 0.1.
So label the area CM DM G with 0.1.
Next do the ‘likes 2 animals’ areas, by subtracting 0.1 Tip: If they like 3
0.1 from each of the given proportions. animals, they’II also be
SoC NDNG’=0.3,CNGND’=0, andDAGNC =0.1. in the ‘likes 2 animals’
Then complete each circle by making sure that the proportions add up bits. So subtracting this
proportion gives us the
to the proportion given in the question for each animal.
proportions of liking just
CND/nNG =0.5 —-(0.3 + 0.1 + 0) = 0.1, 2 of the animals.
DNC’ NG’ =0.6 - (0.3 + 0.1 + 0.1) = 0.1, and
GNC’ND’ =0.4-(0.1 + 0.1 + 0) =0.2.
Finally, subtract all the proportions from 1 to find the proportion who
like none of these animals: C’ M D’M G’ = 1-0.9 = 0.1.

One person who completed the survey is chosen at random. Tip: When you select a
person at random, the
b) Find the probability that this person likes dogs or cats (or both). above proportions equal
the probabilities of that
This is represented by the shaded area. person being in each
The probability of the person being in this
area is equal to the proportion of people
E>
Ae
area of the diagram.

in the area. So you just need to add up oa

the numbers. \Gaa) Tip: You could also do
this by subtracting the
So, P(likes dogs, cats or both) probabilities outside
=0.14+0.3+0+01+0.1+0.1 =O7 circles C and D from 1.

C) Find the probability that this person likes gerbils, but not dogs.
This is represented by the shaded area.
So add up all the probabilities in this area.
So, P(likes gerbils, but not dogs)
=0+0.2 =O2

Chapter 2 Probability
 d) Find the probability that a dog-lover also likes cats.
= This is the proportion of dog-lovers who also like cats.
Tip: You can also D 1 C is represented by the shaded area,
think of this as ‘the
probability that the which forms a proportion of 0.3 + 0.1 = 0.4
person likes cats, given of the whole group of people.
that they like dogs’. But we're only interested in the ‘likes dogs’ circle,
This is an example of a not the whole group, so that means we need to
conditional probability. divide by 0.6.
Conditional probabilities
are covered in detail on So, P(dog-lover also likes cats) = az = 2 = 5
pages 65-67.

Exercise 2.1

QI For events A and B, P(A) = 0.4, P(B) = 0.5 and P(AN B) = 0.15.
Q1 Hint: You're given
the information as a) Draw a Venn diagram to represent events A and B.
probabilities, rather than b) What is P(A’ MB’)?
numbers of outcomes
corresponding to Q2 A sixth form college has 144 students — 46 of the students study
each event. So label maths, 38 study physics and 19 study both.
your diagram with
probabilities. a) Represent the information given above using a Venn diagram.
b) Find the probability that a randomly selected student from the
college studies at least one of either maths or physics.
c) Given that a student studies maths, what is the probability that
they also study physics?

Q3 Use the Venn diagram to find the following probabilities:

Q3 Hint: If two circles a) P(LMM) b) P(LAN)
don’t overlap, it means
the events can’t both €} PCC aeN) dy “PUL May N) GOs)
happen. e) PAP) f) P(M’)

Q4 Two hundred people were asked which of the countries Spain,

France and Germany they have visited.
The results are shown in the diagram. is
Find the probability that a randomly selected bopetin
person has been to: leoJ
a) none of the three countries
b) Germany, given that they have been to France
c) Spain, but not France

Q5 1000 football supporters were asked if they go to home league

Q5 Hint: It doesn’t matches, away league matches, or cup matches. 560 go to home
tell you to draw a Venn matches, 420 go to away matches, and 120 go to cup matches. 240
diagram, but that’s the
easiest way to answer
go to home and away matches, 80 go to home and cup matches, and
the question. 60 go to away and cup matches. 40 go to all 3 types of match.
Find the probability that a randomly selected supporter goes to:
a) exactly two types of match b) at least one type of match

Chapter 2 Probability
3. Laws of Probability
There are two main probability laws you need to know — the addition law Learning Objectives:
and the product law (or multiplication law). You'll see how to use these laws
Be able to use the
to find probabilities, and how you can adapt them in different circumstances.
addition law to find
probabilities.
The addition law Be able to recognise
For two events, A and B, there’s'a nice formula linking the union of A and B mutually exclusive
and the intersection of A and B — the addition law: events.
Be able to use the
P(A U B) = P(A) + P(B) — P(AN B) product law to find
probabilities.
Be able to recognise
You can see why this is true, using Venn diagrams. independent events.
" Be able to understand
and use tree diagrams.

Tip: If you didn’t

P(A) a8 P(B) rs P(A U B) + P(A MB) subtract P(A 9 B) from
P(A) + P(B), you’d be
counting it twice —
So to get P(A U B) on its own, you need to subtract P(A M B) from P(A) + P(B). ce re oe
The addition law is really useful for finding missing probabilities — as long as from B.
you know three of the values in the formula, you can rearrange the formula to
find the remaining probability. ipawOlameniel teen
the addition law called
Example i —_———<—$<$—<—<$<—<—————— the ‘sum law’. It'll be on
the exam formula sheet,
For two events A and B, P(A U B) = 0.75, P(A) = 0.45 and P(B’) = 0.4. but you should learn it
a) Find P(A nm B). anyway. |
= To use the formula, you need to know P(A), P(B) and P(AUB).
You’re missing P(B), so start by finding that.
P(B) = 1 — P(B’) = 1 —0.4 = 0.6.
= Now rearrange the addition law formula to make P(A /N B) the subject.
P(A U B) = P(A) + P(B) — P(A MB) = P(A N B) = P(A) + P(B) — P(A U B)
= And substituting in the probabilities gives:
P(A NB) = 0.45 +0.6-0.75 = 03
b) Find P(A’ - B’). Tip: Watch out for when
: aeis the complement ofA U B. 3B you can use the formula
= AB’ P(A) = 1 — P(A). |
s So P(A 0B’) =1-P(AU B)=1-0.75 = O25
AOE AUB
c) Find P(A NB’). | | THALAR EI AG
= Event A is made up of areas A B and AN B’ — see Tip on the right.
= So P(ANB’) = P(A) — P(A 0 B) = 0.45 — 0.3 = 0.15

d) Find P(A’ U B).

= It’s easiest to do this by drawing a Venn diagram.
= Use the probabilities you’ve worked out above. 0.25

=» So P(N UB) =0.3 + 0.25 + 0.3 = 0.85 NUB

Chapter 2 Probability
 On any given day, the probability that Jason eats an apple is 0.6, the
probability that he eats a banana is 0.3, and the probability that he eats
both an apple and a banana is 0.2.
a) Find the probability that he eats an apple or a banana (or both).
« Let A be the event ‘eats an apple’ and B be the event ‘eats a banana’.
* You want to find P(A U B), so use the addition law:
P(A U B) = P(A)+ P(B) — P(A.G.B) = 0.6 +0.3—-02—07
= So P(he eats an apple or a banana, or both) = 0.7
b) Find the probability that he either doesn’t eat an apple,
or doesn’t eat a banana.
* You want to find P(A’ U B’). You can either do this using a Venn
diagram, or you can use the addition law by replacing A with A’ and
B with B’. Like this: :
Tip: A’ U B’ is also the » P(A’ UB’) = P(A’) + P(B’) — P(A’ 9 B’)
complement of A M B, so = [TS PAS Te PB) fl SPA)
you could work out: =(1-0.6) + (1 —0.3)-—(1 —0.7) = 0.4+ 0.7 -0.3 =0.8
P(A’ U B’) = 1 — P(A B) # So P(he either doesn’t eat an apple, or doesn’t eat a banana) = 0.8
=1-0.2
= 0.8

Exercise 3.1
Hint: Remember, you Ql HPA) =0.3,.P(8) = 0.5 and PYA 1) By) =0.15, tind:
can always draw a Venn a) P(A’) b) PRA. B) Che PAs
diagram to help you. |

Q2- -H P(A) =.0.36, P(B) = 044and P(A B)= 0:27, find:

Q2 d) Hint: You have a) P(B’) b) P(A UB) c) PAB) dy” PIA UB)
all the information you
need to use the addition
Q3 A car is selected at random from a car park. The probability of the
law, replacing B with B’.
car being blue is 0.25 and the probability of it being an estate is
0.15. The probability of the car being a blue estate is 0.08.
a) What is the probability of the car not being blue?
b) What is the probability of the car being blue or being an estate?
c) What is the probability of the car being neither blue nor an estate?

Q4_ If P(X UY) = 0.77, P(X) = 0.43 and P(Y) = 0.56, find:
a) P(Y’) b) P(X NY) Cc) PX iY) dP”)

Q5_ If P(C’ UD) = 0.65, P(C) = 0.53 and P(D) = 0.44, find:
a) PCD) b) PIC aD) Cc) PC Gy) d) PCD)

Q6 The probability that a student has read ‘To Kill a Mockingbird’ is

0.62. The probability that a student hasn’t read ‘Animal Farm’ is
0.66. The probability that a student has read at least one of these two
books is 0.79. Find:
a) The probability that a student has read both the books.
b) The probability that a student has read ‘Animal Farm’ but hasn't
read ‘To Kill a Mockingbird’.
c) The probability that a student has read neither of the books.
Chapter 2 Probability
 Mutually exclusive events
Events can happen at the same time when they have one or more outcomes
in common. For example, the events ‘I roll a 3’ and ‘I roll an odd number’,
both happen if the outcome of my dice roll is a ‘3’. Events which have
* no outcomes in common, can’t happen at the same time. These events are
called mutually exclusive (or just ‘exclusive’.
Tip: A Venn diagram
= IfA and B are mutually exclusive events, then P(A 2 B) = 0. for mutually exclusive
= And since the intersection is zero, a Venn diagram would show the events A and B might
events as non-overlapping circles. look like this.

coe
S
We defined the addition law on page 61 as P(A U B) = P(A) + P(B) — P(A B).
When A and B are mutually exclusive, we can substitute P(A B) = 0,
to give a slightly simpler version.
For two events, A and B,
where A and B are mutually exclusive:
Tip: With n exclusive
And you can write a general form of this for n exclusive events.
events, AU, aX,
For exclusive events PLS ese: none of the events can
happen at the same time
P(A, UA, U... UA) = P(A,) + P(A,) +... + P(A) as any of the others.
In other words, only
one out of A,, ..., A, can
eae. happen at a time.
Example

A card is selected at random from a standard pack of 52 cards. Find the

probability that the card is either a picture card (a Jack, Queen or King),
or the 7, 8 or 9 of clubs.

= Start by defining the two events. Let A be the event ‘select a picture
card’ and B be the event ‘select the 7, 8 or 9 of clubs’.
= You want to find the probability ofA or B, P(A U B). The card can’t be
both a picture card and the 7, 8, or 9 of clubs, so A and B are mutually
exclusive, which means that P(A U B) = P(A) + P(B).
=» Using the formula for equally likely outcomes:
P(A)=
15 and P(B)=
3
12 outcomes where event 3 outcomes where event
happens out of a total of 52 happens out of a total of 52

= So the probability ofA or B is PA U B) = P(A) + P(B)

See)
1
eben ans
= This means P(card is either a picture card or the 7, 8 or 9 of clubs) = +

‘To show whether or not events A and B are mutually exclusive, you just
o
need to show whether the intersection of A and Bis zero or non-zer
—j.e. whether P(A MB) = 0 or P(A N B) # 0.

Chapter 2 Probability
 a) For two events, A and B, P(A) = 0.38, P(B) = 0.24 and P(A UB) = 0.6.
Tip: You could also Show whether events A and B are mutually exclusive.
show that
= The easiest way to do this is to use the addition law to find P(A / B).
P(A U B) # P(A) + P(B)
P(A U B) = P(A) + P(B) — P(A B)
— that’s the same = P(A MB) = P(A) + P(B) — P(A U B) = 0.38 + 0.24 — 0.6 = 0.02
as showing that
P(A M B) # O. = So, P(A B) #0, which means that A and B are not mutually exclusive.

b) For two events, A and B, P(A) = 0.75 and P(A M B’) = 0.75.
Tip: Drawing a Venn Show whether events A and B are mutually exclusive.
diagram can be helpful = This one looks abit trickier, but you just need to think about the
to see areas.
different areas that make up event A.
P(A) = P(A. B) + P(A mB’)
= P(A. B)= P(A) — PIA B) = 0.75 —0.75— 0
So, P(A M B) = 0, which means that A and B are mutually exclusive.

Exercise 3.2
Q1__If X and Y are mutually exclusive events,
with P(X) = 0.48 and P(Y) = 0.37, find:
a) P(XNY) b) P(X UY) CPX TY)

Q2_ P(L) = 0.28, P(M) = 0.42 and P(N) = 0.33. If the pairs
of events (L and M) and (L and N) are mutually exclusive,
and P(M MN) = 0.16, find:
a) P(LUM) b) P(LUN) c) P(MUN)~ d) P(ILMA MAN)
e) Draw and label a Venn diagram to show events L, M and N.

Q3 Dave is planning his evening. The probabilities that he will go

bowling, to the cinema and out for dinner are 0.17, 0.43 and 0.22
respectively. Given that he only has time to do one activity, find:
a) The probability that he either goes bowling or to the cinema.
b) The probability that he doesn’t do any of the 3 activities.

Q4 For events A, B and C, P(A) = 0.28, P(B) = 0.66, P(C) = 0.49,

P(A U B) = 0.86, P(A U C) = 0.77 and P(B U C) = 0.92.
Find each of the probabilities below and say whether each pair of
events is mutually exclusive.
a) P(AMB) Db). PIA 1.C) GC). PIB. ©)

Q5__ For events C and D, P(C’) = 0.6, P(D) = 0.25 and (CN D’) = 0.4.
a) Show that C and D are mutually exclusive. b) Find.P(C U D)

Qé A box contains 50 biscuits. Of the biscuits, 20 are chocolate-coated

and the rest are plain. Half of all the biscuits are in wrappers.
One biscuit is selected at random from the box.
If P is the event ‘the biscuit is plain’, and W is the event ‘the biscuit
is
In a wrapper’, show that the events P and W are not exclusive.
Chapter 2 Probability
 The product law — conditional probability
A probability is conditional if it depends on what has already happened. Tip: Events A and B
The probability that an event B happens, given that an event A has already might be occurring at
_ happened, is called the conditional probability of ‘B given A’, written P(B| A). the same time — i.e.
P(B |A) is the probability
You can work out the probability of B given A, using this formula: of B occurring, given
that A occurs.

P(B| A) = an
Tip: The probability of
A given Bis:
Here’s an explanation of where this formula comes from... P(A 1 B)
P(A|B) = P(B)
Event A has happened and for BJA to happen, B will also happen.
= If you know that A has already happened,
then the only remaining possible outcomes c
must be the ones corresponding to A: 6
And the only remaining possible outcomes ee
corresponding to B also happening a
must be the ones in A 1 B.
ei
Using the probability formula
(and assuming event A has already happened):
number of possible outcomes corresponding to B
PUB A= total number of possible outcomes
— number
of outcomes in AB
number of outcomes in A

So if n(A A B) = number of outcomes in A © B, and

n(A) = number of outcomes in A:
mn(As00B)

Now, if we divide the top and bottom of the fraction by n(S)

(the total number of outcomes inS$), “its value doesn’t change,
but we can write it in a different way:
n(A 1) B)
n(A 1 B) = als) = ea) ~«_ the formula
P(B
|A) = n(A n(A P(A in the box
- WS) above

And if we rearrange this formula for conditional probabilities, we get our

second important probability law, called the product law. For events A and B:
Tip: The product law
is given in this form
on your exam formula
P(A 2 B) = P(A)P(B| A) sheet. But you can
easily rearrange it into
= So this law says that to find the probability that A and B both happen, the form given above
you multiply the probability of A by the probability of B given that A to find conditional
has happened. probabilities.
B:
= Or you can write this the other way around, by swapping A and
P(A 2 B) = P(B)P(A|B)..
Chapter 2 Probability
 For events A and B: P(A) = 0.6, P(B) = 0.5, PAM B) = 0.3, P(B’| A) = 0.5.

a) Find P(A|B).
Using the formula for conditional probability:
AB i 8-3 = 0.6
P(A|B) = Cm

b) Find P(A mB’).

Tip: Just replace B with Using the product law: P(A B’) = P(A)P(B’ |A)
B’ in the product law. ORO Os
= 03

The Venn diagram below represents two events, C and D. The numbers
of equally likely outcomes corresponding to the events are shown.

a) Find P(C|D).

« Here you’re given the numbers of outcomes, so you can find the
Tip: Doing it this way probability by using the original probability formula from page 53.
is a bit simpler than
using the conditional = Once event D has happened, there are 5 possible
probability formula. outcomes corresponding to event C happening «—____5 outcorhes
But if you did use the
formula, you’d work out:
and 12 possible remaining outcomes in total. inCnD

P(C MD)=mp>> and number of outcomes corresponding toC 5

P(D) = ta *S0rBUC| Di = tral rumbene oessibacutetrace
And so, P(C| D) =
CCID )i. oy5 2 5.
PD) 2 12 aS— Find P(D|C’).
“= Once event C hasn’t happened, there are 7 possible
outcomes corresponding to event D happening «—____7 outcomes
and 12 possible remaining outcomes in total. inDOAC

= So, (D/C) = number of outcomes corresponding toD _@

total number of possible outcomes 12

Chapter 2 Probability
Vikram either walks or runs to the bus stop. The probability that he walks
is 0.4. The probability that he catches the bus is 0.54. If he walks to the
bus stop, the probability that he catches the bus is 0.3.

a) Draw a Venn diagram representing the events W, ‘Vikram walks to the

bus stop’, and C, ‘Vikram catches the bus’.
* Okay, this one looks a bit scary, but it’ll make more sense when you
write down the probabiliti k :
P ia aa ie ncn Tip: The probability that
P(W) = 0.4, P(C) = 0.54 and P(C|W) = 0.3 Vikram catches the bus,
given that he walks to
= To draw the Venn diagram, you need to find the intersection C 9 W. the bus stop, is 0.3.
And you can do that using the product law:
Peri Wy = PIC
|W)PW) = 0:3°« 0.4 = 0.12
= Now you can draw the Venn diagram:

P(W) — P(C NW)

ik me P(C) — PPC N W)

b Find the probability that Vikram catches the bus, given that he runs to
—

the bus stop.

= This is the probability, P(>C |W’).
+3 P W’ Tip: He either walks or
= Using the formula for conditional probability: P(C |W’) = Wy ) runs to the bus stop, so
= P(CAW)=0.42 and P(W’)=1-0.4=0.6 ene
pe = “ED
. So,P(C| Ww) (Ca. = 82
42 = 07

= So P(Vikram catches the bus, given that he runs to the bus stop) = 0.7

Qi If P(G) = 0.7, P(H) = 0.63 and P(G 9 H) = 0.24, find:

a) P(G |H) b) P(H|G)
Q2 P(A) = 0.68, P(B’) = 0.44, P(C) = 0.44, P(A MB) = 0.34,
P(A NC) = 0.16 and P(BN C’) = 0.49. Find:
a) P(B|A) b) P(A|C) c) P(C’|B)

Q3 Ina group of eleven footballers, five are over 6 feet tall.

Two of the three players who can play in goal are over 6 feet tall.
One of the players is selected at random.
a) If the player is over 6 feet tall, what is the probability that they can
play in goal?
b) If the player can play in goal, what is the probability that they are
over 6 feet tall?
Chapter 2 Probability
 Q4 The Venn diagram below shows the numbers of students studying
Maths, English and Art, from a group of 100 students.
One of the students is selected at random.

3 If the student is studying Art, what is the probability that they are
also studying Maths?
S If the student is studying English and Maths, what is the
probability that they are also studying Art?
os If the student is not studying Maths, what is the probability that
they are studying English? .
d Find P(A|E’).
e SSFind P(M |AN E).
Q5_ Given that P(X) = 0.44, PCY’) =0-72,, PZy-=0'61, PIXIY= 0.75,
P(Z|X) = 0.25, P(Y NZ’) = 0.2 and P(X NY |Z) = Z, find:
a) P(Y) by PX) cy PXNZD
d) PLY |Z4 e) PIXNYNZ)

Independent events
If the probability of an event B happening doesn’t depend on whether an
event A has happened or not, events A and B are independent.

= For example, if a dice is rolled twice, the events A = ‘first roll is a 4’ and
B = ‘second roll is a 4’, are independent, because the number rolled on
the second roll doesn’t depend on the number rolled on the first roll.
Tip: If the first card isn’t = Or, suppose a card is selected at random from a pack of cards, °
replaced, then A and B then replaced, then a second card is selected at random. The events
are not independent A = ‘first card is a 7’ and B = ‘second card is a 7’, are independent
(B is conditional on A). because P(B) is unaffected by what was selected on the first pick.
IfA happens, P(B) = 2,
but if A doesn’t happen,
4 If A and B are independent, then P(B) is the same, whether A happens or not.
P(B)i= By: And that means you have the following results:

= P(B |A) = P(B |A’) = P(B). Similarly, P(A | B) = P(A | B’) = P(A).

= The conditional probability formula becomes: P(B |A) = P(B) = P(AP(A)B)

= The product law becomes: P(A B) = P(A)P(B |A) = P(A)P(B)

So, for two events, A and B, where A and B are independent:

P(A M B) = P(A)P(B)

Chapter 2 Probability
 V and W are independent events, where P(V) = 0.2 and P(W) = 0.6.
a) Find P(V MW).
Just put these probabilities into the product law for independent events:
P(V 1 W) = P(V)P(W) = 0.2 x 0.6 = a

b) Find P(V UW).

You know the probability of each event and the probability of the
intersection, so you can use the addition law to find the union:
P(V UW) = P(V) + P(W) — P(V NW) = 0.2 + 0.6 — 0.12 = 0.68

To show that events A and B are independent, you just need to show that one
of the following statements is true:
= P(B|A) = P(B) [or P(A|B) = P(A)] Tip: If one of these is
true, then the other one
m P(A) x P(B) = P(A NM B) will also be true.

A scientist is investigating the likelihood that a person will catch two

infectious diseases, after being exposed to one and then the other.
The probability of catching the first disease is 0.25, the probability of
catching the second disease is 0.5, and the probability of catching both
diseases is 0.2.
Show that the events ‘catch first disease’ and ‘catch second disease’
are not independent.
Tip: You could also
» Let A = ‘catch first disease’ and B = ‘catch second disease’. show that:
= There are several different ways to show the result, but it’s probably P(B|A) =
easiest to compare P(A) x P(B) with P(A M B). P(A OB) er
» P(A) = 0.25, P(B) = 0.5 and P(A N B) = 0.2 P(A) 0255 us
Pie PB) = 0225 0S] O125 00.2 += p(B = 05)

So, since P(A) x P(B) # P(A 12 B), the events ‘catch first disease’ and or P(A|B)=
‘catch second disease’ are not independent. P(A 1B). SFOs
ar oe 0.4

22 PUA) = O25.

For events A and B, P(A) = 0.4, P(B |A) = 0.25 and P(A’ M B) = 0.2.
a) Find: (i) P(ANMB), (ii) P(A’), |A),
(iii) P(B’
(iv) (B| A’), (v) P(B), | B).
(vi) P(A

(i) Using the product law: P(A M B) = P(A)P(B |A) = 0.4 x 0.25 = 0.1
(ii) P(A’) = 1-— P(A) = 1 -0.4 = 0.6

Chapter 2 Probability
 Now at this stage, things are starting to get a bit trickier. It'll help a lot if
you draw a Venn diagram showing what you know so far.
* You know from the question that:
P(A) = 0.4, P(B| A) = 0.25 and P(A’ M B) = 0.2
# And you've found:
P(A 1 B) = 0.1 and P(A’) = 0.6 P(A) — P(A NB) Ee 0}

(iii) Using the conditional probability formula: PUA’ 1B)

‘ PBA”)
Leeman eam
You can see from the Venn diagram that P(B’ M A) = 0.3, so:
Tip: Another way of . P(BAA) ores
getting P(B’|A) is to do P(B’ |.A) = “PAys caida 0.75
1 —P(B|A), since BJA
a sean Cl | (iv) Using the conditional probability formula:
P(B AD 3302.

Tip: You can also find | (v) You can see from the Venn diagram that:
P(B) by doing: | P(B) = (BNA) +P(BNA) =0.1 +0.2 = 03
P(B|A)P(A) + P(B|A’)P(A’) |
That's because... And finally...
P(B |A)P(A) = P(B iA),
=PBA) ,
P(BIA)P(X) (vi): P(A | B) B 8-8
And... ;
P(BMA) + P(BNA’) = P(B)
b) Say whether or not A and Bare independent.
Again, there are different ways you can do this. For example, if you
Tip: Or you can show: compare the values of P(B|A) and P(B), you can see that:
P(A| B) # P(A), or
P(A) x P(B) # P(A B) P(B|A) = 0.25 # P(B) = 0.3, so A and B are not independent.

Exercise 3.4

Q1_ If X and are independent events, with P(X) = 0.62 and P(Y) = 0.32,
calculate P(X 1 Y).

Q2 P(ANB) = 0.45 and P(B’) = 0.25. If A and B are independent events,

what is P(A)2

oar ~Q3__ Events M and N are independent, with P(M) = 0.4 and P(N) = 0.7.
Q3 Hint: IfM and N Calculate the following probabilities:
are independent, then
M and N’ are also a) PMN) b) P(MUN) c) P(IMAN’)
independent.
vernon erm Q4 Acard is picked at random from a standard pack of 52 cards.
The card is replaced and the pack is shuffled, before a second card is
picked at random.
a) What is the probability that both cards picked are ‘hearts’?
b) Find the probability that the ‘ace of hearts’ is chosen both times.

Chapter 2 Probability
 Q5 For events A, B and C: P(A) = o P(B) = i, P(C) = are

C)= -and PIBNO =.

P(ANB)=4, PAN
Show whether or not each of the pairs of events, (A and B), (A and C)
and (B and C) are independent.

OGaex , Y and Z are independent events, with P(X) = 0.84, P(Y) = 0.68 and
P
(Z) = 0.48. Find the following probabilities:
Jer (oye wi Z) ey PY |Z) “dy PZ" IN"). ~e) PY TX’)

Q7 Jess, Keisha and Lucy go shopping independently. : The probabilities :

7 Hint: Th duct
that they will buy a DVD are 0.66, 0.5 and 0.3 respectively. e ‘araereten
a) What is the probability that all three of them buy a DVD? events applies for any
b) What is the probability that at least two of them buy a DVD? TO

Tree diagrams
Tree diagrams show probabilities for sequences of two or more events.
Here’s a tree diagram representing two trials. There are two possible
results for the first trial — events A and A’, and there are two possible
results for the second trial — events B and B’.

‘End’ probabilities
1st trial 2nd trial
P(BJA) -~B P(AMB)
Tip: The ‘end’
Probabilities P(A) A <i
P(B’|A) ~B’ P(A MB’) probabilities are the
on each set x probabilities of the
P(BIA)-B PIYOB
of branches P(A) tet ( ) different sequences of
add up to 1. A, ee

P(B’|A) BY P(A’ 7B’) Answers to questions

are often found by
adding the relevant ‘end’
= Each ‘chunk’ of the diagram represents one trial. probabilities.
= Each branch of a ‘chunk’ is a possible result of the trial.
= To find the probability of a sequence of events, you
multiply along the branches representing those events.
= The total of the ‘end’ probabilities is always 1.

Tree diagrams for independent events

Tree diagrams for independent events look a bit simpler than the one above.

= Events A and B have no effect on each BB

other, so P(B|A) = P(B), and so on. P(A) A
ae ne E
= So for the diagram above, the second set
of branches would look slightly different: P(A) A’ ce
P(B’) ~ B’

Chapter 2 Probability
 eT Cs
purple.
A bag contains 10 balls, 6 of which are red and 4 of which are
One ball is selected from the bag at random, then replaced. A second
ball is then selected at random.

a) Draw a tree diagram to show this information.

There are two trials — ‘1st ball selection’ and ‘2nd ball selection’.
Each trial has two possible results — ‘red’ and ‘purple’.
The probability of selecting a red ball on each pick is 0.6 and the
probability of selecting a purple ball on each pick is 0.4.
Tip: Here, the So you can draw a tree diagram like this:
selection is done ‘with
replacement’ — so the
result of the second R, = ‘first ball is red’
selection is independent R, = ‘second ball is red’
0.6 R1 0 p
of the result of the first. a 5 P, = “first ball is purple’

/) 04 P,
0.6 R, P, = ‘second ball is purple’
Tip: You can answer 0.6+0.4=1
questions like b) and 0.4 P
c) without using a tree
diagram, but very often
if you draw one, it helps
to make things clearer. a— Find the probability that both balls are red.
There is 1 ‘path’ along the branches that gives the result ‘red and red’.
Tip: This is the same as Multiply along the branches R, and R, to give:
using the product rule P(R, A R,) = 0.6 x 0.6 = 0.36
for independent events
R, and R,.
Find the probability that one ball is red and the other is purple.
There are 2 ‘paths’ along the branches that give the result
‘red and purple’ — (R, and P.) and (P, and R,).
Tip: This is the same
as using the product Multiply along these pairs of branches to give:
rule for independent P(R, MP.) = 0.6 x 0.4=0.24 and P(P, MR) =0.4 x 0.6 = 0.24
events twice, followed
by the addition rule Now, you want to find the probability of (R, and P,) or (P, and R,),
for mutually exclusive so you add these two probabilities together:
events. P(1 red and 1 purple) = P(R, 1 P,) + P(P, A R,) = 0.24 + 0.24 = 0.48

Tree diagrams for dependent events

When you’re dealing with events that depend on each other, the
probabilities on the second set of branches are conditional on those on
the first set of branches (see the diagram on the previous page).
This means you have to be a bit more careful when you're labelling
your tree diagram.

Chapter 2 Probability
Example 1
A box of 6 biscuits contains 5 chocolate biscuits and 1 lemon biscuit.
George takes out a biscuit at random and eats it. He then takes out
another biscuit at random.

a) Drawatree diagram to show this information.

« There are two trials — ‘1st biscuit selection’ and ‘2nd biscuit selection’.
Let C, = ‘biscuit i is chocolate’ and L, = ‘biscuit i is lemon’, for i = 1, 2.

» The probability of selecting a chocolate biscuit on the first pick is 2 and ee ey

is done ‘without
1 6
the probability of selecting a lemon biscuit on the first pick is é. replacement’ — so the
result of the second
» The probabilities for the second biscuit are conditional: selection is conditional
° If the first pick is chocolate, then: P(choc) = + — i.e. P(C, |C,) = =, on the result of the first.
and P(lemon) = ‘—i.e. P(L, |C,) = 1 | oh Sal av ae
e If the first pick is lemon, then there areno lemon biscuits left, so:
P(choc) = 1 — i.e. P(C, | L,) = 1 and P(lemon) = 0 — i.e. P(L, [|L)=0

= So you can drawa tree diagram like this:

Tip: If George took out

2 PIC ce a third biscuit, you'd get
5 ee a third set of branches,
6 Pt, IG George can’t pick a looking like this:
5 2 second lemon biscuit, 3
11 P(C, SI |L,) so y you only y need a a 4
Stee
6 L eee oe for chocolate. ; a +
< ay it 1 e

ee ‘
b) Find the probability that George takes out two chocolate biscuits.
= There is 1 ‘path’ along the branches that gives this result.
= Multiply along the branches a and C, to give:
P(C, NC) 2= PIC,)P(C,(C,)= Sy2 x 3 202
S053

= So P(George takes out two chocolate biscuits) = 5

c) Find the probability that the second biscuit he takes is chocolate.

= There are 2 ‘paths’ along the branches that give the result ‘second
biscuit is chocolate’ — (C, and C,) and (L, and C,). eae se
é : ip: Another way to fin
= You've ane found P(C, 9 C,), so find P(L, M C,) in the same way: pps Discuitic ee
1 iiss to fin P(2nd biscuit
to find
= Pct) =¢ <i
eer | 6 is lemon) and subtract
= Now add the ites x:the two ‘paths’ together:
et this probability from
1. Sometimes finding
P(2nd biscuit is chocolate) = P(C, N C,) + P(L, 1 C,) = +
\|
WIN oO|— On the complement and
subtracting it from 1 is
an easier way to find a
probability. /

Chapter 2 Probability
 Horace is either late for school or on time for school, and when he gets to
school he is either shouted at or not shouted at. The probability that he’s
late for school is 0.4. If he’s late, the probability that he’s shouted at is
0.7. If he’s on time, the probability that he’s shouted at is 0.2.
Given that Horace is shouted at, what is the probability that he was late?

=
It’s best to take complicated questions like this step by step.
You're given information about two events:
Let L = ‘Horace is late’ and let § = ‘Horace is shouted at’.
* Start by writing down the probability you want to find: that’s P(L |S).
P(LNS
Using the conditional probability formula: P(L |S) = -
Tip: Be careful with
questions like this... * So you need to find P(L 4S) and P(S) — and the easiest way is by
The question tells you drawing a tree diagram using the information in the question:
the probability of S
conditional on L (and L’).
But you need to think of
the situation the ‘other eh
way round’ — with L
0.4 gr P(S"|L) = 1 - P(S|L)
conditional on S. Ss
So don’t just rush in. PL)=1-P)—>06 ~,, 92
0.8 S’ PSS 1 =P(S 1)

Tip: In general, if B
depends on A, then A
= Using the tree diagram:
depends on B. P(L M S) = P(L)P(S|L) = 0.4 x 0.7 = 0.28
Here, S depends on L, P(S) = P(LM S) + P(L’ MS) = 0.28 + P(L)(S|L)
so L depends on S. = 0.28+0.6x0.2
= 0.4
PLAS — 928 _ 97
* So,P(L|S) =“)
This means P(Horace was late, given he is shouted at)= 0.7. >

For events M and N: P(M) = 0.2, P(N |M) = 0.4 and P(N’ |M’)
= 0.7.
'a) Draw a tree diagram representing events M and N.

= You need two sets of branches — one for event M and

one for event N.
Since you're told the probability of M, but not that of N, show
M on the
first set and N on the second set.
Each pair of branches will either show M and its compleme
nt M’,
or N and its complement N’.

74s Chapter 2 Probability

 Draw the tree diagram and write on as much as you know so far:

P(N|M) 9.4 N

02,.-—M ae N’

N
oe
M’

P(N’|M’) 0.7 yp

* Now use this information to find the remaining probabilities:

P(M) = 1— P(M) = 1-0.2 = 0.8 . Tip: This is just using the
P(N’ | M) = 1 — P(N [Mj =1.=.0.4 = 0.6 fact that the probabilities
Rr = Pt ee - ( on each set of branches
P(N |M’) = 1 - P(N’|M’) = 1 -0.7 = 0.3 add up to1—see
p71. |

= Finally, add these probabilities to the tree diagram:

04-7 N
M
es 0.6 nv
N
oy
0.7 N’

b) Find P(N).

= P(M)P(N|M)
P(N) = P(MA.N) + P(M’ 9 N) MeO + P(M)P(N |M’) Tipe Thete ae spain
ee OF giving N, sofind the
a, i : probability of each, then
-Cs add them together.

c) Find P(M’ | N’).

* Remember, If N’ is conditional on M’, then M’ is conditional on N’.

= 7 Fe
Using the conditional probability formula: N’)= teeP(N)
P(M’ |iNes
iSabdigu malas Md
» Using the probabilities you found above:
P(N’) = 1 — P(N) = 0.68
P(M’ 7.N’) = P(M’)P(N’ |M’) = 0.8 x 0.7 = 0.56

" So,P(M’Pa | N) = eee 056 1ns

SOT? = O88 = O0|Or
oes — N

Chapter 2 Probability
 Exercise 3.5
Ql The probability that Jake will win two consecutive erie
darts matches is shown on the tree diagram. ist”
a) Explain whether the events ‘wins 1st match’. ei (ote
and ‘wins 2nd match’ are independen t? rae Win
b) Find the probability that Jake will win: <a
(i) both matches (ii) at least one match

Q2 A game involves rollingafair, six-sided dice and tossing a fair coin.

A player wins if they roll a ‘6’ and the coin shows ‘tails’. ete
a) Complete the tree diagram by showing - ee
the probability on each branch. Ga Tails
b) Find the probability that a person oe ges
wins the game. we,

Q3 The probability that a randomly selected Year 13 student has passed

their driving test is 0.3. The probability that they intend to go to
university, given that they have passed their driving test, is 0.75.
a) Assuming that ‘passed driving test’ and ‘intend to go to university /

are independent, draw a tree diagram to show this information.

b) Find the probability that a randomly selected student hasn’t
passed their driving test and does not intend to go to university.
c) What is the probability that a student picked at random does not
intend to go to university?

Q4 A restaurant has found that if a diner orders a roast dinner, the

probability that they order apple pie for pudding is 0.72. If they
order a different main course, they order apple pie with probability
0.33. The probability that a diner orders a roast dinner is 0.56.
By drawing a tree diagram, find the probability that a randomly
selected diner will order apple pie for pudding.

Q5 A group of people were asked about their mobile phones. 62% of

them own a smartphone and 53.9% of them have a contract costing
more than £25 a month. Of the people with a smartphone, 29%
have a contract costing £25 a month or less.
Use a tree diagram to find the probability that a person from the
group owns a smartphone, given that their contract costs more than
£25 a month.

Q6 A game involves picking two balls at random from a bag containing

12 balls — 5 red, 4 yellow and 3 green — where the first ball isn’t
replaced. A player wins if they pick two balls of the same colour.
a) Draw a tree diagram to show the possible results of each pick.
b) Find the probability that a player wins the game.
c) The game is changed so that the first ball is replaced before the
second one is picked. Is a player more or less likely to win now?

Chapter 2 Probability
Review Exercise — Chapter 2
ITD

Q1_ Afair, six-sided dice and a fair coin are thrown and a score is recorded.
If a head is thrown, the score is double the number on the dice.
If a tail is thrown, the score is the number on the dice plus 4.
a) Draw a sample-space diagram to represent all the possible outcomes.
b) What is the probability of scoring more than 5?
c) Ifa tail is thrown, what is the probability that the score is an even number?

Q2 Half the students in a sixth-form college eat sausages for dinner and 20% eat chips.
2% eat sausages and chips together.
a) Draw a Venn diagram to show this information.
b) Find the percentage of students who eat chips but not sausages.
c) Find the percentage of students who eat either chips or sausages but not both.

Q3 Arabella rolls two fair, six-sided dice and calculates her score by adding
the two results together.
a) What is the probability that her score is a prime number?
b) What is the probability that her score is a square number?
Let P be the event ‘Arabella’s score is a prime number’ and S be the event
‘Arabella’s score is a square number’.
c) Explain whether or not the events P and S are mutually exclusive.
d) Find. P(P U.S).
Arabella carries out the experiment twice. Let S, be the event ‘score from first pair of rolls is
a square number’ and S, be the event ‘score from second pair of rolls is a square number’.
e) Explain whether or not the events S, and S, are independent.
Dey Find P(S, 1 5,).

Q4 Aschool orchestra is made up of pupils in either the upper or the lower school.
40% of the musicians are boys. Of the boys, 30% are in the upper school.
Of the girls in the orchestra, 50% are in the upper school.
a) Draw a tree diagram to show this information.
b) A musician is chosen at random. Find the probability that the musician is:
(i) in the lower school, given that they are a boy
(ii) a girl in the lower school
(iii) in the upper school

Chapter 2 Probability
 Exam-Style Questions — Chapter 2

A soap company asked 120 people about the types of soap (from Brands A, B and C) they
bought. Brand A was bought by 40 people, Brand B by 30 people and Brand C by 25.
Both Brands A and B (and possibly C as well) were bought by 8 people,
B and C (and maybe A) were bought by 10 people, and A and C (and maybe B) by 7 people.
All three brands were bought by 3 people.

a) Represent this information in a Venn diagram. (5 marks)

b) Ifa person is selected at random, find the probability that:

(i) they buy at least one of the soaps (2 marks)

(ii) they buy at least two of the soaps (2 marks)

(iii) they buy soap B, given that they buy only one of the soaps fan)

A jar contains counters of 3 different colours. There are 3 red counters, 4 white counters
and 5 green counters. Two random counters are removed from the jar one at a time.
Once removed, the colour of the counter is noted. The first counter is not replaced before
the second one is drawn.
a) Draw atree diagram to represent this situation. (3 marks)

b) Find the probability that the second counter is green. (2 marks)

) Find the probability that both the counters are red. (2 marks)

d) Find the probability that the two counters are not both the same colour. Gene

Event J and Event K are independent events, where P(J) = 0.7 and P(K) = 0.1.
a) Find:
Gy Pik)
(1 mark)
Gi) POW)
(2 marks)
b) IfL is the event that neither J or K occurs, find P(L|K’). Bane)
marks

Chapter 2 Probability
When Albert eats in a certain restaurant, he always eats either chicken or beef for his main
course. And he always eats either sponge pudding or ice cream for his dessert.
The probability that he eats chicken is 1 the probability that he eats ice cream given that he
eats chicken is z, and the probability that he eats ice cream given that he eats beef is 3

a) Find the probability that he eats either chicken or ice cream, but not both.
(3 marks)
by) Fj ae :
) ind the probability that he eats ice cream. (3 marks)

c) Find the probability that he has eaten chicken,

g iven that he is eating
ating ice
i cream : (3 marks)

A film club with 20 members meets once a week. 14 of the members go every week and
13 plan to renew their membership for another year. Of those planning to renew their
membership, 10 go every week.
One member of the club is selected at random.
a) Find the probability that the person selected plans to renew their membership
and goes to the club every week. seman)

b) Find the probability that the person selected goes to the club every week,
but doesn’t plan to renew their membership. (2 marks)

Show whether or not going to the club every week and renewing membership
are independent. (ame)

For a particular biased dice, the event ‘roll a 6’ is called event B. P(B) = 022.
This biased dice and afair, six-sided dice are rolled together.
a) Write down the probability that the biased dice doesn’t show a 6. men)

b) Find the probability that at least one of the dice shows a 6. Cane)

c) Find the probability that exactly one of the dice shows a 6,

ene)
given that at least one of them shows a6.

Chapter 2 Probability
 Chapter 3 Discrete Random Variables

1. Probability Distributions
Learning Objectives: Probability distributions show the probability of a discrete random variable
Understand what is taking certain values. They can be used to make predictions about the
meant by a discrete outcomes of random experiments.
random variable.
Be able to find and Discrete random variables
First things first, you'll need to know what a discrete random variable is:
distributions and
probability functions. = A variable is just something that can take a variety of values —
Be able to find and
its value isn’t fixed.
= A random variable is a variable that takes different values with
distribution functions. different probabilities.
= A discrete random variable is a random variable which can
only take a certain number of values.
Tip: Remember that
discrete data can only A discrete random variable is usually represented by an upper
take certain values case letter such as X. The particular values that_X can take are
(unlike continuous represented by the lower case letter x.
data which can take all
the values in a certain These examples should help you to get used to the difference between x and_X.
range). See page 3.
Rolling a fair dice and recording the score: Tossing a fair coin twice
= X is the name of the random variable. and counting the number
Tip: Notice that in
these two examples, Here it’s ‘score on dice’. of
heads:
the discrete random ® x is a particular value that
XYcan take. = Xis ‘number of heads’.
variable can only take a Here x could be 1, 2, 3, 4, 5 or 6. = x could be 0, 1 or 2.
few different values. The
possible values are all
whole numbers — but
they don’t have to be. Probability distributions and functions
A probability distribution is a table showing all the possible values a discrete
random variable can take, plus the probability that it’ll take each value.

Draw the probability distribution table for xX,

where X is the score on a fair, six-sided dice.
. Xan take the values 1, 2, 3, 4, 5 and 6,
List all the possible values
each with probability z. that X can take here.

This notation means ‘the probability The probability of each number

thatXtakes the value x’. being rolled on afair dice isz.

Chapter 3 Discrete Random Variables

A probability function is a formula that generates the probability of X taking
the value x, for every possible x. It is written P(X= x) or sometimes just p(x). Tip: Even though it’s
A probability function is really just another way of representing the described as a ‘formula’,
information in the probability distribution table. the probability function
can just be a number —
you'll see ones like this
Examples ari 250 F Sulerisy in the examples.

a) A fair coin is tossed once and the number of tails, X, is counted.

Write down the probability function of X.
= To write down the probability function, you need to work out the
possible values, x, that
XYcan take and the probability of each value.
The outcome can either be heads or tails, so X can either take the value
0 (if it lands on heads) or 1 (if it lands on tails).
1
The probability of each outcome is 5° Tip: The probability of
x being any value other
* Now you can write down the probability, function: than 0 or 1 is zero — it’s
er . List the possible values impossible.
os : of x after the ‘formula’.

b A biased coin, for which the probability of heads is - and tails is el

—
a
is tossed once and the number of tails, X, is counted.
Write down the probability function of X.
= Again, the outcome can either be heads or tails, so XYcan
either be 0 or 1. The probability of heads, P(¥= 0) is 5
and the probability of tails, P(X= 1), is t:
= This time, the probabilities are different for different values ofx
— <0 it’s best to use two ‘formulas’, one for each x value.
Write the probability function as a bracket like this:

Put each value ofx next

to the ‘formula’ which
gives its probability.

There’s an important rule about probabilities that you'll use in solving

lots of discrete random variable problems:
The probabilities of all the possible values that
a discrete random variable can take add up to 1.
For a discrete random variable _X:

We can check this works for the fair coin in Example a) above:
+P= N=5h+z=
YSPX=xn = YP K=x =PX=0) 1Y
4
all x x=0,1

The only values The probability of

ofx are O and 1. each outcome is $.

Chapter 3 Discrete Random Variables

 You can use the fact that all the probabilities add up to 1 to solve problems
where probabilities are unknown or contain unknown factors.

The random variable X has probability function

P(X=x)=kx x=1, 2,3. Find the value of k.

So Xhas three possible values (x = 1, 2 and 3), and the probability of

each is kx (where you need to find the unknown f).
It's easier to understand if you write out the probability distribution:
x hap 2 3
Tip: It'll often help
to write down the
probability distribution
table when solving Now just use the rule: })P(X=x) = 1
problems like these — allx
that way you won't miss
Here, this means: k+2k+3k=6k=1
out any values.
> k=

You may be asked to find the probability thatX is greater or less than a value,
or lies between two values. You just need to identify all the values that XYcan
now take and then it’s a simple case of adding up all their probabilities.

Sein ete —ee

a ee ee)
Find: a) the value of a, b) P(2<X <4), c) the mode

a) Use })P(X
all x
= 1 again. RC
From theitable?” 04-0:.2: 502430 Deane
=> 0.8+a=1
=> hae 02

b) This is asking for the probability that

Tip: Careful with the
‘X is greater than or equal to 2, but less than 4’.
inequality signs — In other words the probability that X¥= 2 or X= 3.
you need to include
X = 2 but not X= 4. Just add up the probabilities: The events X= 2 and X= 3 are
P(2 <X < 4) = PUG? 073) mutually exclusive so you can
add the probabilities. See p63.
Tip: The mode of a = PLY= 2) + P(X= 3)
random variable is the =0.3 + 0.2 =
value you'd expect
to occur most if you
repeated the experiment c) The mode is the most likely value — so it's the value
lots of times — the most with the highest probability. The highest probability
likely value. in the table is 0.3 when_X = 2, so the mode = 2.

Chapter 3 Discrete Random Variables

When it’s not clear what the probability distribution or function should be, it
can be helpful to draw a sample-space diagram of all the possible outcomes
and work it out from that. For more on sample-space diagrams, see page 54.

An unbiased six-sided dice has faces marked 1, 1, 1, 2, 2, 3.

The dice is rolled twice.
Let X be the random variable ‘sum of the two scores on the dice’.
a) Find the probability distribution of X.
= To find the probability distribution, you need to identify all the
possible values, x, thatXcould take and the probability of each.
The easiest way to do this is to draw a sample-space diagram
showing the 36 possible outcomes of the dice rolls:

Score on rolt 1
$441 242 3
So 2 4
G1) 279 4
S122 4
S913 3 5
DOs
WY
3 5
314 4 6

= From the diagram you can see that there are only five values that
A can takes = 23 2417.0:

=» Since all 36 outcomes are equally likely, you can find Tip: See p53 for the
the probability of each value by counting how many equally likely outcomes
times it occurs in the diagram and dividing by 36. formula.
9 out of the 36 outcomes give a score of 2.
SoP\X=2)= =F
12 out of the 36 outcomes give a score of 3. Tip: Don’t forget to
change the fractions into
So P\X= 3) = 42 =4 their simplest form.
Similarly,
PUX= 4)=70 = 3, PUY= 5) = se = 4,

Tip: You should always

= So the probability distribution is: check that all the
probabilities in your
table add up to 1 — if
they don’t you’ve done
something wrong.

b Find P(X< 5).

—

Xis strictly less than 5, in other

This is asking for the probability that
r.
words X takes values 2, 3 or 4. Just add the probabilities togethe

PIX <5) = PIX =2) + P= 3)+ PIX =A= L434 = BE

Chapter 3 Discrete Random Variables
 A game involves rolling two fair dice. If the sum of the scores Is
greater than 10 then the player wins 50p. If the sum is between 8
and 10 (inclusive) then they win 20p. Otherwise they get nothing.
a) If X is the random variable ‘amount player wins’,
find the probability distribution of X.
* There are three possible amounts of money to be won, so there are
three possible values thatX can take: 0, 20 and 50.
= For each x value, you need to find the probability of getting a sum of
scores which results in that value of x.
P(X= 0) = P(Sum of scores < 8)
P(X= 20) = P(8 < Sum of scores <10)
P(X= 50) = P(Sum of scores > 10)

* To find these probabilities, draw a sample-space diagram showing

the 36 possible outcomes of the dice rolls.
Mark on your diagram all the outcomes that give each value of x.
Score on dice 1
+| 1 Z 3 4
112 3 4
N

9 21% 4 5
ous 4 5 6
oe 5 6 i
&5 6 7. 8 10
6 be 8 97410 bis?

= 21 out of 36 outcomes give a sum of scores which is strictly less than 8,

so P(X= 0) = P(Sum of scores < 8) = 21
"369° 12
= 12 out of 36 outcomes give a sum of scores between 8 and 10 inclusive,
so P(X = 20) = P(8 < Sum ofscores < 10) =42 = 4 :

= 3 out of 36 outcomes give a sum of scores strictly greater than 10,

so P(X= 50) = P(Sum of scores > 10) = & - a

= Using this information, draw the probability distribution:

b) The game costs 15p to play. Find the probability of making a profit.
A player will make a profit if they win more than 15p — so
if they win 20p or 50p.
= So the probability of making a profit is
POXS15) = PX 720) 4 Pax = 50) 3+ = 5

Chapter 3 Discrete Random Variables

Exercise 1.1

Q1__ For each of the following random experiments, identify:

(i) | The discrete random variable, X.
Xcan take.
(ii) All possible values, x, that
a) Tossing a fair coin 4 times and recording the number of tails.
b) Picking 4 balls from this bag
without replacement and recording
the number of orange balls selected.
c) Rolling a fair four-sided dice twice
and recording the sum of the scores.

Q2_ Afair six-sided dice is rolled. Write down the probability distribution
for the following random variables:
a) A =‘score rolled on the dice’.
b) B= ‘1 if the score is even, 0 otherwise’.
c) C= ‘5 times the score rolled on the dice’.

Q3 a) The random variable X has probability distribution:

rov-o[o2
[oaforfe
(i) Finda (ii) Find P(X2 2)

b) The random variable X has probability distribution:

z (ead Ortinl6onp. 2514036
P(x=x)| & | & | & | & | k |k
(i) Find k (ii) Find P(X= 5)
(iit) Find PX= 10) = (iv) _Find P(3 <.X< 15)
(v) Find P(X is divisible by three)

Q4_ For each of the probability functions in a) to c) below:

(i) Find k. (ii) Write down the probability distribution of X. Q4 Hint: Remember
that when a probability
a) P(X=x)=he x= 1, 2,3 function is written in
brackets, different values
b) P(X =x) = k x=1,2,3 ofx have probabilities
ae given by different
co PX =x) = E ime formulas.
k(8—x) x =5,6,7 cae

Q5_ An unbiased four-sided dice with possible scores 1, 2, 3 and 4 is

rolled twice.
Let Ybe the random variable ‘product of the two scores on the dice’.
a) Write the probability distribution of X.
b) Write the probability function of X.
¢) Find P3 <x =10)

Chapter 3 Discrete Random Variables

 The cumulative distribution function
The cumulative distribution function, written F(x), gives the
probability that Xwill be less than or equal to a particular value, x.
It’s like a running total of probabilities.
To find F(x,), for a given value x,, you add up all of the probabilities of
the values X can take which are less than or equal to x,.

Tip: You'll sometimes

see this written as
Fx) = Spe. | F(x,) =.PCX Sx). ae
C= Xe Xp

Remember that p(x) is

just the same as writing
P(X =x).

The probability distribution of the discrete random variable H is

shown in the table:

Draw upa table to show the cumulative distribution function F(A).

= There are 4 values of h, so you have to find the probability that H

is less than or equal to each of them in turn.
It sounds trickier than it actually is — you only have to
add up a few probabilities...

= Start with the smallest value of h:

F(0.1) = P(H< 0.1) — this is the same as P(H= 0.1), *
since H can't be less than 0.1. So F(0.1) = +

F(0.2) = P(H< 0.2) — this is the probability that H = 0.1 or H = 0.2.

So,F(0.2) = P= 0.1) + P= 0.2)=f+1= 1

= P(H = 0.1) + P(H = 0.2) + PH = 0.3)= 4+4+4

wl On

Chapter 3 Discrete Random Variables

 = P(H = 0.1) + P(H = 0.2)
+ P(A = 0.3)
+ P(H = 0.4)
|

This isn’t coincidence — P(X< largest value of x)

is always 1 because it’s the sum of all the
possible probabilities, which you know is 1.

= Finally, put these values in a table, and you're done...

Sometimes you'll be asked to work backwards — you can work out the
probability function, given the cumulative distribution function.
The probability that Xis equal to a certain x value is the same as the
probability thatXis less than or equal to that x value, but not less than
or equal to the next lowestx value.
To describe these x values, it can be useful to use the notation x,. Tip: You might have
For example, if Xcan take the values x = 2, 4, 6, 9, then you can seen this x, notation
label these as x, = 2, x, =4,x, =6 and x, = 9. in C1 — it's used for
sequences and series.
Using this notation, the probability that ‘Xx= x, is the same as the probability 4 be ee
that ‘X is less than or equal to x,, but NOT less than or equal to x,_,’.
This clever trick can be written:

AX) ) = POX Sy) = F (xx) — F(%- i)

The formula below gives the cumulative distribution function F(x)

for a discrete random variable X:
F(x) = kx, forx = 1, 2,3 and 4
Find k, and the probability function for X.

Questions with an unknown almost always want you to

use the fact that all the probabilities must add up to 1.

= To find k&, you know that Xhas to be 4 or less, so:

P(X< 4) = 1 +————— This is using the fact
Substitute 4 in F(4) = 1 that all the probabilities
for the ‘x’ in ‘kx’. 1 must add up to 1.
= 4k=150 k= |

Chapter 3 Discrete Random Variables

 Now you can use this to work out the probability function:

= First, work out the probabilities of X being less aan or equal

to 1, 2,3 and 4. This is easy — just substitute k = Z into
F(x) = kx for each x value.

F(1)=P(¥< 1)=1xk= ~
FQ)= Pixs 2=2K R= ;
F(3) = P(X <3)=3xk= enSS I ES) lA Bs i
BIO
luo
—

= But you need to find the probabilities of X being equal to 1, 2, 3 and 4.

Rips : Inpinta thi omipie| This is the clever bit,
it, jjust use:
x, = 1,25 =2, x, =3 and
X= iF

PLY = 4) = PWW< 4) - P(X<3) =1 ee Think about it...

3 1 1 hats <4

Pid = 3) = PIX S3)-PXS 2) 4-5-4 .. but nots 3,

Pe 2) Pe ee Pike 4= + = 5 _..then it has to be 4.
ae neil
P= Ty PX ae 4 Becausex doesn't take
any values less than 1.

= Draw out the probability distribution using this information,

so you can read off the probability function.
The probability distribution ofX is:

Exercise 1.2

Ql Each of a)-d) shows the probability distribution for a discrete random

variable, X. Draw upa table to show the cumulative distribution
function F(x) for each one.

poo |0.1 |
|x |-2|-1[o]1
[2]
ia
|x [2/4/98
[i6[32]64)
1] ts ta a fs
veo] 3/3
[8]ro]3535
Chapter 3 Discrete Random Variables
Q2 Each of a)-b) shows the probability distribution for a discrete random
variable, X. For each part, draw up a table showing the cumulative
distribution function, F(x), and use it to find the required probabilities.

a
Find i) .PXs 3) Find( P(X< 0)
i: P(X> 0)

Q3 The discrete random variable X has probability function:

PX=x)= 4, Boe ee 4,0,oO Q3 Hint: This time
you've been given the
a) Draw upa table showing the cumulative distribution probability function
function, F(x). instead of the probability
distribution, but the
PRE ee Ge 3 eau’ SRG eke7) method is just the same.

Q4 Each table shows the cumulative distribution function of a discrete

random variable, X. Write down the probability distribution forX.

X a2 aly OI
ECON Or OZ

Q5 The discrete random variable X has the

cumulative distribution function:
Q5 Hint: Find the

roa)o3|« [oe
probabilities P(X= x) for
each x as a function of
a, and then use the fact
Given that P(X= 2) = P(¥= 3), tilat ick Xe)
draw a table showing the probability distribution of X. to work out the value
of a.

Q6 For each cumulative distribution function, find the value of k, and

give the probability distribution for the discrete random variable, X.

a) F(x) =
(x 25
+ ky / awl pee p>
Q7 Hint: You need
osha
3

b) Fo) = 20 to find P(X

< x) — the
probability that the
©) Fx) =2%-", a1 3 largest score shown by
the two dice is no more

The discrete random variable X = ‘the larger score showing when a

than x. This means that
Q7 both rolls must score
pair of fair six-sided dice are rolled’. no more than x, so
a) Show that the cumulative distribution function, F(x), is given by: you're looking for the
354, 5,6
probability that two
F(x) = x, ogee,
independent rolls each
b) Hence, find the probability distribution forx. score no more than x.

Chapter 3 Discrete Random Variables Vy

 2. Expected Values, Mean
and Variance
Learning Objectives: You’ve seen in Chapter 1 how to find the mean and variance of a data set.
= Be able to calculate Here, you'll be finding these measures for discrete random variables.
the mean and Instead of working them out from observed data, you'll be finding theoretical
variance of discrete values for the mean and variance based on the probability distribution.
random variables.
® Be able to calculate
The expected value
the expected value The expected value of X, E(X)
and variance of
functions of discrete Every discrete random variable, X, has an ‘expected value’ (or mean) ECX).
random variables. In theory, it’s what you'd expect the mean of the data to be if you took
Use the discrete
lots of readings. In practice, the mean is unlikely to be exactly E(X), but it
uniform distribution should be pretty close. ;
to find the mean and
If the discrete random variable X can take values x,, x,, x...
variance of discrete
uniform random
then the expected value of X is:
variables.

Tip: Remember that x, So all you do is multiply each x-value by its probability and add them all
is just notation for all together.
the different x-values,
X,, X, etc.

SO >)%PLX= 2)
= Gee ®)
The probability distribution of X, the number of daughters in a family of 3
+ P(X= x) +...
children, is shown in the table. Find the expected number of daughters.
And p, is just short for
P(X= x).

Tip: The expected value Just add up all the x, p’s:

(or mean) of a random
variable is sometimes EX)= Yixp. = [0 x 3]+[1 x 3]+[2 x2 +[3 x4
labelled yp.
=0+% 3

Tip: It sounds abit

weird to expect 1.5
daughters —
but all it means is that
if you check a large
number of 3-child
families, the mean will
A discrete random variable, X, has the probability distribution shown
be close to 1.5. _in the table. Given that E(X) = 2.7, find a and b.

= There are two unknowns to find, so you'll need two bits of

information. The first is given in the question, E(X) = 2.7.
E(X) = Dixp =[1 x 0.2] +[2 x 0.3] 4+ 3a4 4b
= 2.7 =0+.8
3a+ 4b
So ~9i=s36-—45

Chapter 3 Discrete Random Variables

 = It might not be obvious what the second bit of information is —
nothing else is given in the question. But there is something that is true for
every discrete random variable — all the probabilities must add up to 1.
DP = 1
S0 0.5+ a+- b= 1's0 Baepae0r5
Now you just need to solve the simultaneous equations
Tip: This is the
Chrengec and Opeureqs ‘substitution’ method
of solving simultaneous
equations. In general
(1) gives a = 0.5 — b and substituting this into (2)gives: it means isolating one
of the variables (getting
3(0.5 —b) + 4b =1.9
a or b on its own) and
15304b = 1.9 substituting it into the
(oer 1-9 other equation. You can
also solve simultaneous
b =0.4 equations using the
‘elimination’ method
Substituting b = 0.4 into (1)gives a=0,1 (which involves adding
or subtracting them to
get rid of one unknown).

The expected value of X’, E(x?)

It is sometimes useful to work out the expected value of ‘X?’, E(X?) Tip: You'll be using E(x?)
— for example, when calculating the variance of X (see p93). to work out the variance
You can work out a probability distribution for X*, just as you can for _X, a lot in this chapter
and then calculating the expected value is easy. so make sure you get
comfortable with it here.

Find the probability distribution of X’.

The probability distribution table for X? is similar to the one for X.
You just need to square the.x, values and keep the probabilities the same.

The distribution of X? is:

. If the discrete random variableX can take values x,, x,, X3,.-.
then the expected value of X° is:

Chapter 3 Discrete Random Variables

 For the discrete random variable defined in Example 1 on the
previous page, find the value of E(X*).
Using the formula:
Tip: Notice that E(X’) is
not the same as [E(X)]’. a > Pi => O x

Exercise 2.1 :

QI Find E(X) for each of the following discrete random variables.

Q1 Hint: You might 2 ,

want to write out the ay PWax= 02 x= 0,1,2,3,4 = 4% x=1,2,3

b) PW=x) 14
probability distribution
to help you find E(x) — 0.1) eS 2 0.1, s=-2
but you don’t need to. c) roan = fo x=3 a eur=n-f02 x==1, 12
OS wins1 Osurx=30

Q1 Hint: For parts e) e) P(X=x)=hv+2) x=1,2,3,4,5 1) Pixies) Sak

X
wey 276,42
and f), you'll need to
work out the value of k Q2__ Each table below shows the probability distribution of a discrete random
first. eee: X. Use the information given to find the required values.

re 0
If E(X)= 6.5 find the value of p.

[en Fd Ble By
Q?2 b) Hint: Find afirst. |Pix=x)
105]02|a }02
If E(X) = 7.5 find the values of a and p.

Q3 Each table below shows the probability distribution of a discrete

random variable, X. Use the information given to find a and b for each.

eee Pm =| 3 [7|8 ae
Povss) 02]a |ot]b Pars fot]
a| bs}.|
where a=o225 where E(X) = 7.8

Q4 In parts a)-c), the tables show the probability distribution for a discrete
random variable,
X. For each part, find: (i) EQ) (it) EQ)

Chapter 3 Discrete Random Variables

Variance Tip: The variance of a
Variance is a measure of dispersion — it measures how spread out data set is given by
values are from the mean. dix — x)
variance =
Just like you can find the mean (j) of a discrete random variable, you
‘can also find its variance. This measures the spread of the random — so the variance is
variable’s distribution about its mean. Again, it’s a theoretical value — ‘the average squared
deviation from the
it’s the ‘expected variance’ of a large number of readings.
mean’ — see page 31.
:
The graphs below show the probability distributions
of two discrete random variables, X and Y:
Tip: E[(X— py]
= Do(% — wy
0.5 Bd 0.5 Y
because to work out
0.4 ee A an expected value, you
.I 0.3 as
ll 0.3 always:
eal? = 0.2 (i) swap the random
variable in the
0.1 0.1 expression for each
of its possible values
hie WAND , je ees — so here you need
x y to find (x, — 4)? for
If you were to collect a large number of readings of both X and Y, you’d the various possible
expect the readings of Xto be less spread out than the readings of Y. values x,,
(ii) multiply each of
But... even if you don’t collect any actual readings of X or Y, you can still
those values by the
work out the variance of X and Ybased on their probability distributions. probability (p,) of x,,
And you'd expect the variance of Xto be smaller than that of Y.
(iii) sum the results.

Now then... suppose a discrete random variable X can take values x,, ye x,,Mor x Agrees
The variance of X is ‘the expected squared deviation from the mean (i) Tip: You can remember
— in other words: — pny] = Dae — YD; (see Tip above).
Var(X) = E[(X¥ the variance formula as:
But there’s a different way to write the formula for the variance ‘mean of the square
that’s much easier to use: minus the
square of the mean.’
VarlX) = EX?) — [EX = Yxtp, — [xpi
Tip: Remember...
(i) P(X¥=x) =p,
In fact, the formula for the variance can be written in various ways.
For example, on the formula sheet you'll get in the exam, it’s written as: (ii) E(X) = w= Dpxpi
(iii) E(X?) = >) xp:
Var(X) = (4 — BYP(X = x) = Dox P(X =x) —

But these formulas are all equivalent — they'll always give the same answers.
Tip: You'll see on
And you can always take one of these formulas and ‘turn it into’ any of the p96 that
others — just by rearranging it, and using the definitions on the right. E(aX + b) = aE(X) +b
For example:
This tells you that
you can ‘break apart’
Var(X) = E[(X— 47] an expected value
= ELX? — 2uX+ y’] (multiplying out the brackets) involving addition, or
multiplication by a
= E(X?)— 2uE(X) + wv? (see the bottom Tip for more info)
number (not by another
= £(X2)—2[EQO))? + (E(x)? (using the definition 4 = E(X)) variable, though). This
means you can write
=r E[X? — 2X+ p’] as
E(X?) — 2wE(X) + pv’.

Chapter 3 Discrete Random Variables

 x=0,1, 2, 3, 4.
X has the probability function P(X =x) =k(x+1)
Find the mean and variance of X.

= First you need to find k.

Work out all the probabilities and remember they add up to 1:
P(X=0)=kx(0+1)=k

Similarly, P(X = 1) = 2k
PX= 2.23%
P(X= 3) = 4k
P(X= 4) = 5k
Go k+ 2k+ 3k+4k =1,1.
45k andso k= ae
1,=
15k

= Now use the formulas — find the mean, E(X), first:

ECX) = >ioop: : :

=[ox]+[1x A) +2 xsltB xis] tl4 x a5]

_40_ &
Ns) 3.

= For the variance you also need E(X”):

Tip: Work out the
beta) = ee
variance in steps to
avoid making mistakes. é
= [ox + [Px B)+ [2 * a) + xsl +l?
You know E(X), so now =, (UO
find E(x’). 1 3

Then put them both into

the formula for variance. # And finally:
Var(X) -= E(X?) — [EQO? _26_/37_
= <2 — [3] = @
|

The discrete random variable X has the following probability distribution:

rites srl Belida

Pix=x |02] « |05]0.2|
If E(X) = 5.2, find the values of a and p and the variance of X. -
» First find the value of a.
Use the fact that the sum of the probabilities must be 1:

0.2+a+05+0.2 =1
>0.9+a =1
=> a=0.1

Chapter 3 Discrete Random Variables

" Now use E(X) = 5.2 to find the value ofp:
E(X) = [1 x 0.2] +[4 x 0.1] + [p x 0.5] +[8 x 0.2]
=2.2.40.5p
So 5.2 =2.2 +0.5p yates
== 050
=> Bee
o

* Now work out E(X’) so you can calculate the variance.

E(X?)= [1* x 0.2] + [4° x 0.1] + [6? x 0.5] + [8? x 0.2]

=[1 x 0.2]+[16 x 0.1] + [36 x 0.5] + [64 x 0.2] = 32.6

So Var(X)
= EQ’) — [E(X)P
= 32.6 —5.2?= 32.6— 27.04 =—aaas

Exercise 2.2

Q1 ‘For each of the following probability distributions,

calculate the mean and variance of X.

“ ee ee : pete
C)
Leee eee
Q2__ Each of the probability functions below defines a discrete random
variable, X. Calculate the mean and variance of X.

a) P=)= x= 12,354,500 -b) PUX=x)= 3 oe

Q3 For each of the probability distributions below:

(i) Finda. (ii) Find the mean and bari OiLX.

a) x pie) Bay 0
P(Xx=x) |0.2/a)0.4| 0.1 =
P(X ra 0.1

Q4 Each of the probability distributions below describes a discrete

random variable, X.

* EatsIf EX)= 5.2, find the value of p and the variance One

If E(X) = 5.7, find the values of a andp and the variance of X.

Q5_ Each of the probability functions below describes a discrete random

variable, X. For each: (i(i) Findk (ii) Find E(X) (iii) Find Var(X)

ee es) by Pan=2 x23) 4,5,6

Chapter 3 Discrete Random Variables
 Expected value and variance of a
function of X
Tip: A function of X
is just an expression You can find the expected value and the variance of any function of X, g(X).
involving X.
If the function of
Xis linear (i.e. of the form aX + b, where a and b are
Tip: On the exam any numbers), there are two helpful rules that will speed things up:
formula sheet, you're
given the formula:
E[g(x)] E(aX+ b) = aE(X) + b Var(aX+ b) = a’*Var(X)
= Dia(x)P(X= x)|
for a function of X, g(X).
If you were to put the
Example 1
function g(X) = aX+ b
into this formula and
simplify, you’d get the If E(X) = 3 and Var(X) = 7, find:
rule for E(aX+ b) shown a) E(2X +5) b) Var(2X+ 5) c) E(2—4x)
here.
d) Var(X
—2) e) Var(-)

a) The function of Xis 2X +5, so leta=2 andb=5 and

Tip: On p91 we were use the formula for E(aX+ b):
actually finding E(g(X)),
where g(X) = X?. E(2X +5) =2E(X)+5=(2x3)4+5=111
vag
E(X)
=3

b) The function of Yis 2X +5, so leta=2 andb=5 and

use the formula for Var(aX+ b):
Var(2X+ 5) = 2?Var(X) = 4 x 7 = ae
A
Varixi= 7

c) The function ofX is 2-4X. Don’t get confused about it being

the wrong way round — it's just the same as —4X+ 2.
So let a= —4 andb = 2:

E(2 — 4X) =-4E(X) +2 =(4 x 3) +2 =8e@

d) The function of Xis X—2, so leta=1 andb= —2:

Var(X— 2) = 1? Var(X) = 7

e) The function of X is 5. This is just the same as

Tyg O,
1 2
S o leta — =5 andb=0:

Chapter 3 Discrete Random Variables

 Example 2

eta asta Fab «|

x)
Find: a) k, b) E(X), ©) Var(X), d) E(3X-1), e) Var(3X—1)

a) Using the fact that the probabilities add up to 1 —

0.1+0.2+034+02+k=1,andso k=0.2

b) Now you can use the formula to find E(X):

E(X) = Domp
= (2x 0.1)+ Gx 0.2)+ (4x 0.3)+(5 x 0.2)
+ (6 x 0.2)=

c) Next work out E(X’) so you can use it to find the variance:
EX = > ae
me a0 is be 01a 14 G0 |
+[5? x 0.2] + [6° x 0.2]
==» (,9:2
Then the variance is easy:
Var(X) = E(X?) — [E(X)]? = 19.2 — 4.2? ="

d) Using the formula for E(aX + b):

EGX—1)
= 3EQ)-1=3 x4.2-1 =Giee

e) And finally: Var(3.X— 1) = 3?Var(X) = 9 x 1.56 = 14.04

Exercise 2.3

Q1. Xisa discrete random variable with E(X) = 4 and Var(X) = 3.

Write down the mean and variance of the following random variables:
edhe be Z = 5X CW 22% 7 d) Wshai2x

Q2 Xisa discrete random variable with probability distribution:

Q2 Hint: The
probability distribution
E(X) = 3.2 and Var(X) = 1.56. of a function of X is
easy to find — just
For each of the random variables in a)-d) below, find:
apply the function to the
(i) the probability distribution, values of x, but keep the
(ii) the mean and variance directly from the probability distribution, probabilities the same
given.
(iii) the mean and variance using the values of E(X) and Var(X) (just like we did with
a) Y=3X+4 = b) Z=3X-4 ey V=20=3X “d) W=20+4 3X x2 on p91):

Chapter 3 Discrete Random Variables

 Q3 a) The random variable X is given by the probability distribution:
rr oe as

(i) Find E(X) and Var(X).

(ii) Y=4X +3. Find E(Y) and Var(Y).
(iii) Z=50—2X. Find E(Z) and Var(Z).

b) The random variable X is given by the probability distribution:

(i) Find E(X) and Var(X).

(ii) Y=7-2X. Find E(Y) and Var(Y).
(iii) Z= 7+ 2X. Find E(Z) and Var(Z).

Q4 The random variable X is given by the probability distribution:

Q4 Hint: Don’t get
confused by the letters m
and c in this one — just 0.3
Bets Gee Rae | a) Find E(X) and Var(X).
r e
MCCA ie b) ¥=26-mX. If E() =0, find mand Var().
preem c) Z=3X—c. If E(Z) = 30, find'eand Var(Z).

The discrete uniform distribution

Sometimes you'll have a random variable where every value ofX is equally
likely — this is called a uniform distribution. For example, rolling a normal,
unbiased dice gives you a discrete uniform distribution.
The probability distribution of a random variable with a discrete uniform
distribution always has the same shape as the one below. This diagram
shows the probability distribution for X, where X takes consecutive whole
number values between a and 5 with equal probability.

There are b—a+ 1 possible values, each with

the same probability. Since all the probabilities
must add up to 1, this probability ispl eae

a a+1 a+2 a+3°" sb x

These numbers are consecutive whole

numbers — there are b—a+1 of them.

Chapter 3 Discrete Random Variables

You can use the same formulas for E(X) and Var(X) as you have done for all
other discrete random variables throughout the chapter — so you don’t need
to learn any new ones.
But for a discrete uniform distribution X which can take consecutive whole
number values a, a+ 1, a+ 2,..., b, you get nice simple formulas for the mean
and variance. These will save you some time if you can remember them.

Mean = E(x)= 44°

Variance = Var(X) = Cee te)

1D

eli
olom
Find the mean and variance of the score on an unbiased,
standard six-sided dice. c

If X is the random variable ‘score on a dice’, then_X has

the discrete uniform distribution shown in the table:

The smallest value ofx is 1 and the biggest is 6 —soa=1andb=6.

Now just substitute these values into the formulas:

Mean Pedi
= tkvie \otO weeees 3.5

Ps aot ss 2 Tip: The symmetry of

Variance = ( = 1) | = (6-1 ~ I) 1 = # = O40 3s the distribution should
tell you where the mean
is — it has to be halfway
between 1 and 6.

Example 2
A lottery involves a ball being picked at random from a box of 30 balls
numbered from 11 to 40. The random variable X represents the number
on the first ball to be picked. Write down the probability function of X,
and find its mean and variance.
Each ball has a probability of<n of being picked first,
so the probability function Is:
Pave= 1112 40
The smallest value is 11 and the largest is 40, so a= 11 and b= 40.
Using the formulas:

Mean = 248 = 1440 _3_ 25.5

era 40 Lih- ealy=1 ape

809 3 s.f.
74.9 to
apres ioe aa =| 12 )

Chapter 3 Discrete Random Variables

 Exercise 2.4

Ql A random variable, X, has a discrete uniform distribution and can

take consecutive integer values between 1 and 5 inclusive.
Draw a table showing the distribution of X, and find E(X) and Var(X).
pe
A random variable, Y, has a discrete uniform distribution and can
take consecutive integer values between 12 and 15 inclusive.
Draw a table showing the distribution of Y, and find E(Y) and Var(Y).

Q3 A random variable, Z, has a discrete uniform distribution given by the

probability function:
Q3 Hint: Recall that
the standard deviation ies in Cielo
is the square root of the
variance. Draw a table showing the distribution of Z, and find the mean and
standard deviation of Z.

Q4 A random variable, D, has a discrete uniform distribution given by

the probability function:
Q4 Hint: The values
of d aren’t consecutive PD a) a=,
3, Deen
whole numbers. D|—

Find: a) The mean of D b) The variance of D

Q5 A random variable, X, has a discrete uniform distribution given by the

Q5 Hint: Part b) may
look tricky, but just
probability function:
break down the event
|X — u| <a into all KX=A 2p, 4,5;,,0,7,0
possible cases and a) Find the mean, uw, and variance, o?, of X.
add up the separate
probabilities. b) Calculate P(X — n| < o)

Q6 A random variable, Y, has a discrete uniform distribution and can

take the value of any single-digit non-negative integer (including
zero).
a) Find the mean, u, and variance, Oo, Oty.
b) Calculate P(|¥-yl| > 0)
c) Calculate P(|Y¥—-p| > 20)

Q7 A random variable, Z, has a discrete uniform distribution

and
can take consecutive integer values from 1 to 7.
a) Find the mean and variance of Z.
b) Calculate the mean and variance OLY =) = 7:

cy
Chapter 3 Discrete Random Variables
Review Exercise — Chapter 3
Ql The probability distribution ofYis:

menleal Pe
a) Find the value of k. ) Find P(Y< 2).

Q2 The probability distribution for the random variable Wis given in the table.
Draw up atable to show the cumulative distribution function.
0.2 | 0.3 | 0.4 | 0.5

Q3 Each probability function below describes a discrete random variable, X.

In each case, draw a table showing the cumulative distribution function,
and use it to find the required probabilities.
(x= 2)
ar X= xh Kale) oO, 4D Find (i) P(X< 3) (ii) PU <x 3)
22

DEP =x) = =a ss 4158 6,7,Ore Find (ip PX <a) (i) PG <x 37)

Q4 The cumulative distribution function for a random variable Ris given in the table.
Work out the — distribution for R. Find PO < R < 1).

Fir)= P(R <r)

Q5 The discrete random variable X has a uniform distribution, P(Y= x) = k forx = 0, 1, 2, 3, 4.

Find the value of k, and then find the mean and variance of X.

Q6 A discrete random variable X has the probability distribution shown in the table,
where & is a constant.

a) Find k b) Find E(X) and show that Var(X) = a — 1)

c) Find E(2X— 1) and Var(2X

Q7 A discrete random variable X has the probability distribution shown in the table.

a) Find E(X)
ofaas
5 ffosfo b) Find Var(X) ;
Q8 Hint: You're looking
’ : / : for F@)= P(X sx). Find
Q8 Xis the random variable ‘highest score showing when three fair . the probability that a
six-sided dice are rolled’. Find the cumulative distribution function single fair six-sided dice
for X, and then use it to find the probability distribution for X. bac a ccoremomnore than
x and remember that dice
rolls are independent.

Chapter 3 Discrete Random Variables

 Exam-Style Questions — Chapter 3

In a game a player tosses three fair coins.

If three heads occur then the player gets 20p; if two heads occur then the player gets 10p.
For any other outcome, the player gets nothing.

a) IfX is the random variable ‘amount received’,

tabulate the probability distribution of X.
(4 marks)

The player pays 10p to play one game.

b) Use the probability distribution to find the probability that the player wins
(i.e. gets more money than they pay to play) in one game. (2 marks)

A discrete random variable X can only take values 0, 1, 2 and 3.

Its probability distribution is shown below.

eeeee ae

a) Find the value of k. (1 mark)

b) Draw upa table to show the cumulative distribution function for X.

(4 marks)
¢) Find P(X 2).
(1 mark)

The random variable X takes the following values with equal probability:
0,123 245576. Ss 9
a) Write down the probability distribution of X.
(1 mark)
b) Find the mean and variance of X.
(3 marks)
c) Calculate the probability that _X is less than the mean.
(2 marks)

Chapter 3 Discrete Random Variables

A discrete random variable X has the probability function:
P(X= x) = ax for x = 1, 2, 3, where a is a constant.

a) Show that a= =. (1mark)

b) Find E(X). ne
c) If Var(X) = 5 find E(X?). (2 marks)

d) Find E(3X + 4) and Var(3.X + 4). (3 marks)

The number of points awarded to each contestant in a talent competition is given by

the discrete random variable X with the following probability distribution:

Find E(x). (2 marks)

Find E(6X + 8). (2 marks)

Show that Var(X) = 1. (4 marks)

Find Var(5 — 3.X). (2 marks)

The discrete random variable X is given by the probability distribution:

Given that E(X) = 1.3:

a) Finda and b. (5 marks)

b) Find E(72X+ 3). (2 marks)

c) Show that Var(X) = 1.01. (3 marks)

d) Find Var(4 —X). (2 marks]

Chapter 3 Discrete Random Variables

 Chapter 4 Correlation and Regression

1. Correlation
Learning Objectives: Correlation is all about how closely two quantities are linked.
Be able to plot scatter For example, as one quantity grows, the other might grow as well.
diagrams using paired Or it might shrink. Or there might be no pattern at all to what it does.
observations of two
variables. Scatter diagrams and correlation
Be able to tell from Sometimes variables are measured in pairs — maybe because you want to
a scatter diagram
investigate whether they’re linked.
whether two variables
are positively These pairs of variables might be things like:
correlated, negatively = ‘my age’ and ‘length of my feet’,
correlated, or not
correlated. = ‘temperature’ and ‘number of accidents on a stretch of road’.
Be able to calculate Data made up of pairs of values (x, y) is called bivariate data. You can plot
a product moment bivariate data on a scatter diagram — where each variable is plotted along
correlation coefficient. one of the axes. The pattern of points on a scatter diagram can tell you
Be able to use something about the data.
coding to simplify = For example, on this scatter
the calculation €25

of correlation
diagram, the variables ‘my age’
coefficients. and ‘length of my feet’ seem linked
SRO TPT EEO PERETTI
OIRO — you can tell because nearly all the
points lie close to a straight line.
= As| got older, my feet got bigger
of
Length
feet
(cmy
and bigger (though | stopped
Tip: The line of best fit measuring when | was 10 years old).
doesn’t need to actually
pass through any of = The line of best fit on this scatter
the data points — it :
FE
diagram lies close to most of the
just needs to be near to points. My age (years)
them.
# The circled point doesn’t fit the pattern of the rest of the data at all —
so the line of best fit doesn’t need to pass close to it. A point like this
could show a measurement error (like here), or just a ‘freak’ observation.

820 oe = It's a lot harder to see any connection

S a between the variables ‘temperature’
° 16 ic = and ‘number of accidents’ on this
5 scatter diagram — the data seems
25 12 = scattered d pretty much everywhere.
E
Z = You can’t drawaline of best fit for this
data — there isn’t a line that lies close
to most of the points. (It would be hard
to draw aline lying close to more than
0 ee:
about half the points.)
Temperature (°C)

E
Chapter 4 Correlation and Regression
Correlation is all about whether points on a scatter diagram lie close to a
straight line.

= Sometimes, as one variable gets bigger, the

other one also gets bigger — in this case,
the scatter diagram might look like this. ———>
= Here, a line of best fit would have a
ss i Ip: O Or you Can say
Tip:
positive gradient. there’s a positive
= The two variables are positively correlated. correlation between
h
them.

= But if one variable gets smaller as the

a Other one gets bigger, then the scatter
diagram would look like this.
m In this case, a line of best fit would
have a negative gradient. Tip: Or you can say
= The two variables are there’s a negative
negatively correlated. correlation between
them.

= And if the two variables are not

linked at all, you’d expect a —__» Tip: Or you can say
random scattering of points. there’s no correlation
= It’s impossible to draw aline of between them.
best fit close to most of the points.
= The variables are not correlated.

Exercise 1.1
Q1 Owen asked a group of people how far from the town centre they
live, and how many visits they made to the cinema in the previous
year. The table shows his results.
|Distance from centre (km) |0.6 |7.0 |1.4 sufio ufo
|Cinema visits |14 ens ea 11

a) Plot a scatter diagram to show this data.

b) Describe the type of correlation shown.

Q2 This table shows the average length and the average circumference of
eggs for several species of bird. (All distances are measured in cm.)
|Length 6.2 |2.1 sf s7fiol 7.2145 |68 |
|Circumference 14 |6.3 Aes 24 18.7|11.3|18.4|
a) Plot a scatter diagram to show this data.
b) Describe the type of correlation shown.
c) One of the measurements was recorded incorrectly.
Use your scatter diagram to determine which one.
Chapter 4 Correlation and Regression
 Si, 5,, and S,,,
Tip: The formula for
Suppose you've got a set of bivariate data written as 7 pairs of values (x, y).
variance on page 31 is:
_ Sx
d(x — XY The formula for the variance of the x-values can be written:| variance, = >>
variance =
2

Tip: The forms

S= sy 2) and
ou
Similarly, the variance of the y-values can be written:| variance, _= =~
Sy»
5, = yy i (29)

are easier to use.

where |S.=
—
i(y—yy =
Rag Vee Se (29)

Tip: The form There's also a similar quantity S. that involves both x and y.
i
8,= Soy — 2A?
is easier to use.
Dy he NYY) = ae Dxdy

To get a feel for what S,, means, look at the scatter diagrams below.
Tip: The formulas for The dotted lines show the value of x (the mean of the x-values) and y (the
S., Sand S., will be mean of the y-values).
important for the rest of
this chapter. = This scatter diagram shows a pair of variables Seen
(x and y) that are positively correlated — so
most of the values fall in the shaded squares.
Tip: On the graph,
= For each point in the top-right square,
there are two points for
which (x— x )(y — p) is
(x-x)y- y)>O0 (sincex> x andy> y).
negative, but there are # Similarly, for each point in the bottom-left
far more for which square, (x — x )(y— y) > O (sincex < ¥ andy< jy).
(x — ¥)(v — 7) is positive. = This means that for positively correlated variables,
So overall, S_xy is positive. S,= D(x — ¥)(y — 7) will be positive.

= In this scatter diagram, x and y are negatively

Tip: There are two correlated. Again, most values fall in the shaded
points for which squares.
(x — ¥ (vy — y) is positive,
but there are far more
For each point in the top-left square,
for which (x — ¥)(v — 7) x= 2)\y= 7) <0 (sincex <a butiyeary ),
is negative. And for each point in the bottom-right square,
Overall, S_yy is negative. (x-— x)(y— y) < O (sincex > ¥ buty< 7).
So for negatively correlated variables,
S.,= Di(x — ¥)(y — J) will be negative.

Tip: The positive and

negative values of = In this scatter diagram, x and y are not correlated.
(x — x )(y — y) largely
cancel each other out. ® In this case, about ie the points give a positive
value for (x — ¥)(y — y), while the other half give
What's left after all this a negative value.
cancelling will be close
to zero. # So Sy = >)(x —X)(y —7) ends up close to zero.

Chapter 4 Correlation and Regression

 Example 1
Find S., S.. and S\, for TASS 2.9oi jes. Tip: The formulas for
the data on the right. Te 24 S., S,,and S are on the
formula sheet you'll be
= You're going to need to find })x, Vy, 3x2, Diy? and Yixy. Berns ah
So add rows to the table showing x’, y? and xy.
And add an extra column to show the total of each row.

Prlkaeie aks 9 [3a [38 | dix =147

y 5 8
ui ie le et a8 Tip: Always draw a
1.21 | 2.25 | 5.29 | 8.41 | 9.61 | 14.44[ >2x? = 41.21 table — it'll help you
y | 25 | 64 121 289 | 576 > y S452| avoid mistakes.
xy oD 2 46.4 2a OZ

= Now you can use the formulas for S_, Sand S..
Tip: Remember... 7 is
= Sin DY
2) ='533 4 — a 81 _ @yere the number of pairs of
data values.

Sa = x earch 14.7? = 5195

— SE = 41.21- 47

el 1331-8= 9375
2 2

Exam questions won't always use the letters x and y to label bivariate data.
But you can still use the formulas for S_, S,, and S., by replacing x and y
with the letters in the question.

Example 2 yesh! amass

Use the following summations to find S|, Sand S_.
Sp =15.7, q=36, yp? = 45.55, ig’ = 278, Dipq = 93.2, n=5

= To find S,, use the formula for S,, — replacing x with p and y with q.
a ZPD _ 2 —15.2x 36 _ a9184
= To find S,, use the formula for S.,— but wherever you see x, write p.

Se pp (QP) = 45.55 15:7 — 631748

= Do asimilar kind of thing to find S__.
a D
Ss = vg — 2) 2973 3 - GR

Chapter 4 Correlation and Regression

 Exercise 1.2

Q1_ For the data shown in ea eo

the table on the right:
a) Calculate Yam yp aye panda
b) Use your summations to calculate S.,xy! S.. and 5.

Q2 Find S., Sand S,, in each case below.

a) ee tO > = 167
> =2031 >ixy = 589 i=
b) Six=206 Diy=50 Six? = 4504
Syrr= 326 ey Ol n=10

Q3 Finds ,S and S_, given the information below.

Pq PP q9
Yip =115 Yig=114. Yip?
=2383
Did = 2702 >ipq=1880 n=6

Product moment correlation coefficient

The product moment correlation coefficient (PMCC, or r for short) measures
the strength of the correlation between two variables. It basically tells you
how close to a straight line the points on a scatter diagram lie.

S
The formula for the PMCC involves S., S_andS.;yy | 4 = =
Hy 29 Say

Tip: In reality, you’d

never expect to get a The PMCC is always between +1 and -1.
PMCC of +1 or -1 —
= If all your points lie exactly on a straight line with a positive gradient
your scatter diagram
points might lie pretty
(perfect positive correlation), r = +1.
close to a straight line, = [fall your points lie exactly on a straight line with a negative gradient
but it’s unlikely they’d all (perfect negative correlation), r =-1.
be on it. = Ifr=0 (or more likely, pretty close to 0), that would mean the variables
aren’t correlated.
Tip: The closer the value These graphs give you an idea of ‘what different values of r look like’.
ofr to +1 or—1, the
stronger the correlation
between the two
variables. (Values close
to +1 mean a strong
positive correlation,
and values close to —1
mean a strong negative
correlation.)

Values of r close to zero

mean there is only a
weak correlation (but
see. p113).

eS
Uc
Wess
Chapter 4 Correlation and Regression
The following data show the score (x) of 10 students in a reading test,
and the time in seconds (y) it took them to run 40 metres.
feo aes5 ft Ak ee Me
y| 9.8 5.8 |8. 8.;
Illustrate the data with a scatter ree and = the are moment
correlation coefficient (r) between the variables x and y.
= The scatter diagram’s the easy bit
— just plot the points.
* Now for the correlation coefficient.
From the scatter diagram, the points
look pretty randomly scattered
— so you'd expect a correlation
Oe 2 ae eA ee Om 7
coefficient fairly close to zero.
= You need to find S., Sand S_, so add a few extra rows to your table.

[>[oa[ar [oe [as

x? |12.25]30.25/37.21/17.64] 7.29 3.61 |30.25]14.44|26.01/13.69] >)x* = 192.64

= This gives: Sy= >)xy— a)

2) = 316.62 - Ee? =-0.9

Sea x
: ey
x)oe et = a0
om tales

Sy = ie ==
ux
i = 596.82 —
75.67
0
_= 5264 Tip: This is very close
to zero — there seems
to be little (if any)
=» Now use the formula for r: r = correlation between
these two variables.
=. 0.) _ss _- Gi
J16.24 X 25.284

A scientist collects data on the percentage water content (w) of 8 different

brands of ham, and its price in pounds (p) per 100 g.
The results are summarised below.
Siw =152 Dp =7.73 Siw = 3492 Dp? = 8.0169 D)wp = 130.41
a) Calculate the product moment correlation coefficient (r).

= Here, Sy» = Diwp— LW 2aP = 130.41 — 2 X22 — -16.46

2
Tip: Don’t round your
Vi (dw) =13497 = 12 G4 answers yet — you're
going to use these values
60 Zi — 0.5477875 to calculate r.
SRT 2 _ (DPy ee
Chapter 4 Correlation and Regression Tey
 ss — 16.46 =S — 0.905 (to 3 d.p.)
Tip: When you're asked 20. miTigcige = 604 x 05477875
to interpret a value of r,
always give your answer b) Give an interpretation of this value of r.
in the context of the
question (so here, talk = This is a very high negative value forr(it’s very close to —1).
about ham, price and « This means that for the brands of ham tested, a higher percentage
water content). water content generally means a lower price.

Exercise 1.3

Ql Use the information below to find the product moment correlation

coefficient (r) between x and y.
ee Se ex =:6875
ip Sze yay tS ji 5

Q2 The table below shows the heights and weights of 8 teenage boys.
Height in cm, x |180]171 |182
Weiginkgy
[7067[66[50[or[75[os[56
a) Calculate the values of S,, Sand S_.
b) Find the value of the product moment correlation coefficient (r)
for x and y.

Q3 For the data shown

|p [13]9 [15|10]8 [11[12 |14)
in this table:
PAEBEARRE ES =
a) Calculate >)p, ig, yp’, og? and pq.
b) Use your summations to calculate S.,
Pq
Sands.
PP 99
c) Find the value of the product moment correlation coefficient (7)
forp and q.
>

Q4 The lengths and widths (in cm) of 8 leaves from a tree

were measured. The results are shown below.

a) Calculate S,, S, and S\

(You may use: >) 142258, > wee 2 > lwa181-75)
b) Calculate the correlation coefficient r.
c) Give an interpretation of your value for r.

Q5 A nurse at a health centre did a memory test on some

patients of different ages. Her results are shown below.

79
Testscore,s _[8.9]40]54|2.8|7.1[75]31
[62[04]
Calculate the correlation coefficient r and interpret your result.

a
Chapter 4 Correlation and Regression
 Qé A doctor checked the kidney function of some of his patients to see
if it was related to their weight. Here, 0 indicates no kidney function
at all and 10 indicates that a patient's kidneys are functioning at
maximum efficiency.
The table shows the results.
|Weight inkg w | ,66| 74 | 96 | 83 |79|54] 64 |71 | 88 |
|Kidney function, k[ 7.9 [8.2 |2.3[7.1 [48/91 [8.4[68 [3.71
a) Find the value of the product moment correlation coefficient (r)
for w and k.
b) Interpret the value of this correlation coefficient.

Coded data
Tip: If your data consists
You've seen the idea of coding before (see p37). The idea is to make your life of really huge numbers,
easier by making numbers less fiddly to work with. you could use coding to
The coding you’ll be expected pg produce smaller, more
to use is of the form: A b manageable numbers.

where x is an original data value, u is the corresponding But if your data values
coded data value, and a and b are numbers to be chosen. are tiny, you’d code
them to produce larger
This is a linear transformation — if you were to draw a graph of w against numbers. And so on.
x, it would beastraight line.
Tip: The reason why the
Using a linear transformation to code data values doesn’t affect the value of correlation coefficient
the product moment correlation coefficient at all — r for the coded data is isn’t affected is because
exactly the same as the value of r for the uncoded data. coding a set of values in
this way doesn’t change
the pattern of points
Example ee ee on a scatter diagram.
The effect of coding a
The table below shows the annual salaries (s, in £) of six shipping variable is the same as
clerks, see with the number of years experience (x) they have. relabelling one of the
scatter diagram’s axes.

Salary, s 23 000 | 24 000 | 24 500 | 26 000 | 26 700 | 29 000

s — 20000 Tip: You can code one

The values of s are to be coded such that u = or both of the variables.
1000 ~
a) Calculate the value of the product moment correlation coefficient
between x and wu. Tip: The values of
= Make a new table of values which includes the coded data values uw. 20 000 and 1000 were
Also include the values of x?, v2 and xu — you'll need them to work out chosen so that the coded
Stews anid Se) data ended up less than
10. (Dividing by 100
uu

Experience, x 2 4 | 5 Pe 10 x = 36 would also have been

[000 sensible — the numbers
Salary, s 23 000]24 000] 24 500 Es
would have been bigger
Coded salary, vu} 3 4 4.5 ae but you wouldn’t have
O25 100 250 any decimals.)
ae 4 16m
ue 9 16 |20.25| 36 |44.89] 81 uw? = 207.14
xu Ge eal 5 128) S36 | 90) | you r
=230.1

Chapter 4 Correlation and Regression

 6 x 33.2 —_ 30.9
= This gives: Sa = Yixu - der
2" = 230.1 — 38-42
2 2
xX ee = 258 —32 = 42
Tip: This is the
Su = Yue — (2H) = 207.14.— 33:2 = 23.4333...
e 2
correlation coefficient
between x and u, but
it must also be the
correlation coefficient . Sa 30.9 — 0.985 (to 3d.p.)
P
ee Sess A223 4333 (
between x and s.

b) Hence write down the value of the product moment correlation

coefficient between x and s. Explain your answer.
=» Since s and u are related bya linear transformation,
the product moment correlation coefficient
between x and s must also be 0.985 (to 3 d.p.) .

Q1 Hint: The
transformation
u =x —20 is of the form
_——
weal Exercise 1.4
b /

but with b = 1. Q1__ Find the product moment correlation coefficient for x and y in the
table below by using the coding u =x — 20 andv=y — 60.

Q2 Hint: The |y |64|61|68|67|63|66|

Pape
transformation u=799
is of the form Q2_ Find the product-moment correlation coefficient for x and y in
a x i;=
the table below by using the coding u = a and v = 10y.
100
but with a = 0.
| * | 400|600|500|300|200|900|1100] 800|
Similarly, the
transformation v = 10y is
LP) s7|o2vtre)0.0led3 a0 epelea)aie aay
of the form v = oe
Q3___ Find the product moment correlation coefficient for s and tin the
i a=0andb=
with = on10° table below by using the coding u = Saest and v = 10(t— 5).

Q3 Hint: The variables

won't always be x and y.
Q4 The weights (w) in kg of a number of athletes were recorded, along
with the number of seconds (s) they each took to run 100 m.
The data is shown in the following table.

Time in secs, s
Find the product moment correlation coefficient for w and s using
the coding uv = w- 60 and v= 10(s— 11).

Q5 Use a suitable coding to find the product moment correlation

coefficient for x and y in the table below.
x | 4008 | 4010 | 4011 4021
300_| 700 | 200
Chapter 4 Correlation and Regression
 Limitations of the PMCC

Correlation is not the same as causation

_ Correlation and causation are often confused, but they’re not the
same thing at all.
= A high correlation coefficient doesn't necessarily mean that a change in
one factor causes a change in the other.
For example, the number of televisions sold in Japan and the number of
cars sold in the USA may be correlated, but that doesn’t mean that high TV
sales in Japan cause high car sales in the USA (or the other way round)
= And the graph on the right
shows the number of tooth- tS oO

fillings a sample of 100 adults

have, plotted against their salary. Qoi)

The graph shows that there’s a

positive correlation between Nhas

‘fillings’ and ‘salary’. As the

number of fillings goes up, so = SD

generally does a person’s salary.

oo
= But... that doesn’t mean fillings pounds)
of
(thousands
Salary

cause someone to earn more, or

that earning more causes fillings.
Number of fillings
= The likely explanation is that
both the number of fillings and a person’s salary are linked to another
quantity — age. As people get older, they tend to get paid more, and the
number of fillings they have also tends to increase.
# Similarly, the TV sales in Japan and the car sales in the USA are probably
both related to the overall strength of the global economy.

The PMCC only shows linear relationships

The product moment correlation coefficient (PMCC) is only a measure of
a linear relationship between two variables (i.e. how close they'd be to a
straight line if you plotted a scatter diagram).

= In the diagram on the right, the

PMCC would be pretty low, but
the two variables definitely look
linked.
= It looks like the points lie close
to a parabola (the shape of an x 2
curve) — not a straight line.

Chapter 4 Correlation and Regression

 2. Linear Regression
Learning Objectives: Linear regression is all about finding the best line of best fit on a scatter
® Be able to determine
diagram. Lines of best fit are really useful — you can use them to predict
which variable is the how one of the quantities will be affected by a change in the other one.
explanatory variable
and which is the Explanatory and response variables
response variable.
When you draw a scatter diagram, you always have two variables.
Be able to find For example, this scatter diagram shows the load onalorry, x (in tonnes),
the equation of a and the fuel efficiency, y (in km per litre). sone
regression line.
" Be able to use a
regression line to = The two variables are negatively correlated.
predict values of the = In fact, all the points lie reasonably close
response variable. to a straight line — the line of best fit.
* Be aware of the ® If you could find the equation of this
problems involved
with extrapolating
line, then you could use it as a model to
beyond the existing describe the relationship between x and y. per fa
(km &aeo
Xi
efficiency
Fuel
litre)
Oe Si Oil ie aOic ans0)
data. Load (tonnes)

Be able to transform
a regression equation
Linear regression is a method for finding the equation of a line of best fit
relating coded ona scatter diagram. Or you can think of it as a method for modelling the
variables to an relationship between two variables.
equation relating the
original variables. Before carrying out a linear regression, you first have to decide which variable
is the explanatory variable, and which is the response variable.
= The explanatory variable (or independent variable) is the variable
you can directly control, or the one that you think is affecting the other.
In the above example, ‘load’ is the explanatory variable.
The explanatory variable is always drawn along the horizontal axis.
= The response variable (or dependent variable) is the variable you think
is being affected. In the above example, ‘fuel efficiency’ is the response
variable. The response variable is always drawn up the vertical axis.

Seley ———$___.
For each situation below, explain which quantity would be the
explanatory variable, and which would be the response variable.
Tip: A place’s latitude is
an angle showing how a) A scientist is investigating the relationship between the amount of
far north or south of the fertiliser applied to a tomato plant and the eventual yield.
equator it lies. = The scientist can directly control the amount of fertiliser she gives
The latitude of the North
each plant — so ‘amount of fertiliser’ is the explanatory variable.
Pole is 90° north, while _ ® She then measures the effect this has on the plant's yield — so
the latitude of the South ‘yield’ is the response variable.
Pole is 90° south.
b) A researcher is examining how a town’s latitude and the number of
North Pole days last year when the temperature rose above 10 °C are linked.
= Although the researcher can’t control the latitude of towns, it would be
the difference in latitude that leads to a difference in temperature, and
not the other way around.
= So the explanatory variable is ‘the town’s latitude’, and the response
South Pole variable is ‘the number of days last year when the temperature rose
above 10 °C’.

Chapter 4 Correlation and Regression

 Exercise 2.1
For each situation below, explain which quantity would be the explanatory
variable, and which would be the response variable.
QI ¢ the time spent practising the piano each week
e the number of mistakes made inatest at the end of the week

Q2 ¢ the age of a second-hand car

e the value of a second-hand car
Q3 ¢ the number of phone calls made in a town in a week
¢ the population of a town
Q4 ° the growth rate of a plant in an experiment
¢ the amount of sunlight falling on a plant in an experiment

Regression lines
Here’s the ‘fuel efficiency v load’ scatter diagram from the previous page.
The regression line (which is what I’m going to call the ‘line of best fit’
from now on) is marked on again. The equation of this regression line is Tip: b is the gradient of
y =a + by, where a and b are numbers to be found. the regression line, and
a is its intercept on the
There are also some dotted lines vertical axis.
showing the vertical distance 10 4
between each data value and oo x
the regression line. 3 8] Xx WwW
Each of these small distances ued
is called a residual (e). aie Kye
The residuals show the errors in 2 :
the model (the regression line) sy .
— they show how the real-life ad a. ekas GAR oeL AG
observations differ from what the Load (tonnes)
model predicts.

Residual = Observed y-vaitie — Estimated y-value

ei ss Yi = (a + bx,)

Linear regression involves finding the equation of the line that minimises Tip: This is why this
the sum of the squares of the residuals, >}e7. kind of regression is
sometimes called
This will mean the regression line is as close as possible to your points. least-squares regression.

The formulas below tell you how to find the regression line of y on x
(the ‘...ofy on x’ part means that x is the explanatory variable,
and y is the response variable).

The equation of the regression line of y on x is: Tip: Remember:

8,=Dy
—See n

yx ia (23)
S re

Chapter 4 Correlation and Regression

 eli)(eam a
The data below shows the load on alorry, x (in tonnes), and the fuel
efficiency, y (in km per litre).
a) Find the equation of the regression line of y on x.
Tip: This is the data
used to draw the graphs semeSe eee
on pages 114 and 115. |y |3 [75[38.3]88]69]5.8]
5.6/6.1] 5.1]47|
= First, you need to find S., and S...
Start by working out the four summations x, )iy, >)x?, and >)xy.
Tip: Loads of calculators It's best to draw a table.
will work out regression
sv [va]Dears |
De [ses]ao[easa]selor[sr[a7]Sy eno
lines for you if you
type in the pairs of data
values (x, y).
= oan can enpaina Sy=sa
But you still need to
know this method,
since in the exam you
[nel 2fron sla fsaSy 6505]
66.8
.3
72x
’
might be given just the
= hen tees aes Se
x y

summations
ae ye and =- 17.914
>)xv, rather than the
individual data values.
And) = S.= ut— (Dux) i723
2 = SA Lay
= 22.081

* So the gradient of the regression line is b, where:

=e : beds

= And the intercept of the regression line is a, where:

Tip: Remember, Dee ee 6b zoe
a=y—br=

= 12.54559...= 12.5 tossigrigyn

= This means that the regression line ofy on x is: y= 12.5 —0.811x

b) Plot your regression line on a scatter diagram.

Tip: You don’t have to = A regression line always goes through the point (x,
use the point (x, 7).

To plot a regression line Hever Lx = 44:3 7.93 'and p = dy BOS S96 %68.
you can choose any two
So the regression line
points to plot. Or you
can draw the line using must go through the
the y-intercept and the point (7.23, 6.68). uth
€) So

gradient.
" By puttingx = 0 into the
It’s a good idea to make equation, you can see the
the points you're plotting line must also go through
for the regression line the point (0, 12.5).
look different from your
actual data points. = So draw the regression line efficiency
Fuel
(km
litr
perOaNWAUDNDWO
through these two points. : : ie e 8 iO pagei2
Load (tonnes)

Chapter 4 Correlation and Regression

 c) Calculate the residuals for: (i) x = 5.6, (ii) x = 6.3.
= (i) When x = 5.6, the residual = observed y-value — estimated y-value Tip: A positive residual
means the data point is
= 7.5 —(12.54559... — 0.81128... « 5:6) above the regression line
= —0.502 (to 3 sig. fig.) for that value of x.
= (ii) When x = 6.3, the residual = observed y-value — estimated y-value A negative residual
= 8.8 — (12,54559... — 0.81128... x 6.3) means the data point
is below the regression
= 1.37 (to3 sig. fig.) line.

In the above example, the regression line was y = 12.5 — 0.811x, where x
was the load (in tonnes) andy was the fuel efficiency (in km per litre).
' These values for a (= 12.5) and b (= —0.811) tell you:
= for every extra tonne carried, you'd expect the lorry’s fuel efficiency to fall
by 0.811 km per litre (since when x increases by1, y falls by 0.811).
= with no load (x = 0), you'd expect the lorry to do 12.5 km per litre of fuel
(assuming it was reasonable to use the line down to x = 0 — see p118).

Tip: Here, d is the

explanatory variable,
A taxi company analyses the fares charged byarival. It looks at 20 so d ‘plays the role of x’,
different journeys and records the distance (d, in miles) of the journey and while f‘plays the role
the fare charged (/, in pounds). The summary statistics are shown below. @t y'

did = 210.1, ) f = 354.8, Sid’ = 2953.53, df = 4619.2

a) Calculate the equation of the regression line of fon d.
* The variables aren'tx and y this time, but d and f— so you need to find
Sands.
d
= S,= af— Lads = 4619.2 — 210-1
¥394.8 — 892.026
© S,= ae— (idy = 2953.53 — 210.12
21 = 746.4295
= So the gradient of the regression line is b, where:
ape oo Ubu ae
ae= 1.20 (to 3 sig.fig,
b= ot = Sie regs = 1.1950...
» And the intercept of the regression line is a, where:
ay aepe
os a — (1.1950...) x 220-1 = 5.1
(to39sig. fig.).
20

= So the equation of the regression line is: f= 5.19 + 1.20d

b) Interpret your values of a and b in this context. Tip: Make sure you
= The value of b tells you that the fare will increase by give your explanations
approximately £1.20 for every extra mile travelled in the context of the
(since when d increases by 1, fincreases by 1.20). question — so here you
need to talk about taxis,
= The value of a tells you that a journey of 0 miles costs £5.19 distances and fares.
— this is a fixed part of the fare that doesn’t depend on how
far you travel (so the taxi’s meter will show approximately
£5.19 before you've even gone anywhere).

Chapter 4 Correlation and Regression

 Exercise 2.2
Q1_ Calculate the cn of the ee line ofy on x for the data
shown in the table below.

yom 9vag. 426: Ea

(You may use S.. = 139.4 and S., = 254.9.)

Q2 The latitude (x, measured in degrees) and the mean annual

temperature (y, in °C) were recorded for a number of locations.
The data is shown below.
Latitude,x | 8 |3019 |41|64|12|60|25|52|39|
[Meanannualtemp. y|27|16|2410] 4 |27]5 |22]9 [12]
a) Draw ascatter diagram of the data.
b)) Calculate the equation of the regression line ofy on x.
c) Interpret your values of a and bin this context.
d)) Calculate the residuals when: (i)x=19 (ii)x = 41

Q3__ The HR department of a large company recorded the salary (P, in

thousands of pounds) of a sample of 10 of its graduate employees,
along with the number of years’ experience (e) they had working for
the company. The summary statistics are shown below.
Die = 92, DP = 264, Ye? = 1072, )\P? = 7082, DieP = 2596
a) Calculate the equation of the regression line P = a + be.
b) Give an interpretation of the values of a and b in this context.

Tip: You can only use a

Interpolation and extrapolation
regression line to predict You can use a regression line to predict values of your response variable.
a value of the response There are two forms of this — interpolation and extrapolation.
variable — not the
explanatory variable. This scatter diagram shows &34
the data from the lorry g8
example that started on — ‘
page 114 — with the fuel = 5
Tip: The data for this efficiency of a lorry plotted
example is on page 116. =
against different loads. 5 2
=o ;
0 2 4 6 8 10 12
Load (tonnes)
In the original data, the values ofx were between 5.1 and 9.8 — the yellow
part of the graph.
= When you use values ofx within this range (i.e. values of x
in the yellow
part of the graph) to predict corresponding valuesof y, this is called
interpolation. It’s okay to do this — the predicted value shoul
d be reliable.
= When you use values ofx outside the range of your origin
al data
(i.e. values ofx in the grey part of the graph) to predict corre
sponding
values of y, this is called extrapolation. It’s best not to do this, or at least
be very cautious about it — your prediction may be very unreli
able.
So here, you'd need to be very careful about using a value
ofx less than 5.1
or greater than 9.8.
Chapter 4 Correlation and Regression
= This is because you don’t have any evidence that the relationship described
by your regression line is true for values of x less than 5.1 or greater than
9.8 — and if the relationship turns out not to hold, then your prediction
could be wildly wrong.

Example eee OF Fe

The length of a spring (y, in cm) when loaded with different masses
(m, in g) is shown in the table below.
Mass,m | 200 | 250 | 300 | 400 | 450 | 500 |
[Length,
y|9.8 | 10.7 | 10.8 | 11.8 | 12.4 | 13.2|
a) Calculate the equation of the regression line ofy on m.
= Extend the table to include m and my, so that you can find Sand S._.

aa
a 24775 — oe 730
. ear
ES
Sam = Sym Pas yee = 805 000 —
2100”
i = 70000

= So the gradient of the regression line is b, where:

Ze0 M aes
Ewe)
b =< = sh q99 = 0.010428... = 0.01043 (to 4 sig. fig.)
* And the intercept of the regression line is a, where:
yeaa) ee)
a= Vy
SS ety aro
= 987 _ (0.010428...) x 2100 7.8
# So the equation of the regression line is: y=7.8 + 0.01043m

b) Use your regression line to estimate the length of the spring when
loaded with a mass of: (i) 370g (li) 6708
= (i) m= 370, soy=7.8 + 0.01043 x 370 = 11.7 cm (to 1 d.p.).
= (ii) m= 670, soy=7.8 + 0.01043 x 670 = 14.8 cm (to 1 d.p.).

c) Comment on the reliability of your estimates in part b).

= m= 370 falls within the range of the original data for m, so this is an
interpolation. This means the result should be fairly reliable.
= But m = 670 falls outside the range of the original data for m, so this is
an extrapolation. This means the regression line may not be valid, and
we need to treat this result with caution.

Chapter 4 Correlation and Regression

 Exercise 2.3 ©
Q1 Hint: For part a), Q1 The equation of the regression line ofy on x is y = 1.67 + 0.107x.
use the fact that this is
the regression line of
a) Which variable is the response variable?
yonx. (And see page to:
b) Find the predicted value of y corresponding
115 if necessary.) (Ve =o (ii = 20

Q2 The regression line ofy on x for a set of data in which x takes

values between 2 and 15 isy = 103 —4.57x.
For each value of x below, estimate the corresponding value ofy
and comment on your estimate.
aa = 4 bax = 20 C) =e

Q3 Data on a large sample of new drivers aged between 17 and 35

was collected. It was found that the number of lessons taken (y)
before passing their driving test could be described by the regression
equation y = 1.4x + 7, where x is the driver’s age in years.
a) Use the equation to estimate the number of lessons a typical
20-year-old could expect to take before they pass their driving
test. Comment on your answer.
b) Use the equation to estimate the number of lessons a typical
50-year-old could expect to take before they pass their driving
test. Comment on your answer.

Q4 A pharmaceutical company is testing a product to eliminate acne.

A volunteer counted the number of spots (s) on an area of skin after
d days of treatment with the product, where d took the values 2, 6,
10, 14, 18 and 22.
The equation of the regression line of s on d was found to be
Ss = 58.8 —2.47d.
a) Estimate the number of spots the volunteer had on day 7.
Comment on the reliability of your answer.
b) She forgot to count how many spots she had before starting to use
the product. Estimate this number. Comment on your answer.
c) The volunteer claims that the regression equation must be wrong,
because it predicts that after 30 days she should have a negative
number of spots. Comment on this claim.

Chapter 4 Correlation and Regression

Linear regression with coded data
You could be asked to find a regression line for coded data. Once you've
Tip: Remember, the
found it, you can easily transform it into a regression line for the original data.
idea behind coding is to
make the numbers easier
to work with.

A company collects data on the age (g) and salary (s) of its senior
managers. The data is coded in the following way:
x=g-50 y= s — 50000
1000
If the regression line ofy on x is given byy = 1.24x + 5.44, find the
regression line of s on g.
= Substitute the expressions defining
x and y into the equation of the s — 50 000 = 1.24(g—50) + 5.44
regression line fory on x. This gives 1000
you an equation involving just s and g.
= Rearrange this so it is in the
form s = a+ bg, where a and s — 50000 = 1240g- 62000 + 5440
b are constants. This is the s = 1240¢ —6560
regression line of s on g.

The annual heating bill (4, in £) for 8 office buildings is shown below,
along with the total floor area (f, in m’) of each building. Tip: If you work out
the means of these
variables, you'll find

5500 | 5900 that f—=2500) and

h = 4000. These are
These results were coded in the following way: the numbers that have
been subtracted in
af 2500 y — 4— 4000 the coding. This will
100 100
make some of the terms
a) Find the equation of the regression line of v on u. disappear when you do
the calculations.
» First make a table of the coded data, along with extra rows for wu’ and wv.
Include the total for each row as well.
SAPS #14 95| 94 el Tip: Yu=D 7 =0
-1]15 | 19 | 21 vy=0 |
This is a result of
1 | 81 [361|576| Siw =1754 | subtracting the means
1818 |
1 [135]361|504| do=w of the original variables
during the coding.
2S = ye = 1818 —2 X90 — 1818

ose = 1754-9= 1754

2 2

= So the gradient of the regression line of v on w is b, where:

b = Sw — 1818 _ 1 936488... = 1.036 (to 4 sig. fig.)
Sa | LAoe Tip: a =0 asa result of
u subtracting the means
) v
= And the intercept is a, where a = V— bu = XY — peut i) of the original variables
during coding.
= So the equation of the regression line of von wis: v= 1.036u
Chapter 4 Correlation and Regression
Tip: c and d are the b) Hence find the equation of the regression line of h on f.
coefficients for the = Substitute the expressions y= 1.036u
eee ie oe ree defining u and v into your f— z= 2500
be different trom
They'll : ti h — 4000 _
Got heeNy regression equation. one 1.036 Cae
coefficients for the
regression line of v on u. _" Then rearrange into the form h — 4000 = 1.036f- (1.036 x 2500)
rE h=c+df.
So h = 1410 + 1.036f

Exercise 2.4
Q1. Aset of bivariate data (x, y) has been coded usingp = x — 7 and
q=y-—50. The regression line of g on p is given by g = 40°+ 2p.
Find the equation of the regression line ofy on x.

Q2_ The regression line of g on p is given by the equation g = —0.9 + 0.1p.

Find the equation of the regression line ofy on x if the following
coding has been used: p = x = 5 20 and g = 10y-3.

Q3 An experiment was carried out to see whether people’s weight

affected how long they could hold their breath for. Each person’s
weight (w, in kg) was coded using the equation x = w— 60. The
number of seconds they could hold their breath for (¢) was coded
using y=¢—45.
The regression line for the coded data has equation y = 17.4 — 0.78x.
Find the equation of the regression line of ton w.

Q4 Mrs Brown put out different amounts of birdseed (s, in grams) each
day, and then counted the number of birds (v) in her garden at noon.
She coded her data in the following way:
X= 759 —2andy=v-7
The summary statistics for the coded data are as follows:
+ = 28, Dy= 124, ox? = 140, xy = 618 andn = 10,
a) Find the equation of the regression line of y on x.
b) Hence find the equation of the regression line of v ons.

Q5__ The time (¢, in hours) that seedlings could survive in water containing
different concentrations of salt (s, in mg per litre) was recorded.
The results were coded such thatx = -& and y = ¢-29.
10
The coded data is in the table below.
Ee bd el aaa
yo, ti oer aro
a) Find the equation of the regression line of y on x.
b) Hence find the equation of the regression line of ton s.

Chapter 4 Correlation and Regression

Review Exercise — Chapter 4
Ql The table below shows the results of some measurements concerning alcoholic cocktails.
Here, x = total volume in ml, and y = percentage alcohol concentration by volume.
150 | 160 | 200 | 240 290
oe ee 20
a) Draw ascatter diagram representing this information.
b) Calculate the product moment correlation coefficient (PMCC) of these values.
c) What does the PMCC tell you about these results?

Q2 For each pair of variables below, state which would be the explanatory variable
and which would be the response variable.
a) © the annual number of volleyball-related injuries
e the annual number of sunny days
b) © the annual number of rainy days
e the annual number of Monopoly-related injuries
c) ® aperson’s disposable income
* a person’s spending on luxuries
d) ¢ the number of trips to the loo per day
e the number of cups of tea drunk per day
e) ¢ the number of festival tickets sold
e the number of pairs of Wellington boots bought

Q3 The radius in mm, r, and the weight in grams, w, of 10 randomly selected blueberry pancakes
are given in the table below.
| [48.0| 51.0 | 52.0 | 54.5 ee an
mee 100 | 105 | 108 | 120 1

aMrind: (i) S, =>)6 oe es (ii) Sw = Dyrw— Lr

The regression line of w on r has equation w = a + br.
Find b, the gradient of the regression line.
or
On Find a, the intercept of the regression line on the w-axis.
es Write down the equation of the regression line of w on r.
Use your regression line to estimate the weight of a blueberry pancake of radius 60 mm.
OeeSeer

=> Comment on the reliability of your estimate, giving a reason for your answer.

r and w are as used in

Q4 The variables P and Q are defined as P=r—5 and O= a! where
Question 3.
Find the regression line of Q on P.

Chapter 4 Correlation and Regression

 Exam-Style Questions — Chapter 4
Values of two variables x and y are recorded in the table below.

pe [2s es eee
|y | 050 | 070 | 010 | 082 | oso [ 036 | 016 | 0.80_|
a) Represent this data on a scatter diagram. (2 marks)

b) Calculate the product moment correlation coefficient (PMCC) of

the two variables. (4 marks)

c) What does this value of the PMCC tell you about these variables? (1 mark)

The following times (in seconds) were taken by eight different runners to complete
distances of 20 metres and 60 metres.
mea ee et Te

Plot a scatter diagram to represent the data. (2 marks)

Find the equation of the regression line of y on x, and plot it on your scatter
diagram. (8 marks)

Use the equation of the regression line to estimate the value of y when:
(jae 35; (11) x = 3.88.
Comment on the reliability of your estimates. (4 marks)

Find the residuals for:

(i) x=3.32 (11) x = 3.27.
Illustrate them on your scatter diagram.
»

(4 marks)

A journalist at British Biking Monthly recorded the distance in miles, x, cycled by

20 different cyclists in the morning and the number of calories, y, eaten at lunch.
The following summary statistics were provided:
Sa 310 880 Sy 188.95 S., = 12 666
a) Use these values to calculate the product moment correlation coefficient.
(2 marks)
b) Give an interpretation of your answer to part (a). (1 mark)
ari
A Swedish cycling magazine calculated the product moment correlation coefficient
of the data after converting the distances to km.
c) State the value of the product moment correlation coefficient in this case. (1 mark)
an

Chapter 4 Correlation and Regression

The equation of the regression line for a set of data is y = 211.599 + 9.602x.
a) Use the equation of the regression line to estimate the value of y when:
i)x= x 12.5
(i) ll) x =x= 14.7.
(ii) (2 marks)

b) Calculate the residuals if the respective observed y-values were y = 332.5

and y = 352.1.
(2 marks)

Ten athletes all trained for the same 10 000 m race. On the day before the race, they each
ran different distances. The distance (d, in km) each athlete ran the day before the race, and
the time (¢, in seconds) that each athlete recorded in the race itself were recorded.
The results were then coded using p = d and q = t— 2367.
The values of p and q are shown in the table below.

Fafeara[nas[ans[0 arfasa)977 er)se

Calculate So7 Ss and Sag"
(You may use >)p’ = 571, >)g° = 548586 and >) pg = 843.) (6 marks)

Find the product moment correlation coefficient between p and q. (2 marks)

Write down the value of the product moment correlation coefficient

between d and t. Explain your answer. (2 marks)

Use your answer to part (c) to comment on how the length of a run on the day
before a race affects the performance of the athletes. (2 marks)

A scientist is investigating the link between the fat content of different brands of burgers
and the price. She measures the amount of fat (x, in grams) in 100 g of each type of burger
(when raw), and calculates the price for 100 g of each burger (y, in pence).
Her data is shown in this table.
Brand | A E FiieG | H |
1) 12 [email protected] | 25
106 | 94 Se |
= 1975, ae= 43290 and >)xy= 6446)
Draw a scatter vee showing this data. (2 marks)

Calculate Sand S.. for this data. (5 marks)

The scientist models the relationship between x and y with the equation
rene)
y=a+bx. Use linear regression to find the values of a and b.

Draw the regression line on your scatter diagram. (2 marks)

iced.
Use the scatter diagram to say which brand of burger appears to be overpr
eH)
Explain your answer.

Chapter 4 Correlation and Regression

 The Normal Distribution

1. The Normal Distribution ov

Learning Objectives: In this section you'll be introduced to the normal distribution.
= Know the shape and Many variables can be modelled by a normal distribution,
properties of the and this can be very useful, as you'll see /ater on.
normal distribution.
Know that The normal distribution
probabilities are
shown by the area The shape of a normal distribution ;
under the normal The distributions of lots of quantities in the real world follow a particular
curve, pattern — with most of the data values falling somewhere in the middle,
Be able to use the and only a small proportion taking much higher or lower values.
correct notation to
describe a normal
distribution. = For example, this histogram
shows the distribution of oct
weights of some hedgehogs.
Most of the weights lie close
to the mean weight — with Frequency
Density
Tip: Remember, similar numbers distributed
continuous variables symmetrically above and below. Mean weight
can take any value
within a certain range.
= A quantity like this can often be modelled by a normal distribution.
= A normal distribution is continuous, so can easily model continuous
Tip: It’s called a random
variable because it takes
variables — such as height, weight, length, etc.
different values with
different probabilities. IfX is a continuous random variable that follows a normal distribution,
We covered discrete you can describe the probability distribution ofX using just two measures
random variables in — its mean, i, and variance, o?.
Chapter 3 — it’s a
similar idea here, but the = Whatever the values of \: and o?, the graph of a normal distribution
variable is continuous. always looks like the curve below.
(RRS RRR Rt re cc ccc cca

= The curve is ‘bell-shaped’.

Tip: The vertical axis is
= There's a peak at the mean, .
labelled f(x) because the
equation of the curve = |t’s symmetrical about the mean — so
is a function of x. It’s values the same distance above and
called the probability below the mean are equally likely. = Sea x
density function, but =p
you don’t need to know = You can see that the shape of the normal curve approximately fits
about these in $1. the
shape of the hedgehog distribution above. The peak at the mean
reflects
the fact that values close to the mean are most likely.
Tip: Fora normal = The width and height of the curve depend on the variance of the
distribution, mean = normal distribution. The three graphs on the next page all show
normal
median = mode. distributions with the same mean (), but different variances
RARERCRA REI senate emo Ee eeannenenenonsemmmnene
(c2)

Chapter 5 The Normal Distribution

 = The larger the variance the wider
the curve, so graph 3 has the
largest variance and graph 1 has
the smallest variance.
= The total area under the curve is
always the same, so a wider curve
needs to have a lower height.

The area under a normal curve

The area under a normal curve shows probabilities.
= The total area under the curve
represents the total probability of
the random variable taking one of its
possible values. And since the total
probability is 1, the total area under Tip: You'll see how to
find and use areas under
the curve must also be 1.
the normal curve over
= The probability of the variable taking the next few pages.
a value between two limits is the area
under the curve between those limits.

There are some facts about the area under the curve
that apply to all normal distributions.
Tip: You don’t need to
= 68% of the total area lies within +1
learn these facts, but
standard deviation (+o) of the mean. they should give you an
® 95% of the total area lies within +2 idea of how values close
standard deviations (+20) of the mean. to the mean are much
more likely than those
= 99.7% of the total area lies within +3 further away.
standard deviations (+30) of the mean.

So 68% of observations are within +o of the mean, 95% of observations are

Tip: Remember...
within +20 of the mean and 99.7% of observations are within +30 of the mean.
standard deviation is
the square root of the
variance — so it’s a
Describing a normal distribution measure of dispersion
(how spread out values
lf a continuous random variable X is normally distributed are from the mean).
0?)
with mean w and variance o?, it is written like this: ———> | X ~ N(u,
‘N’ stands for ‘normal’ and ‘~’ is short for ‘is distributed’.

= So going back to our hedgehog weights on the previous page,

we could define a random variable, W ~ N(p, 07),
where W represents hedgehog weight.
= Here, , would represent the mean weight of the hedgehogs and
o would represent the standard deviation of hedgehog weights.
l distribution Z,
‘ The most important normal distribution is the standard norma
about the
which has a mean of 0 and a variance of 1. There's a lot more
standard normal distribution on the next few pages.

Chapter 5 The Normal Distribution

 2. The Standard Normal
Distribution, Z |
Learning Objectives: The most important normal distribution is the standard normal distribution.
= Be able to use tables The ‘special’ random variable Z follows a standard normal distribution — Z is
to find probabilities called the ‘standard normal variable’ In this section you'll see how to use
for the standard tables to find the probability that Z takes a value in a given range.
normal distribution, Z.
= Be able to use The standard normal distribution
tables to find values
for z, when given The standard normal distribution, Z,
probabilities. has mean 0 and variance 1. ——————- | Z ~ N(O, 1)

Below is a graph of the standard normal distribution.

As you'd expect, the curve is symmetrical
about the mean, 0.

Remember that areas under the curve

show probabilities. The shaded area here
shows the probability that Z takes a value
that’s less than or equal to z. So to find this
probability, we need to work out the area.

Working out the area under a normal distribution curve is usually difficult.
But the reason why Z is so important is that there are tables that list areas
under the curve for N(O, 1). The main table you'll use (see page 150) shows
values of the normal cumulative distribution function, ®(z). These are the
areas under the curve to the left of z (the probability that Z < z) for different
values ofz —1.e. ®(@) = P(Z < z).

So you can often find P(Z< z) by looking up the value of z in the table
Tip: The table only gives
and reading off the value for ®(z). And since Z is a continuous distribution,
P(Z < z) forz = 0.
P(Z = z) = 0, which means that P(Z < z) and P(Z < z) are the same thing.
So you can interchange the < and < signs — i.e. ®(z) = P(Z < z) as well.
Tip: For any continuous
random variable x,
P(X= a) = 0 for any
value of a.
That’s because the area Find the following probabilities.
under a graph at a single
point is zero.
a) P(Z < 0.64)
This is a nice, straightforward one — all you have
to do is look up z = 0.64 in the table for ®(z).
Tip: The table for ®(z) is So, P(Z < 0.64) = 7389
on page 150.

Always write down the b) P(Z < 0.1)

full value from the table. Z \s a continuous variable, so P(Z < 0.1) is just the same as P(Z < 0.1).
So again, you just look up 0.1 in the table for ®(z)
So, P(Z< 0.1) = 5398

Er Chapter 5 The Normal Distribution

OK, so finding the probability that Z is less than z, where zis positive, is easy.
And finding the probability that Z takes a value greater than (or greater than
or equal to) z is only slightly trickier.
Using the fact that the total area under P(Z > z) =1-P(Z<z)
_ the curve is 1, we get this definition. ——> Tip: And we also get:
= 1- &() A 2 a) | AV 7)
= 1-()
: (since P(Z <z) = RiZsz))
Find the following probabilities.
a) P(Z > 0.23)
P(Z> 0.23) = 1 — P(Z < 0.23)
= 1-—0.5910 «—— PZ 0.23) = 0.5910 Tip: Drawing a sketch
always helps to make
= 0.4090 things clearer.
b) P(Z= 1.15)
Pee to 1 PZ < 1.15) ‘
= 9 2 ORAS Se Riza AS) = 0.8749

= (01251

Now then, things start to get a bit trickier when z takes a negative value.
To work these out, you need to use the symmetry of the curve. The best thing
to do is to start by drawing a sketch and shading the area you want to find.

Example 3
Find the following probabilities. f@)
a) P(Z>-0.42)
= Shade the area you want to find. ———_
= Because z is negative, you can’t look it up in the aa ty

table — so the method used in Example 2 won't work.

Instead, use symmetry to shade an area of the same f(z)
size involving a positive value of z.
This is the area to the left of z = +0.42. ———_
P(Z
>-0.42)=P(Z<0
weSo .42). >:
And looking up 0.42 in the ®(z) table you get:
P(Z > -0.42) = P(Z< 0.42) =@aGze f(z)

b) P(Z < -1.94)

= Shade the area you want to find. ———_ N

= Again, use symmetry to shade an area of ae

the same size, but involving a positive fe)
value of z instead of a negative one. ieee ee
This is the area to the right of z = +1.94. —————_> alley suyeae ig pent
. ea an area that you can
» SoP(Z< =| 94) = P(Z= 1 9A).

= You still can’t get your answer from the table directly, ss up in the ee
but using the fact that the area under the curve is 1,
P(Z> 1.94) =1-P(Z< 1.94).
aca Fey 94) =PZ = 1.04) 21 PZ < 1,94) =0.9738
P(Z < 1.9 4)
al 009738
= (00262

Chapter 5 The Normal Distribution

 In the next example, you're asked to find the probability that Z takes a value
between two limits. You can do this by subtracting one area from another.

Find the probability that 0.12 < Z < 0.82.

= Shade the area you want to find. ———————>
= To find this area, you need to find the area to
the left ofz = 0.82, then subtract the area to
the left of z= 0.12. That will leave you with
the area between the two values.
= So, P(0.12 < Zs 0.82) =PiZs 0.82) -PiZ S02)
= 0.7939 — 0.5478
= 0.2461

Exercise 2.1 = ss

Q1__Use the table of the normal cumulative distribution function, ®(z), to

find the following probabilities.
a) PZ <1.87)-—b) P(Z<039)— cl PZ 0.99) —Gaat 2a

Q2___Use the table of the normal cumulative distribution function, ®(z), to

find the following probabilities.
a) \P(Z> 2.48) > by -P(Z>"0'85) Mc) UPIZ S23)" Sd) ePizew 17)

Q3___By using sketches and the table of the normal cumulative distribution
function, ®(z), find the following probabilities.

a) P(Z>-3.35) b) P(Z>-0.24) c) P(Z>-1.21)

d) P(Z <-0.62) e) P(Z<-1.14) f) P(Z<-2.06)

Q4 Work out the following probabilities.

ARE
= ZS
= 18) b) 4P(0.7.6 <7?)

Q4 Hint: For parts g) c) P(-1.45 <Z< 0.17) d) PG2.147

<<1 65)
and h), use symmetry
e) P(-1.66 < Z < 1.66) f) P(-0.34 < Z< 0.34)
to find an identical area
for which both limits are g) P(-3.25 <Z<-—2.48) h) P(-1.11 <Z<-0.17)
positive.

Chapter 5 The Normal Distribution

 Using the tables in reverse
So you know how to find probabilities using the (z) table. The other sort of
question you need to be able to answer is where you're given a probability,
e.g. P(Z < z) = 0.5871, and you have to find z. How you go about this
. depends on the information you're given.
If you're given the probability that Z is less than a value of z, P(Z < z),
it's usually best to see if you can find what you need in the table for &(z)
(the normal cumulative distribution function). But you need to read the
table ‘the other way round’.

If P(Z < z) = 0.9554, then what is the value of z?

» Start by drawing a sketch, showing what you know. ——> Tip: Always check that
= OK, so you need to find z for which ®(z) = 0.9554. the answer seems
reasonable. The total
» Using the table, read through the values of ®(z) until nee A probability is 1, so for
you findfin 0.9554. ®(z)(2) =
= 0.9554
0.9554 for z #
i 1.70. 0) + __ The table for an area of 0.9554, z will
SO; if IA Z) = 0.9554, then Z>= EE AO (z) ison p150 be positive and quite far
to the right.

If P(Z < z) is less than 0.5, then z will be negative, and these values aren’t
listed in the table. But you can still use the table — you just need to doa bit
of thinking and sketching to decide what to look up.

= The area, a, to the left of z is less than 0.5.

You can see from the graph that there’s an
area of the same size to the right of the oe
value —z (where ~z is positive). >
Tip: If P(Z<z) <0.5,
= Now you can draw a second graph Z 0 —@ then z will be negative.
showing the area to the left of -z. y
This is the area 1 - a.
= The area 1 —a is greater than 0.5, so you
can look it up in the ®(z) table to find the
Area > 0.5
value of —z.
= Finally, multiply —z by —1 to get z.

If P(Z < z) = 0.2611, then what is the value of z?

= First, draw a sketch, showing what you know, ——>
and mark on the equivalent area to the right of -z.
= Subtract 0.2611 from 1 to get the area to the left of —z. rep
10.261) =.0.7389.
~ Tip: Again, check the
= Draw another sketch showing the new area. ——-> answer seems OK.
= Look up ®(z) = 0.7389 in the table. Riaz) aro SOE
@(z) = 0.7389 forz = 0.64, so -Z = 0.64. should be negative.
So, z=-1 x 0.64 = -0.64. So, if P(Z< z) = 0.6211, then z=—0.64

Chapter 5 The Normal Distribution

 If you're given the probability that Z is greater than a value of Z, P(Z-S 2);
it’s usually best to see if you can use the table of percentage points of the
normal distribution (see page 151).
= The percentage-points table gives the value of z
for some probabilities, p, wherep = P(Z > 2).
= So this time you start off with the probability
that Z is greater than a value of z, and look up
the value of z in the table.
= However, depending on the value of p, you
Tip: The values of p in might not be able to find what you want from
the percentage-points the percentage-points table. In which case, use
table are all nice,
‘round’ decimals.
the ®(z) table and the symmetry of the curve.

elec eee

a) If P(Z > z) = 0.15, then what is the value of z?

= Start by drawing a sketch, showing what you know. ——>
=» Try looking upp = 0.15°
in the percentage-points table «————_ See p151 oz
From the table, ifp = 0.15, then z = 1.0364.
So, if P(Z>z) = 0.15, then = O364

iy— Find z if P(Z = z) = 0.01.

= Start by drawing a sketch, showing what you know. ——>
Tip: Remember, > and =
mean the same for a * Try looking up p = 0.01 in the percentage-points table.
continuous distribution. From the table, ifp = 0.01, then z = 2.3263. 0 Zz
So, if P(Z> z) = 0.01, then aaa ae

c) If P(Z > z) = 0.7734, then what is the value of z2

= Start by drawing a sketch, showing what you know. —>
= p isa ‘horrible’ decimal, so the percentage-points
table can’t help you. And sincep > 0.5, you know that z is negative.
= So use symmetry to shade an area of the same size
that you can look up in the ®(z) table.
This ts the area to the left of 2. ————————————-
= From the table, 6(z) = 0.7734 forz = 0.75.
So, -Z = 0.75 and z =-1 x 0.75 =-0.75. WK.
= So, if P(Z > z) = 0.7734, then = =e7e

Okay, so the ®(z) table contains values for P(Z< z) and the percentage-points
table contains values for P(Z > z). But it really depends on what sort of
decimal the probability is as to which table you should use to find z.

Chapter 5 The Normal Distribution

 Examples

a) If P(Z < z) = 0.9, then what is the value of z? fe)

= Start by drawing a sketch, showing what you know. ——>
= P(Z<z) is anice, ‘round’ decimal, so you'll probably Tip: The ®(z) table
Ie doesn’t contain
need to use the percentage-points table in some way. URE "i P(zZ) = 0.9.
= IfP(Z<z) =0.9, then P(Z > z) = 0.1. So try looking up
And since P(Z< z) > 0.5,
p = 0.1 in the percentage-points table. you need to subtract it
From the table, if p = 0.1, then z = 1.2816. from 1 before using the
So, if P(Z < z) = 0.9, then Z=1.2816 pe pone table.

b If P(Z < z) = 0.05, then what is the value of z?

_—

* This is a nice, ‘round’ probability, so it looks like you're going to need

the percentage-points table. But you can’t use it directly because you
have a probability that Z < z (not Z > z).
= Start by drawing a sketch, showing what you know. ——»
» This shows that the area to the right of —z is 0.05, ae
so try looking upp = 0.05 in the Tip: You don’t always
have to do a sketch if
percentage-points table. roe
you feel you don’t need
From the table, p = 0.05 for z = 1.6449. to, but they do help you
So, - = 1.6449 and z =-1 x 1.6449 = -1.6449. to avoid mistakes.
# So, if P(Z <z) = 0.05, then z =—1.6449

c) If P(Z > z) = 0.0392, then what is the value of z?

= This is a fairly unpleasant decimal, so you'll need to
use the table for ®(z) this time.
Start by drawing a sketch, showing what you know. —>
= 0.0392
=
* Since P(Z > z)= 0.0392, you know that zZ
P(Z < z)= 1 —0.0392= 0.9608. So look this up in the ®&(z) Nile
From the table, ®(z) = 0.9608 for z == 1.76,
= So, if P(Z>z) = 0.0392, then 27a

The final type of example we need to go through is when you‘re given the
probability that Z takes a value between two limits, and you're asked to
find the missing limit.

Examples
a) If P(0 < Z <z) = 0.4452, find the value of z.
= Start by drawing a sketch, showing what you know. ——>
= If you find P(Z < z), you can use the (2) table to get z. N

oy = 0.4452 + P(Z < 0) ~+~_ Smear eos

Hell wel 2.0)
PZ
= 0.9452
=» Using the table, ®(z) = 0.9452 forz =.1,.60.
= So, if P(0<Z<z) = 0.4452, then z= 1.60

Chapter 5 The Normal Distribution

 b) If P(0.8 < Z < z) = 0.20, find the value of z.

= Start by drawing a sketch, showing what you know. ——>

= If you find P(Z < z), you can use the ®(z) table to get z.
P(Z <z) = P(Z< 0.8) + P(0.8<Z<z) , cee
= 0.78814 0.20
=O: 9881 _—- Using the ®(z) table
= Using the table, &(z) = 0.9881 forz = 2.26.
« So, if P(0.8 < Z <z) = 0.20, then 2a

Exercise 2.2

Q1__Use the table of the normal cumulative distribution function, ®(z), to

find z, given each of the following probabilities.
a) PZ = 7) = 06577 b) P(Z<z)= 0.8264
c) P(Z<z) = 0.3783 d) P(Z<z)=0.004
Clue Ze=Z) = 756 f) P(Z>z)=0.9441
g) P(Z>z)=0-4801 h) “P(Z> 2) =020951

Q2__Use the percentage-points table to find z,

given each of the following probabilities.
a) P(iZ>z)= 03005 b) PZ Sz) 02
Oo. PZ=z)= d)ePiZ<2) =0:85
e) PZ <Z)= HN RZ <2) = 0.15
sg) PZ eZ) = h) PZ Sz) = 0.99

Q3 Hint: If you subtract

the given area from 1 Q3__ Find the value of z, given each of the following probabilities.
and divide by 2, you get a) Plz<Z<z)=0.599
the area = z (and < -z).
bb) PUr<Z<2)=0.9426
And that means you can c) Pez<Z<z)=04 d) Pi-z<Z<z)=0.98
easily find P(Z < z).
Q4 Find the value of z, given each of the following probabilities.
Q4 Hint: Remember,
a) PIO Z <2) 0.3869 b) PO <Z<z)=0.4854
P(Z< 0) and P(Z> 0)
both equal 0.5.
~ Q5__ Find the value of z, given each of the following probabilities.
Q5 Hint: One way of a) PZ<Z< 0) =02420 DRG 2.0) =). 1443
doing these is to flip the
interval across the line
Q6 Find the value of z, given each of the following probabilities.
= 0),
a) Pl.5 <Z<z)= 0.0406 by) P5817 2c) =10, 0691
C) PEl3 <2 <z)=0.871 “dy PEO 542725 eaea,
e) Pz2<Z<027)= 054585) Peaz2 | .25) = 0.0949

Chapter 5 The Normal Distribution

3. Normal Distributions and
Z-Tables
All normally distributed variables can be transformed to the standard normal Learning Objectives:
variable, Z. This is really useful, because it means that you can use the Z-tables
= Be able to transform
to find out information about any normal distribution.
normal variables to
the standard normal
Converting to the Z distribution variable, Z.
_Any continuous random variable, X, where XY~ N(u, 02), can be transformed Use Z-tables to find
to the standard normal variable, Z, by: probabilities for
normal variables.
= subtracting the mean (1), and then Use simultaneous
= dividing by the standard deviation (0). equations to find the
mean and variance of
a normal distribution.
If ¥ ~ N(u, 02), then *=! = Z, where Z ~ N(O, 1)
Tip: Remember, the
Once you've transformed a variable like this, you can use the Z-tables. standard deviation is the
square root of the
variance.
Here’s a sketch of the distribution ofX,——————> f(x) 4 X~ N(20, 9)
where X ~ N(20, 9). w=20,0=3
we

To transform X into Z, you need to

transform all values ofX (called x) into
values of Z (called z), using the formula:
ja 140 Wz 20) 23" 26) 29

|
For example: f(z) Z ~ NO, 1)
x= 23. becomes
2= a =

To find a probability for X, rewrite it as a Tip: Here the z-value

probability for Z, then find the eo fopoudine has been rounded to 2
ee
vA
area under the Z curve. Le ere decimal places so we
can use the ®(z) table.
E.g. P(X<25) = P(Z < 22-529) = P(Z< 1.67) = 0.9525 +— From %) This rounding means
Ky Kz table that the probability we
get isn’t quite as precise
as if we’d been able to
use the full value.

X~ N(5, 16), find P(X< 7).

a) If Tip: You'll usually be
p=5,0=1\16=4 given the variance,
= X to Z.
Start by transforming which means taking the
72) =P(Z<0.5)
PX<7)=P(Z< 3 fe) square root to find the
standard deviation.
= Draw a sketch showing the area you need to find. ———>
= Look up z= 0.5 in the table for ®(@).*~ oo, p150 A ]
P(Z< 0.5) = 0.6915

Chapter 5 The Normal Distribution

 b) If.X ~ N(5, 16), find P(X> 9).

= Start by transforming Xto Z. a

Tip: You might not need

ea ap ety pur>9)= P(2>255) =z)

- = Drawa sketch showing the area you need to find. ———>
Oa eas [|
a PiZ>1)=1-PIZ<1)

= 1 -0.8413
= HEI Oe

c) If X ~ N(5, 16), find P(5 < X < 11).

= Start by transforming Xto Z.

<Z<
1-2) =PO<Z<15)
PI5<X<11)=P(252 ave
= Draw a sketch showing the area you need to find. ——>
= To find this area, you need to find the area to the left of
z = 1.5, then subtract the area to the left ofz = 0. Om SES
= So,P(0<Z< 1.5)= PZ <5)
—PGs)
S093 526.00
= 10}4332

As you saw in the previous section, you often need to use the symmetry
Tip: See page 129. of the curve to answer questions on the normal distribution.
sone oe We'll be using the same methods here as in Section 2, but we need to
do the extra step of transforming X to Z first.

Examples

If X ~ N(102, 144):
a) Find P(X> 78).

= Start by transforming X to Z. f(z)

PLY > 78) = P(Z > B—102) - pz > -2.0)
" Draw a sketch showing the area you need to find. —*>
« Use symmetry to shade an area =a 4
of the same size
that you can look up in the ®(z) table.
This is the area to the left of z= +2.0. ———————_» |
= SoOR(Z > —2.0) = PiZ= 2.0).
And looking up 2.0 in the ®(z) table you get: 0-20) oes

Chapter 5 The Normal Distribution

 b) Find P(X< 98).
= Start by transforming
Xto Z.
a To 2 d.p.
P(X< 98) = P(Z < 8102) = PZ 0133)
= Drawa sketch showing the area you need to find. ——> Tip: Round z values to 2
* You can’t look up a negative value ofz in the table, decimal places so you
=0.330 can look them up in the
SO again, use symmetry to shade an area of the f(z) table.
same size. This is the area to the right of z = +0.33. ——>
aso P(Z-<—0.33) = P(Z 0.33).
SNOW, PZ 2 0.33) = 1 =P 2< 033)
= 1-0.6293 sacs
= 0.3707

c) Find P(84 < X < 106).

= Start by transforming X to Z.
Bad ad 06) iP) 02 7 106
= 102
( 06) P( 12 ana 2 ) f(z)
=ri—-1.5 =Z 50333)
* Draw a sketch showing the area you need to find. ——>
» To find this area, you need to find the area to the left (ba z
of z = 0.33 and subtract the area to the left of z = -1.5.
= 50, PEI =Z < 033) = P(Z =< 0.33) —P(Z< =-1.5)
* You can find P(Z < 0.33) directly from the table, but
P(Z < -1.5) takes a bit more working out.
PiZ = 1.5)= PZ > 1.5)= 1 —PZ< 1.5)
= 1-0.9332
= 0.0668
= So, P(Z< 0.33) — P(Z< -1.5) = 0.6293 — 0.0668 = ‘0.5625

You also need to be able to find values, when you're given probabilities.
For example, find a, given that P(X< a) = 0.6. Tip: You can look up
to Z.
= As usual, the first step is to transform _X probabilities in the ®(z)
table, or the
X~ N(u, o”), the equation P(X< a) = 0.6 becomes P(Z
If eeoO )= 0.6. percentage-points table.
= Then you use the same methods as on pages 131-134, and you end up See p131-134 fora
with an equation in a to solve. reminder and some
examples.

Examples —— tance

a) X ~ N(85, 25). If P(X < a) = 0.9192, find the value of a.

* Start by transforming X to Z.
BT 4 a—8>) = 0.9192
= Draw a sketch to show the information. ———————_>
= Use the &(z) table to find z for which ®(z) = 0.9192.
85 — 1.40 See p150
(z) = 0.9192 forz= 1.40, so 2
= Now, just solve for a: Equation in a.

a— 83 -1.44-85=7> a= 92
Chapter 5 The Normal Distribution
 b) X ~ N(85, 25). If P(X< b) = 0.0179, find the value of b.
= Start by transforming Xto Z.
bee ISL aa ny = 0.0179
= Draw a sketch to show the information — the area is less
than 0.5,s0 b a 85 must be negative. Now mark on the
Tip: The equivalent : (ph Dez OD,
preasie to theanehtc) equivalent area to the right o ie
5 nea
Fe * You can see that P(Z 2 Mt = O.0079:
: : b= 0 S5=b 5” @
Clee b
ee So P(Z < 88
=4)
D
= 1 - 0.0179 = 0.9821
5
= Using the ®(z) table, &(z) = 0.9821 for z = 2.10, so 85 = D340,

= Sosolving for b, 8298

5
=2.1085-b=1055 b=745

Hint: Round z values to

Exercise 3.1
2 d.p. so you can look
them up in the table. OT ie NAO 25S) tind a WH P(X< 50) b) P(X< 43 43)

Q2. If X~ N(24, 6), find: a) — P(X> 28) Db) PUCS 25)

Q?2 Hint: The standard
deviation doesn’t have Q3 If X ~ N(120, 40), find: a SS P(X > 107) by" POSS 415)
to be a whole number. |
Q4 If X ~ N(17, 3?), find: Ay Sea AX oy b) POY <2)
Q4 Hint: You'll often ae
see the variance written Q5_ IfxX~ N(50, 5%), find:
as a number squared. a) POLS = 63) by PG eX = 66)

Q6 If X~ N(0.6, 0.04), find:

a) P(0.45 <X<0.55) b) P(0.53 <X<0.58)
Q7 If X~ N(260, 152), find:
a) P(240 < X < 280) b) P(232 <X< 288)
Q8 X-~ N(7O, 16) .
a) Finda if P(X <a)=0.9938 _b) Find dif P(X <b) = 0.7734
Q9 X~N(95, 25)
a) Find m if P\XY>m)=0.0102_b) Find rif PX¥> 1) = 0.2296
Q10 X~ N(48, 100)
a) FindcifP(¥<c)=0.1251 —_-b) -Findd if PX <d) = 0.0096
QTieY= Ni73 06)
a) Find wif P(X¥>w)=0.9177 _b) Find kif P(X> = 0.6664
Qi2 X ~ NWS, 0.25): Find aif PUG a=2 xen8 aa) 0-800;

Qi3 X~ N(170, 40). Find tif P(170 < X < 4) =0.377.

Ql4 X~ N(98, 225). Find v i PAO7-6e2.% = yO) lo,

Chapter 5 The Normal Distribution

The normal distribution in
real-life situations
Now it’s time to use everything you've learnt about normal distributions
‘to answer questions in real-life contexts. These are the kind of questions
that usually come up in exams.
You always start by defining a normally-distributed random variable to Tip: The first step is
represent the information you’re given. Then you use the usual methods always to define the
to find out what you need to know. variable.

eT om a erick adie Soiree

A machine which fills boxes of cereal is set so that the mass of cereal
going into the boxes follows a normal distribution with mean 766 g and
standard deviation 8 g.
a) Find the probability that a randomly selected box of cereal contains
less than 780 g of cereal. Tip: So the variable here |
« First, define a random variable to represent the mass of cereal in a box. is the mass of cereal ina |
box.
If X represents the mass of cereal in g, then X ~ N(766, 64).
* Next, turn the question into a probability for xX. _@
So you want to find, P(X < 780).
= Now you can transform _Xto Z in the usual way.
PUK< 780) = P(Z < 2B0—766) ;,
= (2 <ale/9) ‘ Tip: See p150 for the
®(z) table.
= Draw a sketch showing the area you need to find. ——>
» Using the ®(z) table, P(Z < 1.75) = 0.9599 4
Tip: Always check the
» So P(a random box of cereal contains less than 780 g) = 0.9599 answer seems sensible.
Using the rules on p127,
you know that 97.5% of
b) Find the probability that a randomly selected box of cereal contains values are less than
2 x o above the mean
—

between 780 g and 790 gof cereal.

(and 2 x 8 + 766 = 782).
= Again, turn the question into a probability for X. So this answer seems
So you want to find, P(780 < X < 790).
about right.
= Next, transform X to Z.
P(780 < X < 790) = p(BO—766 Ze 290 708)
Sy 5 2 3.0)

= Draw asketch to help you see this area. ———————

This is the area to the left of 3.0, minus the area to Z

the left of 1.75 (which you found in part a)).

O 17s 30

meso) Pit.7> = 2 < 3.0) = PZ< 3.0) _PiZs j275)

From part a)
= Using the ®(z) table, P(Z < 3.0) = 0.9987.
So, P(Z< 3.0) — P(Z< 1.75) = 0.9987 — 0.9599 = 0.0388
# This means:
P(a random box of cereal contains between 780 g and 790 g) = 0.0388

Chapter 5 The Normal Distribution

 Here’s another example. In part b), we need to use the symmetry of the curve
to work out what area to look up in the table.

_ The times taken by a group of people to complete an assault course are

normally distributed with a mean of 600 seconds anda variance of
105 seconds. Find the probability that a randomly selected person took:
a) more than 620 seconds

= Start by defining a random variable to represent the time taken.

If X represents the time taken in seconds, then X ~ N(600, 105).
= Next, turn the question into a probability for X.
So you want to find P(X> 620).
= Now you can transform _X
to Z.
620
PW > 620) = P{/Z > = — 600
(X> ) ( AOS

= P(Z> 1.95) +——_______ Jo 2 d.p.

= Draw a sketch to show the information. —---—>
= So, PZ S91.95) Sal — PiZO)

='(00256
» This means:
P(a randomly selected person took more than 620 seconds) = 0.0256

b) fewer than 575 seconds

= Again, turn the question into a probability for X.
So this time you want to find P(X< 575).
=» Next transform _X to Z.
2/5.— 600
Pix 5/5) =')( 7 <2
( V¥105
= P(Z< -2.44) «—_______ Jo 2 d.p. f(z)
=" Drawa sketch to show the information. You can’t ——»
look up a negative value ofz in the table, so use
symmetry to shade an area of the same size.
"= OK so, P(Z < -2.44) = P(Z > 2.44) ata See
=1-P(Z< 2.44)
='1:= 0.9927 = 00073
* This means:
: P(a randomly selected person took fewer than 575 seconds) = 0.007
3

Real-life normal distribution questions also ask you to find values

,
when you're given probabilities.
Remember there are two tables you can use to look up probab
ilities
— the table of the cumulative distribution function, ®(z),
and the
percentage-points table.

re
ane
Beam Chapter 5 The Normal Distribution
The forces needed to snap lengths of a certain type of elastic are normally
distributed with ,.= 13 N and o = 1.8N.

a) The probability that a randomly selected length of elastic is snapped by

a force of less than a N is 0.7580. Find the value of a.
Start by defining a random variable to represent the force needed. Tip: It’s neater to write
the variance as 1.87,
If F represents the force needed in N, then F ~ N(13, 1.82).
rather than 3.24.
Next, turn the question into a probability for F.
So you know that P(F < a) = 0.7580.
Tip: Use a sensible
Next, transform F to Z.
letter for the variable —
PF <a)=P(Z<4 ve3) = 0.7580 it doesn’t have to be X.

Draw a sketch to show the information. ————>

Use the ®(z) table to find z for which ®(z) = 0.7580. Tip: P(Z <z) isn’ta
‘nice’ decimal, so you'll
®(z)= 0.7580
forz=0.70, so a1 - 0.70 need to use the ®(z)
Now, just solve for a: table.

41-07 Sa-13=1.265 oT OG

—
Find the range of values that includes the middle 80% of forces needed.
It’s difficult to know where to start with this one. So it’s a good idea to
sketch the distribution of F, to show the range you need to find.

Ay (1 —0.8)+2=0.1

K Middle 80% of values

gay [Salisticn lie between a and b.

Now you can write 2 probability statements — 1 for a and 1 for b.

Pi@= a) = 0.1 and Pi> b) =0.1.
Next, transform F to Z.
z>b=13
pir <a)=P(z<2=13) - 0.1 and PF > b)= P(Z> 2-19)= 0.1

Draw another sketch to show this information.

So, now you need to use the tables to find a and b.
The easiest one to start with is b, because you can use A Tip: P(Z> z) is a ‘nice’
the percentage-points table to find the value of z for ot decimal, so you'll
seam; mia Ta need to use the
> z) = 0.1.
which P(Z
percentage-points table.
z = 1.2816. a
From the table, p = 0.1 for P

So, pe Se iO
Tip: The range is
13
And using symmetry, “-g= = —1.2816 = a= 10.69 symmetrical about z = 0,
ean Oe eens
Sica 7.8
So the range of values is 10.69 N to 15.31 N

Chapter 5 The Normal Distribution

 Exercise 3.2

Q1 The lengths of worms in a certain area are found to follow a normal

distribution, with mean 8.4 cm and standard deviation 3.1 cm.
a) What is the probability that a randomly selected worm is shorter
than 9.5 cm?
b) What is the probability that a randomly selected worm is longer
than 10 cm?
c) What is the probability that a randomly selected worm has a
length of between 5 cm and 11 cm?

Q2 The lengths of time taken by a group of blood donors to replace their

red blood cells are modelled by a normal distribution with a mean of
36 days and a standard deviation of 6 days.
a) It takes Edward 28 days to replace his red blood cells. Find the
probability that a randomly selected donor from the group takes
less time than Edward to replace their red blood cells.
b) 6.3% of the group take longer than Bella to replace their red
blood cells. How long does it take Bella?

Q3 The ‘personal best’ times taken by athletes at a sports club to run

400 m are known to follow a normal distribution with a mean of 51
seconds and a standard deviation of 2.1 seconds.
a) Gary’s ‘personal best’ time is 49.3 seconds. What percentage of
the athletes have a slower ‘personal best’ time than Gary?
b) The athletes with ‘personal bests’ in the top 20% of times are
selected for a special training programme. What time do they
have to beat to be selected for the programme?

Q4 A particular type of toy car uses two identical batteries. The lifetimes
of individual batteries can be modelled by a normal distribution with
a mean of 300 hours and a standard deviation of 50 hours.
a) What is the probability that a battery of this type lasts less than
200 hours?
b) What is the probability that a battery of this type lasts at least 380
hours?
Q4 Hint: See p68 for c) Stating any assumptions you make, find the probability that both
a reminder of finding of the batteries in a car last at least 380 hours.
probabilities of multiple
events. d) The probability that a randomly selected battery lasts more than
160 hours, but less than A hours, is 0.9746. Find the value of h.

Q5 The masses of the eggs laid by the hens on farmer Elizabeth’s farm
are assumed to follow a normal distribution with mean 60 g and
standard deviation 3 g.
a) The probability that a randomly selected egg has a mass of at least
60 — m grams is 0.9525. Find the value of m to the nearest gram.
b) Farmer Elizabeth keeps the lightest 10% of eggs for herself and
uses them to make sponge cakes. Find the maximum mass of an
egg that could end up in one of farmer Elizabeth’s sponge cakes.

Chapter 5 The Normal Distribution

 Finding the mean and standard deviation
of a normal distribution
Remember that any normally distributed variable, X ~ N(u, 02), can be
" transformed to the standard normal variable, Z, by subtracting the mean, 1,
and dividing by the standard deviation, o.
So X and Z are linked by the equation: Z = x = Ee
So far you've used this relationship to find probability facts for X, when the
mean and standard deviation have been known. But you can use exactly
the same approach to find the mean and standard deviation when they’re
unknown... as long as you already know some probability facts.

If the random variable X ~ N(u, 4) and P(X< 23) = 0.9015, find i.

* Okay, start by transforming the probability you're given for X into a

probability for Z. The mean is unknown, so just leave it as . for now.

Py 23)= P(Z< Zo 2
ph
)=0.9015
f(z)

# Draw asketch to show the information. ————————

= Now, if you use the ®(z) table to find z for which fs
®(z) = 0.9015, you can form an equation in . —— O 23-
2
From the table, ®() = 0.9015 forz=1.29. coo, p150
23 —p
oe Equation in .
* Now solve this equation for 1. Tip: As always, check
that the answer seems
BSH 21295 23-.-253> am about right. You know it
must be a bit lower than
23, 50 20.42 seems OK. |

Example 2

If the random variable X ~ N(1, 42) and P(X > 19.84) = 0.025, find .

« Transform the probability you're given forX into a probability for Z.

Le
PUK> 19.84) = P(Z> eS) = 0.025
= Draw a sketch to show the information. ————————->
= From the percentage-points table, p = 0.025 for z = 1.96.
So, 19°84
w+ = 1.96 See p151 0 Ee

= Now solve this equation for iL.

pe OE = 1,96 = 19,84 =p = 7.84 > p= 12

Chapter 5 The Normal Distribution

 In the first two examples we found the mean, but you can find the
standard deviation (s.d.) in exactly the same way.

If the random variable X ~ N(53, 0?) and P(X < 50) = 0.2, find o.
= Again, start by transforming the probability you're given forX into a
probability for Z. The s.d. is unknown, so just leave it as o for now.
<20 =23) Ah
PUX < 50) = P(Z Bie o3)=02
= Draw asketch to show the information. The area —___»
is less than 0.5, so mark on the equivalent area to
the right 3
i of — i= ae
oO es =(=3/6) = 3/01
=02
See z

= You can see that P(Z = 3) = 0.2.

= From the percentage-points table, p = 0.2 for z = 0.8416.

3 fa
See pts!
So, aa 0.8416 <«~—_ Equation in o.

= Now solve this equation for o.

Tip: Values of o should 2 = 0.8416 > (ages *+— To3s.f.

always be positive. -~

When you’re asked to find the mean and the standard deviation, the method is
eee renee ol a little bit more complicated.
can solve simultaneous You start off as usual, but instead of getting one equation in one unknown to
equations by adding or solve, you end up with two equations in two unknowns, «. and o. In other
subtracting them to get | words, you have simultaneous equations, which you solve to find j: and o.
rid of one unknown. |

The random variable X ~ N(u, 0”).

If P(X < 9) = 0.5596 and P(X> 14) = 0.0322, find 1 and o.

= Let's start with the first probability for X,

and transform it into a probability for Z.

PIX <9) = P(Z <2 —#) = 0.5596

= Now, if you use the ®(z) table to find z for which ®(z)
= 0.5596, you can form an equation in ,1 and o.
From the table, @(z) = 0.5596 for z = 0.15. See p150
So, 2 = 0.193 9-110:
a Ven Ce :
1st equation in. and o

Chapter 5 The Normal Distribution

= Now do the same thing for the second probability for X.
Transforming to Z, you get:

eerie F) = 0.0322
Tip: p = 0.0322 isn’t in
= Draw a sketch to show the information. ——--—> the percentage-points
table, so it might help to
= You can see from the graph that: Cente use a sketch.

P(Z < a Aiywll P(Z 5 ax ==i} =f0.032 20.9678

= Now use the ®(z) table to find z for which ®(z) = 0.9678.
From the table, ®(z) = 0.9678 for z = 1.85.
os 14 — p
—~=1.85
> 14-p== ls85o0 <+_ Ind equation in LL
and o

= So now you have your simultaneous equations:

9-w=0.150 @ It helps to number the equations,
14-p=1.855 @ so you Can refer to them later.

= Each equation has one ‘1’, so you can subtract them to get rid of 1,
which will leave you with an equation ino to solve.
@ - @ gives 14-9 -—p-() = 1.850 - 0.150
=> 5 = 1.76
= 0 =2.941176...> 6=2.94 + Jo3f.

= Finally, find 1 by substituting o = 2.94... back into one of the equations.

Using equation @,
9-p=0.150>p=9-0.15 x 2.94... = Bee Ge +— Jo36¢

Exercise 3.3

Q1_ For each of the following, use the information to find pL.
a) X~ N(u, 6?) and P(X< 23) = 0.9332.
b)X~ N(u, 82) and P(X< 57)= 0.9970.
c) X ~ Niu, 100?) and PX> 528)= 0.1292.
d) X ~ N(u, 0.42) and P(X< 11.06) = 0.0322.
e) X~ N(u, O.0”) and PU> 1.52)= 0.9938.
ee 4: Q2?2 Hint: Start by
renmeete
Q2 X~ N(u, 3.52). If the middle 95% of the distribution lies between en
Ehowinenhe Geinbun
6.45 and 20.17, find the value of 1.
of X.

Chapter 5 The Normal Distribution

 O35 For each of the following, use the information to find o.
a) X~ N(48, 0?) and P(X< 53) = 0.8944.
b)) X ~ N(510,
( 0?) and P(X< 528) = 0.7734.
c) X~ N(17, 0?) and P(X> 24) = 0.0367.
d) X ~ N(0.98, 02) and P(X< 0.95) = 0.3085.
e) X~ N(5.6, 0?) and P(X> 4.85) = 0.8365.

Q4 X ~ N(68, 02). If the middle 70% of the distribution lies between 61

and 75, find the value of o.

Q5 For each of the following, find 1 and o.

aX = Nia, 07), P(X <30) 2079192"and PX <36) = 09252.
Q5 Hint: You need
b) X~ N(u, 02), PX< 4) = 0.9332 and P(X < 4.3) = 0.9987.
to form and solve
simultaneous equations c) X~ Niu, 02), PX< 20) = 0.7881 and P(X< 14) = 0.0548.
in and o. d) X ~N(uyro?), PX< 696). = 0.9713 and Pig2592) =.0 2420:
e) X~ Niu, 02), PX> 33) = 0.1056 and Pix >21) = 0.9599.
f) X~ N(p, 07), P(X> 66) = 0.3632 and P(X< 48) = 0.3446.

Q6 The volume of vinegar contained in bottles of vinegar is modelled by

a normal distribution with a standard deviation of 5 ml. Over time, it
is found that 71.9% of bottles contain less than 506 ml of vinegar.
a) Find the mean volume of vinegar contained in the bottles.
b) The label on each bottle says it contains 500 ml of vinegar.
What percentage of bottles contain less than 500 ml?

@7 The heights of a population of 17-year-old boys are assumed to

follow a normal distribution with a mean of 175 cm. 80% of this
population of 17-year-old boys are taller than 170 cm.
a) Find the standard deviation of the heights of the 17-year-old boys
in this population. |
b) One 17-year-old boy is selected from the population at random.
Find the probability that his height is within 4 cm of the mean
height.

Q8 In a particularly wet village, it rains almost continuously. The daily

rainfall, in cm, is modelled by a normal distribution. The daily
rainfall is less than 4 cm on only 10.2% of days, and it’s greater than
7 cm on 64.8% of days.
Find the mean and standard deviation of the daily rainfall.

Chapter 5 The Normal Distribution

Review Exercise — Chapter 5
Q1__ Find the probability that:
a) Z=— 0.84 Bb) 2295
c) Z> 0.68 d) Z=2 1.55
e) Z<-2.10 t) Ze 0,01
2) One h) Z< 0.64
i eZ O23 () O2L0'
Z.= 0:50
=
kK) 062 =Z 1.10 Il) -0.99<Z=-0.74

Q2_ Find the value of z if:

ay Ze = 2)—= O13) DimRZ <2) =0.5871
C) nee ZO 0509 CPZ 2)= 0.01
ey P! 22)=0:4013 . i eP(Z 2) =]0:995
g) Pi-z<Z<z)=0.5034 hy POIs <Z <7) 03917

Q3 If X ~ N(50, 16) find:

a) P(X <55) b) P(X< 42) c) P(X> 56) d) P(47<X<57)

Q4 IfX~N(5, 72) find:

a) P(X <0) b) P(X <1) c) P(X>7) d) P(2<X¥<A)

Q5 X~N(80, 15) |
ERGs a) = 0.99, finda. Q5 Hint: |X— 80] <b means
b) If P(LX— 80| <b) = 0.8, find b. Lae eae
Q6 The mass of items produced by a factory is normally distributed with a mean of 55 grams and
a standard deviation of 4.4 grams. Find the probability of a randomly chosen item having a
mass of:
a) less than 55 grams b) less than 50 grams c) more than 60 grams

Q7 The mass of eggs laid by an ostrich is normally distributed with a mean of 1.4 kg and a
standard deviation of 300 g. If 88.3% of the eggs laid by this ostrich have a mass of less than
a kg, find the value of a.

Q8 X~ Niu, 10) and PW< 8) = 0.8925. Find p.

Q9 X~ Ni, 82) and P(X> 221) = 0.3085. Find p.

Q10 X ~ N(11, 0?) and P(X< 13) = 0.6. Find o.

Q11 X~ N(108, 0) and P(X= 110) = 0.9678. Find o.

Q12 X~ N(j1, 0°).

The random variable
If PX< 15.2) = 0.9783 and P(X> 14.8) = 0.1056, then find . and o.

Chapter 5 The Normal Distribution

 Exam-Style Questions — Chapter 5

The exam marks for 1000 candidates can be modelled by a normal distribution
with mean 50 marks and standard deviation 15 marks.
a) One candidate is selected at random. Find the probability that they scored less
than 30 marks on this exam. (3 marks)

b) The pass mark is 41.

Estimate the number of candidates who passed the exam. (3 marks)

Find the mark needed for a distinction if the top 10% of the candidates achieved
a distinction. (3 marks)

The random variable X follows a normal distribution with mean \1 and standard deviation 6.
The probability thatX takes a value of less than 50 is 0.123.
a) Find the mean of this distribution. (4 marks)

ib) "Find POV 71).

(3 marks)
c) Find the value ofa such that P(u—a<X<y+a)=0.8. (4 marks)

The lifetimes of a particular type of battery are normally distributed with mean LL
and standard deviation o. A student using these batteries finds that 40% last less
than 20 hours and 80% last less than 26 hours.
a) Find and o.
(7 marks)
b) Find the probability that a randomly selected battery of this type has a lifetime _
of at least 15 hours.
(3 marks)

Chapter 5 The Normal Distribution

The random variable X has a normal distribution with mean 120 and standard deviation 25.

a) Find P(X> 145). (3 marks)

marks
b)) Findind the value of
the value of7j such that P(120 < X <7)/) =
= 0.4641 (4 marks)

The diameters of the pizza bases made at a restaurant are normally distributed.
The mean diameter is 12 inches, and 5% of the bases measure more than 13 inches.
a) Write down the median diameter of the pizza bases. (1 mark)

b) Find the standard deviation of the diameters of the pizza bases. (4 marks)

Any pizza base with a diameter of less than 10.8 inches is

considered too small and is discarded.
c) If 100 pizza bases are made in an evening, approximately how many would you
expect to be discarded due to being too small? (3 marks)

Three pizza bases are selected at random.

d) Find the probability that at least one of these bases is too small. fen)

A garden centre sells bags of compost. The volume of compost in the bags is normally
distributed with a mean of 50 litres.
a) If the standard deviation of the volume is 0.4 litres, find the probability that a
randomly selected bag will contain less than 49 litres of compost. (a mane)

b) If 1000 of these bags of compost are bought, find the expected number of bags
containing more than 50.5 litres of compost. ier

A different garden centre sells bags of similar compost. The volume of compost, in litres,
that
in these bags is described by the random variable Y, where Y ~ N(75, 6”). It is found
.
10% of the bags from this garden centre contain less than 74 litres of compost
cjaeeFind 0: (3 marks)

Chapter 5 The Normal Distribution

 S1 Statistical Tables
The normal distribution function
The cumulative distribution function 6(z) is tabulated below.
This is defined as 0(z) = ale dt.
oo 5 ix 1 Zz 12

$1 Statistical Tables
 $1 Statistical Tables
Percentage points of the normal distribution
The z-values in the table are those which a random variable
Z ~ N(O, 1) exceeds with probability p16. PZ > Zz) == oC ee:

S1 Statistical Tables
 e Q2 a) First you need to work out how many people are
Chapter 1 ss Data represented by each square unit on the graph
— use the information that the bar for 30-45
seconds represents 54 contestants.
1. Representing Data Bar for 30-45 seconds:
Exercise 1.1 — Data basics Width = 45 — 30 = 15 and height = 1.8
So area = 15 x 1.8 =27
See ee — this represents 54 contestants.
b) Mileage, Number of doors, Cost of service So each square unit represents
Q2 a) Number of medals won last season, Shoe size 54 + 27 =2 people.
b) Height, Mass Or frequency = 2 x area’
Q3 a) Th ere are no : gaps bet
etween Possl ible
ie heights.
elg S
Bar for 10-30 seconds:
Width Ak 10 = 20 et height = 03 ‘

as So area = 20 x 0.3 = 6
— this represents 6 x 2 = 12 contestants
{em)_|_ (cm) fem) So 12 contestants’ auditions lasted less than
tears
b) Now you need to add on the frequencies
represented by the other bars as well.
2 | 1708 180) ter) 175 Area of ‘45-55’ bar = 10 x 2.6 = 26, which
180 | 190 represents 26 x 2 = 52 contestants.
Area of ‘55-60’ bar = 5 x 3.0 = 15, which
represents 15 x 2 = 30 contestants.
: : Area of ‘60-75’ bar = 15 x 1.0 = 15, which
Exercise 1.2 — Histograms represents 15 x 2 = 30 contestants.
Q1 First add columns to the table to show class Area of ‘75-90’ bar = 15 x 0.4 = 6, which
boundaries, the class widths and the frequency represents 6 x 2 = 12 contestants.
densities. So the total number of contestants who
Mass (m, in g) class |Freq- auditioned was:
width }uency uency 12+54+52+30+30+12=190
Q3 a) The area of the bar is 1.5 x 9 = 13.5 cm?.
Ra This represents 12 butterflies.
ol So each butterfly is represented by
13.5 + 12 = 1.125 cm? on the histogram
b) 22.5 + 1.125 = 20, so the frequency was 20.
: c) The first bar, representing a class of width 3, was
1.5 cm wide. So the bar representing the class
53.5 < w < 58.5, which has a width of 5, must be
2.5 cm wide. And because it needs to represent
a frequency of 14, its area must be
14 x 1.125 = 15.75 cm’. This means it must be
15.75 + 2.5 = 6.3 cm high.

Exercise 1.3 — Stem and leaf diagrams

QI a) Add up the number of leaves in each row:
4+8+5+6+5+0+ = 30 choirs
2
rede b) There are 2 leaves of ‘5’ in the row with stem ‘4’.
So 2 choirs had 45 members.
c) 5 choirs
d) 73 members

Answers
Q2 a) 2+5=7 towns Q3 a) 5
b) 10.7°C b) There are 176 ratings in total.
OF 12a Sas °C 176 + 2 = 88, so the median is midway between
O37 a) 0} 8 (1) the 88th and 89th values.
ieee (3) Add a column to the table to show cumulative
Oat 4ae 5. G7 9.6) persuas
3 2 (2) Number of | Cumulative
AM 0 3 (2) customers recwueney
Saab (1)
6 |4 (1) ey ay eee
ce eae)See eee
2a a ae
b) (i) 6.4km (ii) 4 children anne hans|1043)
Q4 a) 60 minutes een eto
b) 17 From the cumulative frequencies, the 88th and
89th values are both 4, so the median = 4.
Cc) 6
Q4 a) 5 seats
There are 35 values altogether.
2. Location: Mean, Median
e e b)

Since 35 + 2 = 17.5 is not a whole number,

round this up to 18 to find the position of the
and Mode median. So the median = 20 seats.
Exercise 2.1 — The mean Q5 a) (i) 4.0 Mbit/s, 5.8 Mbit/s and 6.9 Mbit/s
Q1 The sum of all 12 prices is £13.92. (ii) 6.2 Mbit/s
So the mean price is £13.92 + 12 = £1.16 b) (i) There are 19 values altogether.
Q2 e722 0i= Garo Since 19 + 2 = 9.5 is not a whole number,
round this up to 10 to find the position of the
Q3
[Number ofgoals,x|0 |1 |2 |3 |4 |Total| median. So the median = 5.8 Mbit/s.
prequency fsa] 2 Z104./3 [1] 20 | (ii) There are 22 values altogether.
Pepe el Since 22 + 2 = 11 is a whole number, the
median is halfway between the 11th and 12th
So the mean is 28 + 20 = 1.4 goals
values in the ordered list.
Q4 a) 15x 47.4=711 years So the median = (5.5 + 5.9) = 2 = 5.7 Mbit/s.
b) New total of ages = 711 + 17 = 728
So new mean = 728 + 16 = 45.5 years
Or you could have used the formula with n, = 15, Exercise 2.3 — Averages of grouped data
x = 474, n, =1 and X2 = 17 to get the same answer.
Q1 a) Time (¢, mins) rrequen Mid-por x
ESaSee
Exercise 2.2 — The mode and the median Paces [eas
Q1 a) First put the amounts in order:
£19, £45, £67, £77, £84, £98, £101, £108, £110,
£123, £140, £185, £187, £194, £216, £250,
£500
b) ia = 60, Dade a4)
There are 17 amounts in total. Since 17 + 2 = 8.5
is not a whole number, round this up to 9 to find So estimate of mean = 334.5 + 60
the position of the median. = 5,6mins(fo 1 d.p.). \
So the median = £110. You could add an extra row to the table to show the
column totals. ;
b) All the values occur just once.
Q2 a) 0-2 letters
Q2 a) 6.9%
All the classes are the same width, so you can use
b) First put the rates in order: the frequency to find the modal class ( instead of the
6.2%, 6.2%, 6.3%, 6.4%, 6.4%, 6.5%, frequency density).
6.9%, 6.9%, 6.9%, 7.4%, 8.8%, 9.9%
There are 12 rates in total. Since 12 +2 =6isa
whole number, the median is halfway between
the 6th and 7th values in the ordered list.
So the median = (6.5% + 6.9%) + 2 = 6.7%.

Answers
 b) Add some extra columns to the table: So . = 30, meaning the median must lie in the class
‘270 < t< 300’. Now you need to sketch that class.
letters houses, f G b, 6
270% m 300
arenesA 7
+
30 48
Pree ee wo}; >Eee P.
b,
>
RID
GU}
ri a a2 3 . 2
Finally, solve =!i = =.
mes This g gives:
vf =50, fe = 209 m2270, 30 =27
So estimate of mean = 209 = 50 = 4.18 letters 300-270 Ara 27
c) Since ))f +2 =50 +2 =25, the median is or m = 30 X fl4270 = 274.3 mins (to 1d-p.)..
halfway between the values in this position (25)
and the next position (26) in the ordered list.
So the median must be in the class 3-5. Exercise 2.4 — Comparing measures of
Q3 Add a column showing cumulative frequency to the
table:
location
Q1 a) Median — most employees will earn relatively
Rainfall (r, mm) | Frequency | Cumulative low salaries but a few may earn much higher
frequency salaries, so the mean could be heavily affected by
5 a few high salaries.
40<r<50 I 12 b) Mean — the data should be reasonably
50 <r<60 9 symmetrical so the mean would be a good
60 <r< 80 i 36 measure of location. The median would be good
Bee |oe el as well (for a symmetric data set, it should be
roughly equal to the mean).
c) Mode — make of car is qualitative data so the
So 5 = a = 23, meaning the median must lie in
mode is the only average that can be found.
the class ‘60 < r < 80’. Now you need to sketch that
d) Mean — the data should be reasonably
class. symmetrical so the mean would be a good

60 ai
ay m
>
80
measure of location. The median would be good
as well (for a symmetric data set, it should be
2 23 36 roughly equal to the mean).
4 as > e) Median — most employees will perhaps travel
fairly short distances to work but a few employees
Finally, solve 2 = @. This gives: may live much further away. The median would
bh bp not be affected by these few high values.
60s ees 2 The mode is unlikely to be suitable in b), d) and e) (and
6060 — 3621 possibly a) as well) because all the values may well be
or m = 20 x a+ 60 = 62.7mm(to1d.p.). different.
Q2 There is a very extreme value of 8 that would affect
Q4 a) Estimated mean = 16740 + 60 = 279 minutes the mean quite heavily.
b) Add a column showing cumulative frequency to
the table:
Frequency, | Cumulative
3. Dispersion and Skewness
ie frequency Exercise 3.1 — Range, interquartile
iaoeremo| 8 |8 | range and interpercentile range
Q1 a) Highest value = 88846 miles

5 OE Pee Oe ane
360< t< 480
7 Lowest value = 3032 miles
So range = 88 846 — 3032 = 85814 miles
b) (i) There are 8 values, and the ordered list is
3032, 4222, 7521, 7926, 30 HIS), 3)\\HESS,
74898, 88 846

Since 4 = 2, the lower quartile (Q,) is

halfway between the values in this position
(2) and the next position (3) in the ordered
list.
So Q, = (4222 + 7521) + 2 = 5871.5 miles.

Answers
 (ii) Since an = 6, the upper quartile (Q,) is Finally, solve 4 = a This gives:
halfway between the values in this position b, 2

(6) and the next position (7)in the a Or 0563-22

list. So Q,= (31763 + 74898)= (= (oy
= 53 330.5 miles. or Qi =4 x FE + 10 = 12.2kg (to 1d.p.).
(iii) IQR = Q, — Q, = 53330.5 — 5871.5
= 47459 miles b) “a =3 x 252 + 4 = 189, meaning the upper
quartile (Q,) must lie in the class ‘14 < w < 18’.
Q2 a) andb)
In town at 8:45 am: Now you need to sketch that class.
The ordered list of 18 values is: b, >
hoa AMA e15) 19 15; 16, 16, 16, 16, 16, 14 a, ? 18
lial eae /-— : 4
97 a, 189 199
So the range= 18-13 =5 mph
b,
Since 4 = 4.5, the lower quartile (Q,) is in
ag ae 5 in the ordered list. So Q,= 15 mph. Finally, solve Z = B This gives:

Since 22 = 13. 5, the upper quartile (Q,) is in Ore 14 2 189 — 97

position 14 in theordered list. So Q, = 17 mph. 18—14° 199 —97
This means IQR = Q, - Q, = 17-15 = 2 mph. or Qs = 4 x Fa 4 14 = 17.6kg (to 1d.p,).
In town at 10:45 am: C) 30 1OKR OF 0 = 17-6
—122
The ordered list of 18 values is: = 5.4 kg (to 1 d.p.).
25, 29, 29,29, 30.30,341,31,31, 32, 33, 34, 34,
Bop SO SG), SKS oS Q4 a) Add a column showing cumulative frequency to
So the range = 39 —- 25 = 14 mph the table:
The lower quartile (Q,) is in position 5 in the
ordered list. So Q, = 30 mph. beetles | frequency
ooNo
The upper quartile (Q,) is in position 14 in the
ordered list. So Q, = 35 mph.
This means IQR = Q, — Q, = 35 — 30 = 5 mph.
On the motorway at 1 pm:
The ordered list of 18 values is:
67, 69, 69, 71, 71, 73, 73, 74, 74, 75, 75, 76, 76, The position of the 20th percentile is
76,78), 78, 88, 95 201 s ae
100 * 200 = 40, meaning that P,, must lie in the
So the range = 95 — 67 = 28 mph class ‘0-5’. Now you need to sketch that class,
The lower quartile (Q,) is in position 5 in the taking care to work out the lower and upper class
ordered list. So Q, = 71 mph. boundaries carefully.
The upper quartile (Q,) is in position 74 in the
ordered list. So Q, = 76 mph. 0 a,
> Be
..
5.5
This means IQR = Q, - Q, = 76-71 =5 mph. 4
0 a, 40 82
Q3 a) Add a column showing cumulative frequency to See
eee
b;
the table:
Weight of Number of | Cumulative Finally, solve 4 = @. This gives:
bi ~ by
baggage (w)
Po —O0 _ 40—0
O0<w<10 55 —0 » 820
10<w< 14 Or Py = 55) == 2.7mm(to1d.p.).
14<w<18
18<w<25

So a = 252 + 4 = 63, meaning the lower quartile

(Q,) must lie in the class ‘10 < w < 14’. Now you
need to sketch that class.

—— ms
10 4, Q, A
B 4
22 ae 63 97
{22 > z >
b,

Answers
 b) The position of the 80th percentile is (vi) The upper fence is
i x 200 = 160, meaning that P,, must lie in
=56+ (1.5 x 21)
Q, + (1.5 x IQR)
S85) EUS
the class ‘16-30’. Now you need to sketch that
b) The value 1 is outside the lower fence, so that is
class. an outlier. Similarly, the values 91 and 99 are
5, — outside the upper fence, so these values are also
15.5 4% P,, 30.5 outliers.
]
154 a, 160 184 C)
>
b,
Finally, solve 4 = ©. This gives:
bby Pheer
0 10 20 30 40 50 60 70 80 90 100
Po— 15.5 _ 160 — 154
30.5.2 wh5e5 ae (64a) 54
Q4 a) (i) There are 18 data values for Pigham.
oF Pao = 15 x B+ 15.5 = 18.5mm Since 18 + 2 = 9, the median is halfway
between the 9th and 10th data values (which
C) So the 20% to 80% interpercentile range are both 35). So the median = 35. °
=P, — P.) = 18.5 —2.7 = 15.8 mm (to 1 d.p.).
(ii) Since 18 + 4 = 4.5, the lower quartile (Q,) is
the 5th data value.
Exercise 3.2 — Outliers and box plots So the lower quartile = 23.
QI IQR=Q,-Q,=31-19=12 (iii) Since 3 x 18 + 4 = 13.5, the upper quartile
Lower fence = Q, — (1.5 x IQR) = 19-(1.5 x 12) =1
(Q,) is the 14th data value.
Upper fence = Q, + (1.5 x IQR) So the upper quartile = 46.
Sil $215
12) =4) (iv) The interquartile range = Q, — Q,
The value 4 is inside the lower fence, so 4 is not an = 46 — 23 = 23
outlier. The value 52 is outside the upper fence, so (v) the lower fence is
52 is an outlier. QO, ~ (125eQk)=23 = (So822)
Q2 a) IQR
=Q,-Q, =37-16=21 =-11.5
(vi) the upper fence,
b) Lower fence = Q, — (1.5 x IQR)
ml r= Geo eal
t= es Q, + (1.5 x IQR) = 46
+ (1.5 x 23)
Upper fence = Q) + (le5 2alQR) = 60.5
= O77 leo xl)
= 6885 b) No, none of the values fall outside the fences.
Since the minimum value and the maximum C)
value both fall inside the fences, there are no
outliers in this data set. elles
C)
ansread BAagaah ES 0 10 20 30 40 50 60 70 80

0 10 20 30 40 50 d) There are 18 data values for Goossea, so the

Q3 a) (i) There are 41 data values altogether, median is halfway between the 9th and 10th data
1.e.n=41. Since 41 +2 = 20.5 is nota values. So the median = 27.5.
whole number, round this up to 21 to find the The lower quartile (Q,) is the 5th data value.
position of the median. So the lower quartile = 15.
So the median = 46 cars. The upper quartile (Q,) is the 14th data value.
(ii) Since 41 + 4 = 10.25 is not a whole number, So the upper quartile = 35.
round this up to 11 to find the position of the
The interquartile range = Oe,
lower quartile.
= ile)
So the lower quartile (Q,) = 35 cars.
The lower fence is
(ili) Since 3 x 41 + 4 = 30.75 is not a whole
Q, - (1.5 x IQR) = 15
—(1.5 x 20)
number, round this up to 31 to find the
position of the upper quartile.
=-15
This means there are no ‘low outliers’.
So the upper quartile (Q,) = 56 cars.
The upper fence,
(iv) IQR = Q, - Q, = 56-35 = 21 cars
Q, #5 XIORYS 35 S15" 20)
(v) The lower fence is =165
Q, - (1.5 x IQR)
=35 =(1.5 x 21) This means that the value of 75 is an outlier (but
= 3.5 cars
the next highest value, 50, is not an outlier),

Answers
 So the box plot looks like this:
—I}- «
Extend the table to include fx, x? and fi’.

0 10 20 30 40 50 60 70 80

Exercise 3.3 — Variance and standard 192

deviation So Dif =2+34+54+74+44+2=23
dDA = 14424445
+704 44424 =221
Q1 a) ae 756 + 755 + 764 + 778 + 754 + 759
6 >) fe’ = 98 + 192 + 405 + 700 + 484 + 288 = 2167
_ 4566
6 — /6i This means =x = af 221
ya

Doe = 756" 755° = 764" a Se we oer

And so variance= vf TS ies)
+ 7787 + 754? + 759?
= 3475138 = 1.8903... = 1.89\(to 3 sig. fig.)

C) variance = Dating pues 24751385) 7617

n 6 oa rate 56-60|61-65]| 66-70 | 71-75
= 68.666... = 68.7(to3sig. fig.).
frequency, f 1 2
d) standard deviation = /variance = /68.666... (i)mid-a 2 R58" Po Poor a ee
= 8.29 (to3 sig. fig.).
e) There are no outliers or extreme values to affect
the standard deviation in a way that would make (iv) fx?
it unrepresentative of the rest of the data set.
b) @ Sf =20
Q2 a) y= 2% _ 480 60 (ii) fe = 1430
(iii) oe = 102 850
Six?
= 35292
Cc) Variance = oars -(34)
variance = 2 i 2 COs oile> De teen
— 102850 — (ey
b) standard deviation = Yvariance = /811.5 DO. 20
= 28.5 (to3 sig. fig.).
30525
Q3 a) Start = oa an extra row to the table for fx.
Q6 a) Add some more columns to the table showing the
class mid-points (x), as well as fx, x* and fx’.

50 <w<60
Then x* = Si tS leo
yo 60 <w< 70
70 < w < 80 1125|5625| 84375
b) Now add two more rows to the table.
80 < w < 90 510 |7225| 43350
xX
90<w<100]2] 95 |190 |9025| 18050
frequency, f
Lr So ys == EXS), ae = 07,0) yee = 266050.

Le Oe Then:

eae Variance =
Be (Ba)
So) fi = 7 32 e364 16 =91 yas
_ 266050 (3870
Ss — ("53)
Cc) variance= Ae == a =e 5e = 0.7475
= 134.95838... = 135kg’ (to 3sig. fig.)

d) standard deviation = /variance = 70.7475 b) Standard deviation = 7134.95838...

= 0.865 (to 3 sig. fig.). = 11.6 kg (to 3 sig. fig.)

Answers
Q7 a) Work out the total duration of all the 23 eruptions Q2 a) The coded data values are 3, 7 and 8.
that Su has timed. The mean of these is y = (3 + 7+ 8)+3 =6.
This is > x =x = 23 = 3:42 =" 76.66 minutes. The standard deviation of the coded values is
Work out the total duration of all the 37 eruptions 7°+8" _ 6? — 2.16 (to 3sig. fig.).
3+
that Ellen has timed.
So the mean of the original data values is
3 .

This is >)y = ny = 37 x 3.92 = 145.04 minutes

x = y + 2000 = 6 + 2000 = 2006.
So the total duration of the last 60 eruptions is:
The standard deviation of the original values is
x + Dy = 78.66 + 145.04 = 223.7 minutes the same as the standard deviation of the coded
This gives a mean duration of: values — this is 2.16 (to 3 sig. fig.).
223.0
con = Def PH8}0)9)b00 LW) The coded data values are 2, 17, 3, 11 and 7.
The mean of these is:
= 3.73 minutes (to 3 sig. fig.). y =(2+174+34+1147)+5=8.
b) Work out the sum of squares of the durations of The standard deviation of the coded values is
all the 23 eruptions that Su has timed — use the (ee
formula for variance. 5
me _ #3 107 = 2% — 3.42? = 5.51(to 3sig. fig.).
2

variance =
n 28
So the mean of the original data values is
So oan == 23) (07-5 3-42-)) = 295.3499 x=—y — 100] s— 100 — 0:08:
Do the same for the 37 eruptions that Ellen has The standard deviation of x
timed — use the formula for variance. = standard deviation ofy + 100,
so standard deviation ofx = 5.51 + 100
2 2!

variance = du _P~> 097? = 2 — 3.92? = 010551 ((0'3 siete?

Soy = 37 XK (0:97 3.92") = 603.3701 C) The coded data values are 7, 2, 20 and 15.
Now you can work out the total sum of squares The mean of these is:
(for all 60 eruptions): y =7 +2204
15) S4=11.
xX + Dy’ = 295.3499 + 603.3701 = 898.72 The standard deviation of the coded values is
So the variance for all 60 eruptions is: J) a
a
aaa
ee
,
variance = 898.72 DIN
60 = 60 ) = 6.96 (to 3 sig. fig.).
= 1.0781... = 1.08 min’
(to3sig. fig.) So the mean of the original data values is
C) This means the standard deviation of the
durations is 71.0781... = 1.0383... ¥ = 7+350=
41+350 =355.5.
= 1.04 min (to3 sig. fig.) The standard deviation of x
= standard deviation of y = 2,
so standard deviation of x
Exercise 3.4 — Coding = 6.96 + 2 = 3.48 (to3 sig. fig.).
d) The coded data values are 10, 7, 4, O and 6.
Q1 a) Since y=x-500, x = y+ 500. The mean of these is: ‘
SO Zea 0 0.5122 y =(104+7444046)25=5.4
standard deviation ofx = standard deviation of y The standard deviation of the coded values is.
= Ary?
(ee
b) Since y =PR 4x,ie¥ =a eo
3 aes
Soa.
ee
= 3.32 (to 3sig. fig.).
Se Of
4ee 2
4 gas
So the mean of the original data values is
Lex Z 01000 —_ x —20000 x = 10y — 8000 = 10 x 5.4 — 8000 = -7946.
C)
sa RGHS ae gaa ae The standard deviation ofx
This means ¥ = 15 x 12.4.+ 20000 = 20186. = standard deviation ofy x 10,
stan. dev. ofx = 15 x stan. dev. of y so standard deviation of x
= 5342081 = 3.32 x 10 = 33.2 (to3sig. fig).
Q3 a) All the values are of the form ‘0.6_’, and so if you
subtract 0.6 from all the values, and then multiply
what's left by 100, you'll end up with coded data
values between 1 and 10.
So code the data values usingy = 100(x — 0.6),
where x is an original data value and y is the
corresponding coded value.
This gives y-values of: 1, 7, 3, 3, 6, 5, 4, 8, 42

Answers
 b = Eve ee eee es eee
Ace 2 :
ae 10 Exercise 3.5 — Skewness
a a A Q1 a) symmetrical
, b) positively skewed
Find the sum of squares of the coded values, . Hl
y?. This is 2 999. c) negatively skewed
ys Ly ; Here, you can see there's a bit of a tail on the left’ of the
Cavaronee Das sta 5/229 68, 33 distribution (though you have to imagine rotating the
n : 10 ; stem and leaf diagram 9O° anticlockwise).
= 4.41 Q2 a) Pearson's coefficient of skewness
This gives a standard deviation of /4.41 = 2.1 =p node
ean 25
©) Since y = 100(x — 0.6), 7 = 100(¥ - 0.6 standard deviation 1.6
esat Sete ee BONE This is positive, so the distribution is positively
This means:
skewed.
= _ 4.3 oes
ai = “1 0.6 100 + 0.6 = 0.643 cm b) Pearson’s coefficient of skewness
Since
y = 100(%
— 0.6), » stmean = median) an jax (le. 13.5) ae 45
stan. dev. ofy = 100 x stan. dev. ofx ~ standarddeviation 1.8 ‘
So stan. dev. ofx = stan. dev. ofy + 100 This is negative, so the distribution is negatively
= 2.1 + 100 =0.021 cm skewed.
Q4 Make a new table showing the class mid-points (x) ©) Pearson’s coefficient of skewness
and their corresponding coded values (y), as well as = “mean mode 14a a
fy, ¥ and fy’. ~ standarddeviation =—-.2.37
Weight (to nearest g) |100-104] 105-109] 110-114] 115-119 This is zero, so the distribution is symmetrical.
equencif 3 d) Pearson’s coefficient of skewness
112 ~ 3(mean = median) 3°x (—9 = 8))
~ standarddeviation 25
=-1.2
This is negative, so the distribution is negatively
skewed.
Q3 a) Quartile coefficient of skewness
One 2 Opa O ie x8+3
= Q;—-Q, 15 =3
> h 675 = 0.167 (to 3sig. fig.).
variance of y = Sp wee = 7 ae 25° This is positive, so the distribution is positively
— 17.1875 skewed.
— ui 7 b) Quartile coefficient of skewness
This means standard BRNO MN Oy 20.4 0) 164 oe ones
4.15 (to3 sig. fig.).
= SO oeOi 194 35
Now you can convert these back to values for x. = — 0,0566 (to 3 sig. fig.).
Sincey =x — 102: This is negative, so the distribution is negatively
¥ =y+ 102 = 108.25 skewed.
pea ° es Bye a Only very slightly skewed, though.
stan. dev. ofx = stan. dev. of y = 4.15 g (to3 sig. fig.). ©) Quartile coefficient of skewness =
Q5_ Using the coding y= * +2. Qs = 205 Om 10 = 2x (= 23) = 28)
ee =O. ON flat = ijn)
Tee = = 0.444 (to 3 sig. fig.).
PO amen =% mee This is positive, so the distribution is positively
And the variance of y is: eeaven
De — i oe — 0.352 = 3.8775 d) Quartile coefficient of skewness =
Qr=2Qrr Cimr 82 ae 24 _ 9
This gives a standard deviation fory Q;— Q;
of 43.8775 = 1.97 (to 3 sig. fig.)
6 FS S085 =A 1.65: This means that the distribution is symmetrical.
And standard deviation of x ae
= standa rd deviati on ofy = 1.97 (to 3 sig. fig.)

Answers
 For Café B:
Q4 a) mean = 2s = w = ts, 50
mean = Ze = me == ff)
There are 8 values, so the median is halfway
between the 4th and 5th values in the ordered There are 14 values, so the median is halfway
list. So the median is £10. between the 7th and 8th values. So the
median is £81.50.
stan. dev. = x a a — 8.5° (ii) The mean is the same for both data sets, and
nA

= 3.20156... = £3.20 (to 3 sig. fig.). the median is only slightly different, so the
average daily sales in both cafés are very
b) Pearson’s coefficient of skewness similar.
3(mean — median) _ 3 x (8.5 — 10)
~ “standard deviation 3.20 b) (i) For Café A:
=-1.41 (to3 sig. fig.) OLI0l ae
stan. dev.= 2 =f 74
C) This is negative, so the distribution is negatively = eOLoe (tonearest penny)
skewed.
For Café B:
Q5 a) mean = 2% See
10
4.9
The mode is 22. stan. dev.= / akah =,/ 22074 7
=e IS bie(to nearest penny)
stan. dev. as
= De
as ase: =Sal 72>:
TO — 24.9°
(ii) The standard deviation is a lot higher for Café
= 12.3486... B, so sales are a lot more varied there (or
Pearson’s coefficient of skewness sales are a lot more consistent in Café A).
_ _mean—mode_ _ 24.9 — 22 c) (i) For Café A:
~ standarddeviation 12.3486...
There are 14 values, so Q, is the 4th value
= 0.235 (to 3 sig. fig.) and Q, is the 11th value from the ordered list.
b) This is small but positive, so the distribution is So Q, = £72 and Q, = £84.
slightly positively skewed. This gives:
Quartile coefficient of skewness =

Exercise 3.6 — Comparing Distributions Os 207-7 Ore

Onn ne
84 — 2 x 80+ 72
B42 72
Ql a) The median is higher for Shop B. This shows that = — 0.333
(to 3sig. fig.).
the prices in Shop B are generally higher. The
median in Shop A is approximately £37 while the For Café B:
median in Shop B is approximately £63, so the There are 14 values, so Q, is the 4th value
difference between the average prices is around and Q, is the 11th value from the ordered list.
£26. So Q, = £61 and Q, = £90.
This gives:
b) Although the ranges in the two shops are quite
similar, the interquartile range (IQR) for ShopBis Quartile coefficient of skewness =
higher than that for Shop A. This shows that the Q; — 2Q,+ Q; = v0 Sez x 81.5
+ 61
prices of shoes in Shop B are more varied than Q;— 0; 90 — 61 :
the prices in Shop A. = — 0.414 (to 3 sig. fig.).
C) The distribution of prices in Shop A is (ii) Both cafés have slightly negative coefficients
approximately symmetrical — this shows that of skewness, so they both have a ‘tail’
prices are fairly evenly balanced either side of the of relatively quiet days, where sales are
mean, without a long tail of really high or really relatively low.
low prices. For Shop B, the distribution of prices
Q3 a) (i) For the men:
is negatively skewed, showing that there is a tail
on the left of this distribution. This means that the mean = 222Fo oe=
aoe 7.3 hours
prices of the shoes are generally relatively high,
There are 10 values, so the median is halfway
but there were some much cheaper shoes as well.
between the 5th and 6th values in the ordered
Q2 a) (i) ForCaféA: list. So the median is 7 hours.
ee hOG Don't forget to sort the list before trying to find the
Me a ata E79
median — it's an easy mistake to make.
There are 14 values, so the median is halfway For the women:
between the 7th and 8th values. So the
median is £80. as
mean = S— = Ta 8.5 hours
Again, there are 10 values, so the median is
halfway between the 5th and 6th values. So
the median is 8.5 hours.

Answers
 (ii) The mean and median are both higher for the Q3 Add a row to the table showing fc.
women, so they get between 1 and 1.5 hours
more sleep per night, on average, than the
men.
b) (i) Forthe men:
Dawe = (MS, ape = 22" so mean
= 22.) |6= 17375
stan. nee = (3-73
Since there are 16 values, the median will be halfway
= 1.42 La (to 3 sig. fig.) between the 8th and the 9th values (which are both
For the women: 1), so median = 1
Mode = 0
stan. dev.= 2 x am” — 8.5? Q4 Extend the table to show the class mid-points (x) and
=' 1.63 foe? (to 3 sig.fig) the values of fx.
(ii) The standard deviation is slightly higher for
the women, so the number of hours of sleep
for the women varies slightly more from the
average than it does for the men.
c) (i) For the men: vf= 60, ¥ fe= 2226, so
Pearson’s coefficient of skewness estimated mean = 2226 + 60 = 37.1 mph.
~.3(mean = median) 3 x (7.3 — 7) Use linear interpolation to find the median.
standard deviation 1.42 Since there are n = 60 values, this means that
n + 2 = 30, and so the median will be in the
= 0.63 (to 2 sig. fig.).
class 35 - 39.
For the women: Now draw a diagram of that class showing the
Pearson’s coefficient of skewness cumulative frequencies at either end, as well as the
= 3(mean — median) | 3 x (8.5 = 8.5) _9 class boundaries.
standard deviation 1.63
eas <a eee
(ii) The distribution for the men is slightly 34.5 gi
ee eS
m 39
positively skewed, so there are a few 12 a, 30 49
relatively high values for the men, while most eee >
are relatively low. The distribution for the b,
women, on the other hand, is symmetrical. So you need to solve: m234,5 _ 30=12
39.5—345 49—12
This gives m = 5 X + 34.5
Review Exercise — Chapter 1 = 36.9 mph (to3 sig. fig.).
Ol 12-8513.2, 13.5, 1413, 14.3, 14.6, 14.8, 15.2, 15.9,
161, 16:1 16.2.6 917.0; .1722. “(all iniem) The modal class is 35 - 39 mph.
Always work out the class mid-points very carefully.
Q2_ Lots of fiddly details here — a table helps you get
them right. 5 MeanULtee 14 size
= 7
Length of calls
Number of calls =122=17.9
(to3sig. fig.).
Lower class boundary
Upper class boundary Stan. dev. = z
ee Bey Acoa1 2

Class width = 5.57 (to 3 sig. fig.).

Frequency density Q6 You need to extend the table here:
Score ate128-134
Frequency, f DD
Class mid-
— Pea | LO 110 duly, 124
618 | 1210 | 2574 | 1116
10609 | 12100|ee 15376 | 17161
Frequency
Density
aNwWBUD
63654 |133100 [301158 |138384| 34322
(OMT mone Ome ml On in
2al onan = =50, >)fe = 5780, and >fx” = 670618
Length of call / minutes yO
so mean = 9 == (5x6
670618 _ 115.6’ == 49
and variance = B70G18 —

Answers ail
Q7 Let
y=x — 20. b) IORH13 5513s:
So lower fence = 5 — 1.5 x 8.5 =-7.75
Then:
y= x— 20 of se—y-, 20 and upper fence = 13.5 + 1.5 x 8.5 = 26.25.
So the values 30 and 50 are outliers.
dy = 125 and Yo yr= 221
C) This means the box plot looks like this:
So y= 13 = Seance 10 dees —{] }— x x
And:
Daa
variance for y = Ta 1.25* = 0.6475 and so OS
+++
10.1 2045
30 35 40) 25> 50
3
Pocket money (£)
standard deviation for y = 0.805 to3sig. fig.
This data is positively skewed. Most of the
But since y = x — 20, this must also equal 15-year-olds earned a relatively small amount of
the standard deviation for x. pocket money. A few got very large amounts.
So standard deviation forx = 0.805 (to 3 sig. fig.). Q11 Quartile coefficient of skewness =
Q:==207-POr 150= 2 x 13256 36
O7=Q; (S036
— ULSD
A possible sketch of the distribution would need to
be negatively skewed — something like this:

Then >)f =20, >)fy = 48, and >)fy” = 416

This gives:
p= 3 2.4
30 % = y 735.5 — 2.4—> 35.5 — 37.9 minutes Q12 a) (i) Put the times in order first:
Iie SA a S , WO, We
Since n + 2 = 10 + 2 =5, the median will be
variance for y = aR — 2.4* = 15.04 and so halfway between the 5th and 6th values.
standard deviation for y = 3.88 to3sig. fig. So the median is 5 minutes.
But since y = x — 35.5, this must also equal (ii) Sincen + 4 = 10 + 4 = 2.5, the lower quartile
will be the 3rd value.
the standard deviation for x. So standard deviation
So the lower quartile is 4 minutes.
forx = 3.88 minutes (to3 sig. fig.). The upper quartile is the 8th value, which is
Q9 IQR
= 88 — 62 = 26, so 1.5 x IQR
= 39. 7 minutes.
So upper fence = 88 + 39 = 127, (ili) IQR = 7-4 =3.
and lower fence = 62 — 39 = 23. So lower fence = 4 - 1.5 x 3 =-0.5
This means that: and upper fence = 7 + 1.5 x 3 =11.5.
a) 124 is not an outlier. So the value 12 is an outlier.
b) 131 is an outlier. b) For Worker B, the ordered list is:
c) 28 is not an outlier. AA OF 7.8, Gp OO ali
Q10 a) Put the 20 items of data in order: So the median and quartiles are:
1, 4) By Dy Dy Dp Bp 0, OT; 7, @, IO, 10, 12, 15, LO, Q, = 6 minutes, Q, = 8 minutes, Q, = 9 minutes.
AQ), BO), SO This gives a lower fence of 6-1.5 x 3 =1.5,
Since n + 2 = 20 + 2 = 10, the median is halfway and an upper fence of9 + 1.5 x 3 = 13.5.
between the values in positions 10 and 11, and This means there are no outliers for Worker B, so
since the 10th and the 11th items are both 7, the your box plots should look like this:
median = £7. - Worker A
Since n +4 =20+4=5, the lower quartile is =fin et
halfway between the values in positions 5 and 6, WorkerB | |
and since the 5th and 6th values are both 5, the
lower quartile is £5. —|
O24
Ce 8 0 12 time in mins
Since 3n + 4 = 15, the upper quartile is halfway
between the values in positions 15 and 16.
So upper quartile = (12 + 15) + 2 = £13.50.

Answers
 c) You could say various things — here, you can b)<x=
choose whether to say something about the
location or dispersion or skewness. And so })x = 42 x 19 = 798 [1 mark]
E.g. the times for Worker B are 3 minutes longer Dae es
than those for Worker A, on average. Variance of x =
190, ae.
The IQR for both workers is the same — generally
they both work with the same consistency. ee
at.Ce eee
= [1 mark] ee
The range for Worker A is larger than that for And so >)x? = (16 + 427) x 19
Worker B. Worker A had a few items he/she
= 33820 [1 mark]
could iron very quickly and a few which took a
long time. C) New >)x = 798 + 32 = 830 [1 mark]
Worker A’s times are positively skewed, whereas So new
xX= we = 41.5 [1 mark]
those for Worker B are negatively skewed.
New ae = 33820 + 32* = 34844 [1 mark]
d) Worker A would be better to employ. The median
time is less than for Worker B, and the upper So new variance = ae = 41.57 = 19.95
quartile is less than the median of Worker B. and new stan. dev. = 4.47 to3 sig. fig. [7 mark]
Worker A would generally iron more items in a
given time than worker B. 60.3 = 3.015g [1 mark]
a) = 0

Exam-Style Questions — Chapter 1 b) Variance = we — 3.015* [1 mark]

1a) You need to find the frequency densities, since = 1.8597... g* [1 mark]
the classes are not all the same width. So standard deviation = 1.36 g (to3 sig. fig.)
Profit, Number of |Class | Frequency [1 mark]
£x million | businesses |width density C) Brand A chocolate drops are heavier on average
4.0<x<5.0 24 1.0 24 than Brand B. Brand B chocolate drops vary in
5.0<x< 5.5 26 0.5 52 mass much less than brand A.
5.5<xX< 6.0 21 0.5 42 [1 mark for each of 2 sensible statements]
6.0<x< 6.5 19 0.5 38
d) Combinedmean = date
6.5<x< 8.0 1S
__ 60:3 5 |OOLX 2795)
The frequency densities are the heights of the e 50
bars. = 2.976 g [1 mark]
b2
Standard deviation for B=1, so ue = DO = |,
and so >)b* = 291.075 [1 mark]
2 b?

Combined variance = web — 2.976°

Density
Frequency
i aa, Eee
Kee ane ee eyOme DMG 2
Profit / £million
[1 mark for correct axes, plus 2 marks if all bars = 1.3449...
[1 mark]
drawn correctly, or 1 mark for at least 3 bars
So combined standard deviation for all 50
correct.]
chocolate drops is 71.3449... = 1.168
b) The distribution is positively skewed — only a
few businesses make a high profit. The modal [1 mark]
profit class is between £5 million and £5.5 Work through each step carefully so you don't make silly
million. mistakes and lose marks.
[1 mark per sensible comment, up to a maximum 4 a) There are 30 males. Since 30 + 2 = 15, the median
of 2.] is halfway between the 15th and 16th values in the
2 a) Let y=x-30. Then ordered list. So take the mean of the 15th and 16th
p= 228 — 12, sox =y + 30 = 42 [1 mark] readings to get
median = (62 + 65) + 2 = 63.5 [1 mark]
variance ofy =208 — 12?= 16 [1 mark],

and so standard deviation of y = 4. But

standard deviation of x = standard deviation of y
and so standard deviation of x = 4 [1 mark]

Answers
 b) The female median is halfway between the 8th
and 9th readings, so median = 64.5. This means
Chapter 2: Probability
the female median is higher than the male
median by 1 mark, so the females scored better
than the males on average.
1. Elementary Probability
Female range = 79 — 55 = 24. Exercise 1.1 — The sample space
Male range = 79 — 43 = 36 -
Q1 a) There is 1 outcome corresponding to the 7 of
The female range is less than the male range. .
Their scores are more consistent than the males’. diamonds, and 52 outcomes in total
[Up to 2 marks available for any sensible So, P(7 of diamonds) = 5
comments] b) There is 1 outcome corresponding to the queen of
You could also have compared the figures for the spades, and 52 outcomes in total.
mean, interquartile range or skewness. So, P(queen of spades) = a
a) Total number of people = 38 c) There are 4 outcomes corresponding to a ‘9’, and
So median is halfway between the 19th and 20th 52 outcomes in total.
So, P(9 of any suit) = s = at
values in the ordered list [7 mark].
19th value = 15; 20th value = 16, 13 '
so median = 15.5 hits [7 mark]. d) There are 26 outcomes corresponding to a heart
or a diamond, and 52 outcomes in total.
Mode = 15 hits [7 mark]
So, P(heart or diamond) = a — :
b) 38 +4 =9.5, so lower quartile = 10th value = 14
38 +4 x 3 = 28.5, so upper quartile = 29th value Q2 a) 6 of the 36 outcomes are prime numbers.
=i So, P(product is a prime number) = ~ oe i
[1 mark for both]
b) 14 of the 36 outcomes are less than 7.
So interquartile range = 17 — 14 = 3, So, P(product is less than 7) = tn — a
and upper fence = 17 + (1.5 x 3) = 21.5.
This means that 25 is an outlier [7 mark]. C) 6 of the 36 outcomes are multiples of 10.

° ~[T}— .~
So, P(product is a multiple of 10) = z = z
Q3 a) E.g.

(2
eeee re
AS Oreo 20) 22 24726
Number of hits [1 mark]
As Q, - Q, > Q, - Q,, and the line extended i 2 3:SA EA Ca eae Oe LO)
from the box is longer on the right-hand side,
the distribution seems to be slightly positively b) There are 5 ways of getting an even number and
skewed. [7 mark]
‘tails’, and 20 outcomes altogether.
D
# P(even number and tails) = 30-4 1
Make sure you can tell positive skew and negative skew
apart — from a histogram and from a box plot. Q4 a) g.
d) lf 25 was removed then the distribution would
be more symmetrical (but still slightly positively
skewed) [7 mark].
Find the total area underneath the histogram using
the grid squares [1 mark]:
24+1.5424+24+154+44+5+4+3+4=25 [1 mark]
So each grid square represents 2 lions [17 mark].
The number of squares for lengths above 220 cm is 7
[1 mark], which represents 7 x 2 = 14 lions [17 mark].
6 of the 36 outcomes are zero.
So, P(score is zero) = Oo een
A. 6
C) None of the outcomes are greater than 5.
So, P(score is greater than 5) = 0
d) The most likely score is the one corresponding to
the most outcomes — so it’s 1.
10 of the 36 outcomes give a score of 1, so:
30 wots

Answers
Q5 Start by drawing a sample-space diagram to show all
d) P(U AM’ AN’) = 1-(0.25 + 0.1 + 0.15 + 0.25)
the possible outcomes for the two spins combined.
= 0.25
Then circle the ones that correspond to the event
‘number on spinner 2 is greater than number on e) P(LU M) = 0.25 + 0.1 +0.15 = 0.5
spinner 1’. E.g.
P(M’) = 0.25 + 0.25 + 0.25 =0.75
10}®@ @ Don't forget to include P(L'’ 1 M’ MN’). You could
also find P(M’) by doing 1 — P(M) = 1— (01 + 0.15).
Q4 a) Number of outcomes in S’M F’ A G’
= 200-(17+
18+ 49+284+11+34+6)=37
Spinner
2
So P(S’ MF’ AG)= ia
b) You're only interested in those people who have
De? oe we || been to France — 49 + 28 + 34+11=122
Spinner1 people. The number of people who have been to
There are 13 outcomes that correspond to the event France and Germany = 28 + 34 = 62.
‘number on spinner 2 is greater than number on So P(G, given F) = oS = 4
spinner 1’, and 25 outcomes altogether.
C) Number of outcomes in SM F’ = 17 + 18 = 35.
So, P(spinner 2 > spinner 1) = i
35
So PS NF) = #2 = 7
Q5 a) Start by drawing a Venn diagram to represent
2. Venn Diagrams the information. If H = ‘goes to home league
matches’, A = ‘goes to away league matches’ and
Exercise 2.1 — Using Venn diagrams C = ‘goes to cup matches’, then:

ao
Q1 a) Label the diagram by starting in the middle with
the probability for AM B. Then subtract this
probability from P(A) and P(B). And remember
to find P(A’ M B’) by subtracting the other
probabilities from 1. So:

The people who go to exactly 2 types of match

are those in HN ANC’, HN CNA and
AMCMH’. That's 200 + 40 + 20 = 260 people.
So P(2 types of match) my= 1000
60 eis
= 50
0.25 b) P(at least 1 type of match) = 1 — P(no matches)
240 760 19
b) P(A B’) =0.25 hi 1000 ~~ 1000 5
@2 a) LetMbe ‘studies maths’ and P be ‘studies
physics’. Start by labelling the area MM P with
19. Then complete the circles by subtracting 19 3. Laws of Probability
from the numbers studying maths and physics. Exercise 3.1 — The addition law
And don’t forget to find the number who don’t
study either subject. So: QI a) P(A) = 1 = PA) = 103 = 07,
b) P(A U B) = P(A) + P(B) — P(A 9 B)
=0.3 +0.5—0.15 =0.65
Cc) P(A 1B) = 1=P(AU B)=1—0:65=0.35
Remember, A’ B' is the complement ofA U B.
Q2 a) P(B’) = 1 — P(B) = 1 — 0.44 = 0.56
HY b) P(A U B) = P(A) + P(B) — P(A OB)
= (1 — 0.36) + 0.44 — 0.27 = 0.81
b) MUP has 27 + 19 + 19 = 65 outcomes.
SoiR(MIe PR) = 65.
144 C) P(A MN B’) = P(A) — P(A B) = 0.64 — 0.27 = 0.37
d) P(A U B’) = P(A) + P(B’) — P(A B’)
c) So you're only interested in the 46 students = 0.64 + 0.56 — 0.37 = 0.83
who study maths. 19 students study maths and
Q3 Let B = ‘car is blue’ and E = ‘car is an estate’.
physics, so P(P, given M) = 2.
a) P(B’) = 1 — P(B) = 1—-0.25 = 0.75
Q3 a) P(LMM)=0.1 b) P(B U E) = P(B) + P(E) — P(B MN E)
= 0.25 + 0.15 — 0.08 = 0.32
b) P(LON)=0
as C) P(B’ ME’) =1 —P(B UE) =1 —0.32 = 0.68
C) aa Ni 02S
Since N doesn't overlap with L, L'O N ts just N.

Answers
 Q4 a) P(Y’)=1-—P(Y) =1-0.56 = 0.44 Q4 a) P(AMB) = P(A) + P(B)— P(A U B)
b) P(X MY) = P(X) + P(Y) — P(X UY) = 0.28 + 0.66 — 0.86 = 0.08
= 0.43 + 0.56 — 0.77 = 0.22 P(A MB) # 0, so A and B are not mutually
exclusive.
@) ROM ING) = 1 SO UING) = 1 = O77 OS
d) PX UY) =1-P(IXNY)=1-0.22 =0.78 B) =PAN Ce Oe ee
0.28 + 0.49 —0.77 =0
Q5 a) PC’ ND)=P(C) + P(D)- PIC UD) P(A 9 C) = 0, soA and C are mutually exclusive.
= (1 —0.53) + 0.44 — 0.65 = 0.26 ©) PIBAC)=P(B) +P(C-PBUC)

b) PC’ 1D’) = P(C) - (Cn D) = 0.66 + 0.49 - 0.92 = 0.23

= 0.47 - 0.26 = 0.21 P(B A C) = 0, so B and C are not mutually
Just Gs Ge Gia DiGi Dy, exclusive.

Sa ee Q5 a) You need to show that P(C ND) = 0.

c)) PIC IE Da = WC) ri) — Pe") P(C)=1-0.6 =0.4.
= Eouaak ee mn P(C ND) = P(C)- PIC ND’) = 0.4 - 0.4 = 0, so C
d) P(CND)=P(C) + P(D) - PC UD) and D are mutually exclusive.
= IC) + (Dis [1 = RC’ 1D] b) P(C UD) = P(C) + P(D) = 0.4 + 0.25 = 0.65
ena Te a OeNS a _ Q6 Out of the total of 50 biscuits, 30 are plain, and
Qe LetM = ‘has read To Kill a Mockingbird’ and A = ‘has 20 are chocolate-coated. Half of the biscuits are
read Animal Farm’. Then, P(M) = 0.62, P(A’) = 0.66, in wrappers, so 25 biscuits are in wrappers. Since
and P(M UA) = 0.79. there are more biscuits in wrappers than there are
a) P(MMA) = P(M) + P(A) — P(M U A) -chocolate-coated ones, there must be some biscuits
= 0.62 +(1)— 0.66) = 0.79. = 0.17 (at least 5) which are plain and in wrappers. So
b) P(M’ NA) = P(A) —P(MNA) = 0.34 — 0.17 = 0.17 events P and W can happen at the same time (i.e.
©) P(M’ AA) =1—P(MUA) =1-0.79 = 0.21 P(P ™ W) # 0), which means they are not exclusive.

Exercise 3.2 — Mutually exclusive events

Q1 a) PXNY)=0 Exercise 3.3 — The product law
b) P(X UY) = P(X) + P(Y) = 0.48 + 0.37 = 0.85 — conditional probability
©) PX AY) =1-PIKUY)=1-0.85 = 0.15 QI a) P(G|H)= on = 0.24 _24_ 8
Q2 a) P(LUM) = P(L) + P(M) = 0.28 + 0.42 = 0.7 You could give the answer as a decimal instead, but
b) P(LUN) = P(L) + P(N) = 0.28 + 0.33 = 0.61 using a fraction means you can give an exact answer.
c) P(MUN)
= P(M) + P(N) — P(M N) b) P(HIG APIGST
ED. Aue oa ke (9
= 0.42 + 0.33 — 0.16 = 0.59 ) PR y= Pie Saat) eae
d) P(LAMNN)=0 P(A MB) a 034
ae
e) Draw 3circles to represent events L, M and N, 28) eT CN SORES eae
making sure that mutually exclusive events don’t PIANC
overlap. As usual, start the labelling with the b) P(A|C)= eon 7 at 7 a oe 3 ;
intersection and work outwards. PAC) ’
©) Cee
P(C'|B) =p OLS = 0.875
=O
Q3 Let F = ‘over 6 feet tall’ and G = ‘can play in goal’.
a) PGIp =e
P(F)
0.13 P(FG) = 7 and P(F) = =
Q3 a) Let B = ‘goes bowling’, C = ‘goes to the cinema’, P(E C) =
and D = ‘goes out for dinner’. All 3 events are So, P(G|F) = on + 5 iv:5
mutually exclusive, so: th
P(B U ©) = P(B) + P(C) = 0.17 + 0.43 = 0.6
oP Ge P(F M G)
Geen :Sa
b) P(doesn’t do B, C or D) = PB’ NC’ ND’) LG) A OD aig ae"
Since either none of B,y C and D happen, / or at P Z
—
least one of B, C and D happen, B’ C’ ND’ and P(F |G) = oo = u = 3
B UCU Dare complementary events. So: 1
PBS Cg DD) = 1 PB Urewp) 10
= 1 — [P(B) + P(C) + P(D)] a p _ MOA)
_ jot ato
= 1-(0.17 + 0.43 + 0.22) Oh P(A) = Bai3s
= 1-0.82 be
= 0.18

e108
Fe Answers
 DYSON Pry) PCAN EAM) te
Se
el) _ 100
755 2
Q5 A and B:
cl ) P(EM M) He P er ae ple
100
ge ras as nn
ey Pee) P(A) x P(B) = = = P(A 1B),
= eM) 700 _ 39 _ 13
P(M’y ss60 *—3O
«C200 so A andB are independent.
100
A and C:
d) P(A|E’) PANE) aa 4 _
OVE = sche =paps 18
eo= 36 100
P(A) x P(C)= 3
; x 78=— AS
308

PCA) <P (C= 5 Fe 2

Adel yee 15)
DOMIAE P(IMMANE) = 00
2

STON ADE) = Toray Rio TS so A and C are not independent.

Q5 a) P(Y)=1-P(Y) =1-0.72=0.28 Band C:
b) P(X NY) = P(X|Y)P( pages 02 Hohe hr ae
eas i ee OS
c) P(X Z) = P(Z|X)P(X) = 0.25 x 0.44
=0.11
P(B) x P(C) = a =-P(B 6 C),
d)DE)
P ; = eR 0.2Ue 20
= ean 0.61 <2 so B and C are independent.
e) PIXMA Y MZ) =P(XNY|Z)P(Z) Q6 a) P(X MY) = P(X)P(Y) = 0.84 x 0.68 = 0.5712
= eRe Oly b) P(Y’A Z’) = P(Y’)P(Z’) = 0.32 x 0.52 = 0.1664
61 * 100 c) Since Y and Z are independent,
= of
TO = 0,07 P(Y |Z RW) = 0:68

There are lots of ways to write an expression for d) Since Z’ and Y’ are independent,
POX MYA) eg) PZ XA YYP(XM Y) or RZ lea = |Z) = 11 =0418 =O. 52
PIX|YN Z)P(YM Z). You have to choose the way that e) SinceY and X’ are independent,
makes best use of the information in the question. P(Y |X= R= OF68
Q7 LetJ = ‘Jess buys a DVD’, K = ‘Keisha buys a DVD’
Exercise 3.4 — Independent events and L = ‘Lucy buys a DVD’.
QI P(X MY) = P(X)P(Y) = 0.62 x 0.32 = 0.1984 a) The probability that all 3 buy a DVD is
Q2 P(A)P(B) = P(A NB),s PUMKAL). Since the 3 events are independent,
p(ay= PAO
PB)
B) 0.45 you can multiply their probabilities together to
~ T=0.25= 0.6 get: PU) x P(K) x P(L) = 0.66 x 0.5 x 0.3 = 0.099
Q3 a) P(MAN) = P(M)P(N) = 0.4 x 0.7 = 0.28 b) The probability that at least 2 of them buy a DVD
will be the probability that one of the following
b) P(MUN) = P(M) + P(N) — P(MMN)
happens: JN KN LorJQAKNU orJNK’ AL or
= 0.4 + 0.7 — 0.28 = 0.82
Jemkeiag
c) P(MAN’) = P(M)P(N’) = 0.4 x 0.3 = 0.12 Since these events are mutually exclusive, you
Q4 a) LetA= ‘Ist card is hearts’ and B = ‘2nd card can add their probabilities together to give:
is hearts’. Then, since the first card is replaced 0.099 + (0.66 x 0.5 x 0.7) + (0.66 x 0.5 x 0.3) +
before the second is picked, A and B are (0.34 x 0.5 x 0.3)
independent events. = 0.099 + 0.231 + 0.099 + 0.051
So, P(A Mm B) = P(A) x P(B) = 0.48
There are 13 hearts ps of the 52 cards, so P(A)
= itq
and P(B) both equal 5
2 Exercise 3.5 — Tree diagrams
So, P(A B) Ee= h5
ligy x ~ ayPras
ie Q1 a) The events are not independent because the
probability that Jake wins his 2nd match depends
b) Let A = ‘1st card is ace of hearts’ and B = ‘2nd
on whether or not he won his 1st match.
card is ace of hearts’. Then, since the first card is
replaced before the second is picked, A and B are b) (i) P(Win then Win) = 0.6 x 0.75 = 0.45
independent events. (ii) P(Wins at least 1) = P(Win then Win) +
So, P(A M B) = P(A) x P(B) P(Win then Lose) + P(Lose then Win)
There is 1 ‘ace of hearts’ out of the 52 cards, so = 0.45 + (0.6 x 0.25) + (0.4 x 0.35)
P(A) and P(B) both equal 55 Oy
Or you could find 1 — P(Lose then Lose).
1 [eegee |
So, BAW By 55% 55 = 5704 -

Answers
Q2 a) 1 Heads Q6 a) Let R, = ‘ball i is red’, Y, = ‘ball i is yellow’ and
2 G, = ‘ball i is green’, for i= 1 and 2. Since the
Ye ae first ball isn’t replaced, the second pick depends
= 2 Tails on the first pick and you get the following tree
S\Not 31 Heads
diagram:
ee | \
2 Tails
Sl 1

b) P(wins) = Pn Tails) = +x
Onan
= 1+
mel =heYf\= RLS!
pS
On \
Q3 a) Let D = ‘passed driving test’ and U = ‘intend to go
to university’. Since D and U are independent, ale
eS
P(U |D’) = P(U|D), and so:
lua \
nN

O75 9)
als 1

INAS
03220 au eS
@)
a
72
Ou
y

xe 025 >
== nN

P(wins) = P(R, OR,) + PCY, NY,) + P(G, 9 G,)

0.7™p’ O:7or-aU)
ee b) ee
025 =U" =x)
+(aT) +GX TD
a2Ole ee Om
be (D wwii 04 O25 — 05175 = F370
137Me?
CPW RiDin
=\(053.
es RD ews)
0.25) 02175
yee is
=139 66
If the first ball is replaced, the probability of
Or, since the events are independent,
winning becomes:
P(U') = P(U'| D) = P(U"| D) = 0.25 (reading from the
tree diagram above).
Qa Let R = ‘orders roast dinner’ and let A = ‘orders apple
pie for pudding’. Then you can draw the following
tree diagram:

Since a is
72 ~ 66’ a player is more likely to win
now that the game has been changed.
You could also answer this question by explaining that
when a ball of one colour is selected, then replaced, the
proportion of balls of that colour left for the second pick
SO BVA) = P(RIGWA) = RUR nA)
is higher than ifit isn’t replaced. So the probability of
= (0.56 x 0.72) + (0.44 x 0.33)
picking the colour again is higher.
= 0.5484
Q5 Let S = ‘owns smartphone’ and let C = ‘has contract
costing more than £25 a month’. Then you know the Review Exercise — Chapter 2
following probabilities: P(S) = 0.62, P(C) = 0.539 and Q1 a) The sample space looks like this:
P(C’ |S) = 0.29, and you want to find P(S|C).
Using the conditional probability formula: Dice
1 2 3 4 5 6
P(S iC)
PS |C)=
rors Eth |e 4 6 8 10 12
Use a tree diagram to help you find P(S MC). You CoinT 6 Hi 8 9 10
don’t need to label all the branches, just the ones that
help you answer the question: b) There are 12 outcomes in total, and 9 are more
O71 C than 5, so P(score more than 5) = =F:3 9

0.62. -S Sea ©) There are 6 outcomes which haveatail showing,

Q20~C' and 3 of these are even, so P(even score given
‘e that a tail is thrown) = 2 = 1
0.38 S!Een ay
Cc

So RiSim) @) = O62) 07 04402 and'so:

_PSAC) _ 0.4402 _

Answers
Q2 a) 50% of the students eat sausages (S), 20% eat (iii) Choosing an upper school pupil means either
chips (C) and 2% eat sausages and chips, so you ‘boy and upper’ or ‘girl and upper’.
get the following Venn diagram: P(B MU) = 0.4 x 0.3 = 0.12
ACG iO) SOS S105 SO sx0)
So P(U) = 0.12 + 0.30 = 0.42

Exam-Style Questions — Chapter 2

1 a) ~The Venn diagram would look like this:

Ss
b) Using the Venn diagram, 18% of students eat
chips but not sausages.

ie
c) Using the Venn diagram, 18% + 48% = 66% eat
chips or sausages, but not both.
Q3 It’s best to start by drawing a sample-space diagram:
[1 mark for the central figure correct, 2 marks for ‘5;
‘7’ and ‘4’ correct (get 1 mark for 2 correct), 1 mark
for ‘287 ‘15’ and ‘11’ correct, plus 1 mark for a box
with ‘47’ outside the circles.]
b) (i) Add up the numbers in all the circles to get
73 people out of 120 buy at least 1 of the
soaps [7 mark].
Or you could do 120 — 47 = 73...
So the probability = x [1 mark]
(ii) Add up the numbers in the intersections
a) to get5+3+4+7=19,
meaning that 19
people buy at least two of the soaps
; [1 mark].
b) 7 of the 36 outcomes are square pes (since 4 Sethe Drobabiliuea Dereon OU eat eset Wo
and 9 are square), so P(square) = 36 See as Tt marke
c) It isn’t possible for her score to be both a prime (iii) 28 + 11 + 15 = 54 people buy only1 of the
number and a square number, so the events P and Gone b 5
S are mutually exclusive. soaps [7 mark], and of these 15 buy soap
[1 mark].
d) Since P and S are mutually exclusive, So the probability of a person who only buys
PS 7SEEN ayo 5 67
36 p=— 22
36 ti
4 : type B is lees
one of the soaps buying = = +5
e) The score from the second experiment is [1 mark].
unaffected by the score from the first experiment, Let R, = ‘Ist red’, W, = ‘Ist white’,
so the events S, and S, are independent. G, = ‘Ist green’, R, = ‘2nd red’, W, = ‘2nd white’
79 49 and G, = ‘2nd green’. Then:
a) rea 36, 36 1296 ALR,
Q4 a) Let B=boy, G= girl, U = ‘in upper school’ and 4, W,
L = ‘in lower school’, then: Wy pe V1 G
0.3-U hy R,

0.4 B : Yi a
Oxiral: See
0.5 U
Y,12 : ;
4 Ww,

0.6 G Z11 G

0.5 SL [3 marks available — 1 mark for each set of 3

b) (i) Reading from the tree diagram, P(L|B) = 0.7. branches on the right-hand side correct]
(ii) PCG OL) = 0.6 x 0.5 = 0.3

Answers
 b) The second counter is green means one of three a) P(chicken or ice cream but not both)
results: ‘red then green’ or ‘white then green’ or = P(C NS) + P(BAI) [1 mark]
‘green then green’. So P(2nd is green) = = (3 x 2)+ (3 x 4)[1 mark]
3 y 5 pats a
(53 * ta)+ Ga * Ta) + aa * Ta) At marl Sein 6: ole
IS i 20 20
2. _ 55. es is? ire ne
= 732 + 732 + 732 = T32 ~ 12 [FPmark b) P(ice cream) = P(C M1) + P(B. 1) [7 mark]
c) For both to be red there's only one possibility: aol? Do:
‘red then red’ [1 mark]. = (3 x §) + ( x g) [1 mark]
P(both red) = 5 x 4 = 25 = x, [1 mark] - oe SO.
= 7s a=
soll mand
d) ‘Both same colour’ is the complementary event Cc) P(chicken |ice cream)
of ‘not both same colour’. So P(not same colour)
= 1 —P(both same colour) [7 mark]. Both same = P(C|I) = on 2 [1 mark]
colour is either R, and R,, or W, and W,, or
G, and G,. So P(not same colour)
pCAh= « and P(l) = 4 (both from b) above)
PEMA Pe TS
erSes eet a) eo ° PCE = 75 * 309 [1 mark]
=1-[3 <4) + (aX A) + Ga * 7)
[1 mark] pen
= 79 [1 mark]
ee OO ee a
=) — i397 = Too eGo tt mark! a) Let E = ‘goes every week’ and R = ‘plans to renew
membership’. Then P(E) = 14 P(R) = A: and
(Alternatively, 1 mark for showing P(R,W, or = 20" ~ 520
R_G, or W,R, or W,G, or G,R, or GW), 1 mark P(E|R)= 73.
=10
for adding the 6 correct probabilities and 1 mark PiRia 2) = P(E |R)P(R) [1 mark]
for the correct answer.)
a) (i) J and K are independent, so
OA
= 3 [1 mark]
x x0
Pin Kyr= RO) Pik)! = O07 < OF) = 0:07
ma=aq'=
eal) 3 [1 mark]
[1 mark]

(ii) PJ UK) = PJ) + P(K) — PU 9K) [7 mark] b) P(E MQ R’) = P(E)— P(E R) [7 mark]
= 0.7 + 0.1 — 0.07 = 0.73 [1 mark] welt alge
b) Drawing a quick Venn Diagram often helps: ye aS 91 1
Cc)

ty
[1 mark]
So going to the club every week and renewing
membership are not independent [7 mark].
Or you could say that P(E) # P(E|R).
Eo27,
a) P(B’) = 1 — 0.2 = 0.8 [1 mark]
P(L|K’) = P(L 9 K’) = P(K’)
Now Ln kK’ = L — think about it — all of L is b) It’ll help if you draw a tree diagram before going
contained in K’, so L 1 K’ (the ‘bits in both L and
any further. If F = ‘roll a 6 on the fair dice’:*
K’ ) are just the bits in L. YEE
Therefore P(L M K’) = P(L)
\2 B Se
=1-P(K UJ) =1-0.73 = 0.27 [1 mark] 76 SE
P(K’) = 1 — P(K) = 1 — 0.1 = 0.9 [7 mark]
And so P(L|K’) = 0.27 + 0.9 = 0.3 [1 mark]
4 Start by drawing a tree diagram to make things easier.
Let C = ‘eats chicken’, B = ‘eats beef’, | = ‘eats ice You want to find P(B U F). Now, either at least
cream’ and S = ‘eats sponge pudding’. one of the dice shows a 6 or neither of them do,
so these are complementary events.
leet So P(B U F) = 1 — P(B’ MN F’) [7 mark]
es = 1 —($x2)= 1 —=4 [1 mark]
yl
4% ~B Or you can use the addition rule:
P(B U F) = P(B) + P(F) — P(BN F)

Answers
 c) P(exactly one 6 | at least one 6)
_ P(exactly one 6 1 at least one 6) b) The probability of the score being even is 2 =
P(at least one 6) pa vared
and the probability of ‘otherwise’ (the score bei SNIA
8
The next step might beabit easier to figure out if odd) is the same. The probability distribution is:
you draw a Venn diagram:

C) C can take 6 values, c= 5, 10, 15, 20, 25,

30 (each score x 5) and each one will have
‘exactly one 6’ NM ‘at least one 6’ = ‘exactly one 6’
(Look at the diagram — ‘exactly one 6’ is the light probability zt. The probability distribution is:
grey area, and ‘at least one 6’ is the light grey
area plus the dark grey bit. So the bit in common e [5] 10] 15]20] 25 |30
to both is just the light grey area.) Ly 13 Oe
Now, that means P(exactly one 6 ™ at least one 6) ol Gr Onl GeleGnlao
= P(BN F’) + P(B’ M F) — this is the light grey area
in the Venn diagram, Q3 a) We Pa w=0240440h-a=1
i.e. P(exactly one 6 fat least one 6) allx
So, a=1-0.2-0.4-0.1 =0.3
RSeso6s ed“6)930
oe Oo ee 8
710 (il) P(X = 2) = P(X = 2) + P(X= 3) + P=4)
[1 mark] =0.44+0.1+0.3 =0.8
P(at least one 6) = =it (from b) above). b) (i) SSPW=x)=6k=1
5
And all of this means... allx
P(exactly one 6| at least one 6) So, k= 2.
Baer
=j07 sa3 [1 mark] aeTo [1 mark] (ii) P(X >5)
= P(X = 9) + P(X = 16) + P(X = 25) + P(X = 36)
iee
6 3

Chapter 3: Discrete Random

(iii) PX= 10) = P(X= 16) + P(X= 25) + PLY= 36)
=jpedert
ee 6T Te
Variables (iv) PB < X < 15) = P(X = 4) + PIX = 9)
eae=2k=F
eee=>.
(v) P(X isdivisible
by 3) = PY = 9or 36)
1. Probability Distributions = P(X = 9) + P(X = 36)
Exercise 1.1 — Probability distributions ihe al
=2h=@=3.

and functions Q4 a) (i) SIP(K =x) =k4+ 4k + 9k = 14k =1

allx
Q1 a) (i) The discrete random variableX
is ‘number of tails’.
(ii) x could be O, 1, 2, 3 or 4.
b) (i) The discrete random variableX
is ‘number of orange balls selected’.
(ii) x could be O, 1 or 2.
b)
c) (i) The discrete random variableX
is ‘sum of the two dice scores’.
(ii) x could be 2, 3, 4, 5, 6, 7 or 8.

C) (i) P(X =x) =k+2k4+3k4+4k

4+3k4+ 2k +k
allx as

x
Son (6k="1 apie
sok = ze.

Answers
 F(S) Ree)
= P(X = 5) + P(X = 4) + P(X = 3)
+ P(X = 2) + PX = 1)
=0.2+0.2+0.3+4+02+01=1
Using all this information, the cumulative
Q5 a) Drawasample-space diagram to show all the en ves va
possible outcomes:
Score on dice 1 a
N Cia wh ea 0.ne0.3 (|
S
7

5 b) Add up the probabilities to work out the

5 values of F(x):
oO
TAs 12) 16 ie Pe eee
5
So the possible values that
Xcan take are T = P&s=1) = PX =—1) + Px=—2)
1, 2, 3, 4, 6,8, 9, 12 and 16. one
‘ Be : Soe
To find the probability of X taking each value, a = =)
count the number of outcomes that give the value a 3 ae 0) ee Oh Beat the=
and divide by the total number, 16. = pete
5 ee5
i a ly
Baa Rei ee.
1G =
a 67 F(1) = Ps 1 1) a,
QUBe I & ayp< aep<
252s aed 3
+ +

B=2)=7, =gre te =A, i oSaioPage

lluia Awot
B= Ole yp = 31 2,
Rexe—10) = 16 8’1 F(2) = P(X <2) 2al
2 1 = PUG = 2) POG Iie Pia =30)
Ho arg”1 Pela ie 6 + P(X = -1) + P(X = -2)
Pik = 16)= 21 = 1 ede ae
yl + 5 tetas: = |
Sr)
eS
ewe
So the probability distribution looks like this: Using all this information, the cumulative
distribution function is:

il c) Add up the probabilities to work out the

8 values of F(x):
Px=x=|% x=4 F(1) = PX s< 1) = PX = 1) = 0.3

Ms Y=
1 On6 F(2) = P(X < 2) = P(XY= 2) + P(X= 1)
16 20.2 02205
c) PB <X<10) F(3) = PW < 3) = P(X = 3) + P(X =2) + PX = 1)
= P(X = 4) + P(X = 6) + P(X = 8) + P(X=
9) = 03 023032078
Bt el he ees oh
= 16%
@ PBeulE eo F(4) SS)
=:|IHooa P(X
<4)
= P(X = 4) + P(X = 3) + P(X = 2) + P(X =1)
Exercise 1.2 — The cumulative =0.24 0.3 40.25 03.= 1
distribution function Using all this information, the cumulative
Q1 a) Add up the probabilities to work out CistriburiopiTag ction ty
the values of F(x): : x 1 2 |3 4
F(1) = P< 1) =P(X=1)=0.1 F(x) | 0:3) |lo.5 [08 1
(2) sR OX ap (XG)
Xe)
= 02 = 0.1 2088 d) Add up the probabilities to work out the
F(3) = P(X< 3) = P(X= 3) + PIX= 2) + P(X=1) vel Oho)
=0.3+0.2+0,1=0.6 F(2) = PX < 2) = P(X=2) = 4
F(4) = P(X< 4) e eae =
= P(X= 4) + P(X= 3) + P(X= 2) + P\X= 1) os - oe Se
=0.2+03 +02 201 =08 hr aay

Answers
 iy = — me ae 8) + P(X= 4) + P(X =2) So the cumulative distribution function
= Bae Se looks like this:

F(16) = P@< 16)

= PO 16) PC = 8) + P(X = 4) + P(X = 2)
le ae oS
16° 6 41sac
FG2) = PH 2)

+ P(X
=4)+P(X =2
aay Sel oe Pe 31
Dettmann Open Aare 2p Se. Here you're just using the fact that
F(64) = P(X< 64) P(X< O) + P(X > O) = 1 (all the
= P(X= 64) + P(X= 32) + P(X= 16) probabilities add up to 1).
as a :+ ee q ¥ Soy Q3 a) Use the probability function to work out the
ch ayUae pealinee ieNy elect ty | F(x).
Se iba A ales. ’
Using all this information, the cumulative EG) Oa) 8
distribution function is: Oe eee b ees
(2) = P(X¥= 2) + CESS gag tats ay

F(3) = P(X¥
= 3) + P(X
= 2) + P(X¥=1)
Sy ieee eee
“88.88
ae
BecauseE you are adding on a constant term of 4
Q2 a) You need to draw up P a table showing 6 the each time, the rest are easy to work out:
cumulative distribution function, so work
out the values of F(x): F(4) = 4 = 3 FS)= 2, EG) 8 = 3,
Fi y= P= 1)=028
F(2) = P(X = 2) + P(X = 1) =0.14+0.3=0.4 F7)=4,8° Fe)=8=1
8
F(3) = P(X= 3) + P(W= 2) + P(X=1) So the cumulative distribution function looks like
= 0.45 + 0.1 + 0.3 = 0.85 this:
F(4) = PX = 4) + P(X = 3) + PX = 2) + PX = 1)
= 0/5 O45 + 0. +: 0.3 =|
So the cumulative distribution function looks like
this:

() Ps 3) = FG) = 0.85 Q4 a) P(Y=5)=P(X¥<5)-PUX< 4)=1-0.9=0.1

(ii) P01 < X< 3) = P(X 3)-P(XS
1). P(X = 4) = P(X <4) — P(X < 3) = 0.9-0.6 = 0.3
= 0.85 = 0.3 = 0:55 P 3) = PX< 3) - P(X ¥ 2)
P 2) =P(X< 2) -P(X< 1)
b) Again, start by working out the values of F(x): p ) P(X < 1)
F(-2)=P(X =-2) =eea5

Patt Peeety
a a
apy = Dia Be 4 ON
= Iee
F(O) = P(X = 0) + P(X =—1) + PX =—2) —
Se eee Always check that the probabilities add up to 1 — if
> 1Oee neOon2> they don't, you know for sure that you've gone wrong.
a qo Waa et b) PX =1) =P(X< 1)-PIX< 0) =1-0.7=03
=i+t+2+5-2 P(X = 0)= P(X <0) - PIX <-1) = 0.7 -0.2 =0.5
: P(X =-1)
= P(X <-1) —-P(X<-2) =0.2-0.1 =0.1
BOS te +2)P(X
=-2)
r= 0) P(X
=-2) = P(X
<- -2) = 0.1 om
P(X=-1)
_ i 1 ;i Ze gl 25 So the probability distribution is:
= 5 Ge 5 oP To ateie sr =

; nee 1 ce
p(x) | 0.1 | 0.1 | 0.5 | 0.3

Answers
 P(X = 2) = P(X < 2) —P(X< 1) = F(2) - F(1)
<32) =1 —
c) P(X= 64) = P(X< 64) - P(X (26 1, Chee We to7 Bop 2 19
=~ 164 (A= 4964 6
=32) = PX<32)
P(X -Ps 16)= $—
P(X = 16) = P(X <16) —P(X< 8) = 37 7
P(X= 8)= PUY<8) Ps 4)= 4-9 =
<4) - ee 2=4-35
=4) = P(X
P(X
P(X = 2) =P(X< 2) = 30
So the probability distribution
Sw is:

= FG) = PX = 3)= DVPA=aS 1

oo[so]s]3]3|a1a
=4)=P(X< 4)-P(X<
1

3) =1-0.8=0.2
S072 ea.
For 2 to the power of something to be equal to 1,
the power must be 0, so3 -k=0 > k=3.
P(X = 3) = PX s 3)—=PXs2)= i a
2) = P(X
< 2) —P(X <1) =a— To find the probability distribution, you need to
POG ai) =e Gai) 083 find the probability of each outcome:
Now P(X
= 2) = P(X= 3) so 0.8-a=a-0.3, P(X= 3) = P(X <3).— PW s2) = FB)—F2)
soZa= 1.1, soa= 0:55. = 90-9)L 92-3 99 _ pat
SOIRCG=2) =) Be 3) = 0225 1 SS 1
S|
So the ee distribution is: De
POXG
2) = POD)
= PX 1) = FO) F(a)
= 92-3) __ 911-3
=) 9-1 9-2
oul eb zeta,4
2 4
Q6 a) F(3) = P(X <3) = HPUX
Il
PX = 1) POS) Ft) a2 aie?
So (3 +k
SF = 133ns vi
4+h=
So the table looks like:
>34+k=5>k=2
To find the probability distribution, you need to
find the probability of each outcome:
P(X= 3) = PW< 3) — PWs 2) = FG) = FQ)
_G+2P_@+27_25
2 DS DiS
4 F(x) means the probability that the larger score on
the dice is no larger than x. This means both dice
=) se ae) Cel) 0) must score no more than x, where x = 1, 2, 3, 4,
C2) leo) 6
i ay ae 5 or 6. The ees) that one dice will score
no more than x is=, so the ee that Hoth
will score no more ohar rid Se Se
6 6nd,
Two dice rolls are completely independent, so the
Cae as can just be multiplied together.
SOIR)== ax = 12a, SG.
36
To find the probability distribution you need to
by FG) BOS 3a) R(X a find the probability of each outcome:
5 allllke
go BEN
1s G+ipa 64 PU =16) = POs BlPOs Oe
2

363
>34+k=4>5>k=1 = 36 25 Sa
36 36 36
To find the probability distribution you need to Sg oeae
P(X = 5) POTS 5) BOGS diate
find the probability of each outcome:
=o = 10293
P(X= 3) = P(X< 3) — P(X< 2) = F(3) — F(2) 36, Gb 936.14 =e
hoa) ea) eso PIX= 4)= Ps 4)- PIX<3) = 2 —
64 64 64 64 64 = dee
36 36 36

Answers
 Q2 a) E(X) =[2 x 0.2] +[5 x 0.3]
~ eeoy a Nie
|S |& |
ale aly AIA
[oO |
als + [6 x 0.1]+[p x 0.4]
= 2.5 + 0.4p
SO 6.5=2.5+04p> p=10.
2. Expected Values, Mean and b) The probabilities add up to 1 so
Variance 0.5 + 02 -a+-02=>i1>a=0.1
E(X) =[4 x 0.5] +[8 x 0.2]
Exercise 2.1 — The expected value + [px a)+[15 x02]
Q1 a) E(X)= [0 x 0.2] + [1 x 0.2] + [2 x 0.2] = 6.6+0.1p
+ [3 x 0.2] + [4 x 0.2] =2
So 7.5=6.6+0.1p > p=9.
b) E(X)= [1 x ay]
14 t[2 x 14
G]+B x3]
14 Q3 a) The sum of the probabilities is 1 so
20s=a
TO" =D) STISCt, 02+a+01+b=15>a+b=0.7
=A
E(X) is 2.5 so
CMEC)= 320i 18 084 |= x 015)
= 1-9 02>, 2a-- 0:35 4b — 225
d) E(X) =[-2 x 0.1] + +1 x 0.2] + [1 x 0.2] = 2a+ 4b=2>a+2b=1
2px Oe Ob0s i)— 02 Rearranging the second equation to get a= 1 — 2b
e) First work out the value of k: and substituting this into the first equation we get:
The probabilities add up to 1 so (1 —2b)+b=0753b=03
3k+4k+5k+ 6k+7k=1
and putting b = 0.3 into the first equation we get
be | = i5=
Dijcs yee i
75° a=0.4.
Now you can work out E(x):
1 1 b) The sum of the probabilities is 1 so
EO x ae(l
D5 + 2)|462 M552
25 t 2) Oi a aseo4-
ON = il =>>asbb
= O06

+[3 x553 + 2)|+[4 x 54 +2) E(X) is 7.8 so

O33 42 Ha se 0 SE 8 == 7/3 = HE)sb BD = 3
+[5 x 565 + 2)| From the first equation, a = 0.8 — b, and
substituting this into the second equation, we get:
=V=34 7(0.8 — b) + 8b = 6.3
f) First work out the value of k: = 5.6 — 70> 80— 6:3
The probabilities add up to 1 so = 5.6-> b = 6.3
Keay oe an a ae
els AS | = b=0.7
Now the first equation is a + b = 0.8, and
ee Dee Sed substituting in b = 0.7 we get a = 0.1.

= (139)= Q4 a) (i) E(X) = [1 x 0.2] + [2 x 0.1] +-[3 x 0.25]

+ [4 x 0.25] + [5 x 0.2]
=O)
sk= 2
(ii) E(X?) = [1 x 0.2] + [4 x 0.1] + [9 x 0.25]
+ [16 x 0.25] + [25 x 0.2]
= 111135
b) (i) E(X) = [3 x 0.2] + [-2 x 0.1] + +1 x 0.25]
+ [0 x 0.25] + [1 x 0.2]
=-—0.85
(ii) EX?) = [9 x 0.2] + [4 x 0.1] + [1 x 0.25]
+ [0 x 0.25] + [1 x 0.2]
— 2265

Answers
 ©) (i) E(X)=[3 x 0.1] + [4 x 0.25] + [5 x 0.15] b) (i) The probabilities must add up to 1:
G0 aa Onle=nle=niai—Onoe
+ [7 x 0.3] + [9 x 0.2]
= 5.95 Gi) EX) = 3 x 0.31 #42 «023!
+ [-1 x 0.3] + [0 x 0.1]
(ii) E(X2) = [9 x 0.1] + [16 x 0.25] + [25 x 0.15]
eo
+ [49 x 0.3] + [81 x 0.2]
= 39) 55 E(X2) = [9 x 0.3] + [4 x 0.3]
+ {1 xi0.3] + [0.0.1]
=4.2
Exercise 2.2 — Variance Var) = 4.2 — (-1.87 = 0.96
1 a) E(X)=[1 x0.2] +[2x0.1)+ 3 x0.2 Q4 a) EX) =[3 x 0.2] + [4 x 0.3] + [6 x 0.1] + 0.4p
a a : mn i i 4 ue aka
seb=2.4+04
pee
E(X2) = [1 x 0.2] + [4 x 0.1] + [9 x 0.2]
+ [16 x 0.1] + [25 x 0.4] = 14
So Var(X) = E(X2) - (EW)? = 14 - 3.4? = 2.44 EUie= 0.2] Oana
+ [49
x0.4]
0.3] + [36 x 0.1
b) EW)=[1 x 5)+B x G+ 6x 9] = 29.8
+18 x ab} +19 x 35] + 10x 351 So Var(X)= 29.8— 5.2?= 2.76
99 b) The probabilities add up to 1:
to3sf)
= 35 (= 3.09 0.2+a+044+03=1 >a=0.1
2 = [1 x ft ‘ll
crqg!+ Bex“3! 0.2] + [4 x 0.1] + 0.4p+ [9 x 0.3]
= [1 x oe
-E(X) bas
E(X*) Zales
+ [64 x =] + (81 x 35] + [100 x 35] 5.7=3.3+04p > p=6
x 0.1]
E(X?) = [1 x 0.2]+ [16
541
oa Oe) + [36 x0.3]
x 0.4] + [81
So Var(X) = E(X?)— (E(X))? = 40.5
= 541_ (99) = -7.33to3s.4) Var(X) = 40.5 — 5.7? = 8.01
32 024 Q5 a) (i) All the probabilities add upto 1:
c) E(X)=[-2 x 0.2]+ ree arola) talear
+ [1x 0.1] + [12 x 0.4]= ot Se ae

Exar =4* 0.2) + i a + [0 x 0.2] 56

+[1x 0.1] + [4x 0.4] =2 (ii) EY) =[3x 5 x 37]4 [4 x a5
mee |
So Var(X)= E(X?) -— (U0) == 2.6 -0.4*= 2.44
lleheal eg
Q2 a) EW)=I x 314 2 51+ 13* 5) mereitengs
+[4x heisxd123 ee
= 1 1
es edie lax Hed x 554"
(ii) EX) = [9 x 553’] + [16
1 + (25x e]=11
+ (16x 2) i
+ [25 x sh] G
So Var(X)
= E(X2) — (E(X))?
=11 —32=2 aes
Tene tag pa Oh
b) a 50|* 2D eagit sal So Var(X) = 19:24 — 4.322 = 0.5776
+ [4x BOs : b) (i) ng eee add upto 1:
E(X?) ug + [4 xae
=“ [1x 35] 9
35] + [9 x 35] kTCR e LEem

+116 x 351=5 z+ g+it+pj=1

So Var(X)= E(X?) — (EU? = 32 - (22) 19
31 (= 0.689 to3 sf) - pati
= #1 Shamed
20 -
Q3 a) (i) The probabilities must add up to1: TG
02 Fa+04 40:1 =1 > a= 0:3. (32) (22) |
(ii) E(X)= [1 x 0.2] + [2 x 0.3] (ECO (9) , 4x (i9)}
+ [3 x 0.4] + [4 x 0.1] :
=2.4 (75)
15 x al (79)
+16 xsl2
E(X?)= [1 x 0.2] + [4 x 0.3] 6s 6
es+ [9 x 0.4] + [16 x 0.1] =260)
19 = 4.21to3sFf.
Var(X)= 6.6 — 2.42= 0.84 Notice that the numbers cancel in each bracket so
you're just left to add together four lots of =

Answers
 C) i) V=20- s

20
3
20
7 25 x {19)} 36 x {19)|
4
rasta
(ii)EV) = (17
x0.1] + 114
x 0.2]
fa
|0.2 0.2

= [igx 3]+ [tgx 4]

2 aeMoi ae + [5 x 0.2] = 10.4

Rs] E(V?) = [289 x 0.1] + [196 x 0.2]

+ [121 x 0.3] + [64 x 0.2] + [25 x 0.2]
= 1222
= 388 (= 18.91035.f,) Var(V) = 122.2 — 10.4? = 14.04
So Var(X) = yp - (8) = 40
(iii) E(V) = E(20 — 3X) = 20
—3E(X) = 20- (Gaxaore)
(1222 403 S43) = 10.4
Var(V) = Var(20 — 3X) = 9Var(X) = 9 x 1.56
= 14.04
Exercise 2.3 — Expected value and
d) (i) W=20+3X
variance of a function of X
Ww 23 AE
Q1 a) E(Y) =E(X¥ + 3)=E(X) +3=44+3=7
Var(Y) = Var(X+ 3) = Var(X) = 3 pW=w) | 0.1 | 02 | 03
b) E(Z) = E(5.X) = 5E(X) =5 x 4= 20 (it) E(W) = [23 x 0.1] + [26
x 0.2]
Var(Z) = Var(5.X) = 25Var(X) = 25 x 3 =75 + [29 x 0.3) + a x 0.2]+ [35 x 0.2]= 29.6
c) ES EX)
=2 EW 7 = 2 A 1 E(W*) = [529 x 0.1] + [676 x 0.2]
Var(W) = Var(2.X— 7) = 4Var(X) = 4 x 3 = 12 + [841 x 0.3] + [1024 x 0.2]
d) BA SEA = 2) = 7 = EO = 7 24)
= =1 + [1225 x 0.2]
Var(V) = Var(7 — 2X) = (-2)?Var(X) = 4 x 3 = 12 = 890.2

Q2 a) (i) y= xe 4 Var(W) = 890.2 — 29.62 = 14.04

i (iii) E(W) = E(20 + 3.X) = 20 + 3E(X)

P(Y=y= = 20+ (3 x 3.2) =29.6

Var(W) = Var(20 + 3X) = 9Var(X) = 9 x 1.56
= 14.04
(ii) E(Y) = [7 x 0.1]+ [10
x 0.2]
+113 x 0.3] + [16 x 0.2] + [19 x 0.2] = 13.6 Q3 a) (i) E(X) = [8 x 0.2] + [10 x 0.3]
+ [15 x 0.1] + [20 x 0.4]
E(Y*) = [49 x 0.1] + [100 x 0.2]
= (4a
+ [169 x 0.3] + [256 x 0.2] + [361 x 0.2]
= 199 E(X?) = [64 x 0.2] + [100 x 0.3]
+ [225 x 0.1] + [400 x 0.4]
Var(Y) = 199 — 13.62 = 14.04
= 225.3
(iti) E(Y) = E(3.X+ 4) = 3E(X) + 4 = (3 x 3.2) +4
Var(X) = 225.3 — 14.17 = 26.49
= 186 F
Var(Y) = Var(3.X
+ 4) = 9Var(X) = 9 x 1.56 (ii) E(Y) = E(4X+ 3) = 4E(X) + 3
=a O4 = (4x 14.1) +3 =59.4
Notice that you get the same values for the mean Var(Y) = Var(4xX+ 3) = 16Var(X) = 16 x 26.49
and variance in parts (ii) and (iii). = 423.84
(iii) E(Z) = E(50 — 2X) = 50 — 2E(X)
b) (eee

erm
ese enc = taexen al) ee
Var(Z)= Var(50— 2X)= (-2)?Var(X)
= 4 x 26.49
=105.96
(ii) E(Z) = [-1 x 0.1] + [2 x 0.2] b) (i) EX)= [-4x 4]+ ]
1 1
+ [5 x 0.3] + [8 x 0.2] + [11 x 0.2] = 5.6 + [2

E(Z) = [1 x 0.1] + [4 x 0.2] = —24 =-1.78t03 sf.

+ [25 x 0.3] + [64 x 0.2] + [121 x 0.2] Qo
- [16 xall
E(X?)
= al
5] +se al
5
= 45.4 x

Var(Z) = 45.4 — 5.6? = 14.04 +i4x 4+oe 351 + 136x

— 4) = 3E(X) — 4 = (3 x 3.2) -4
(iii) E(Z) = E(3X ooo
= 5 = 10.4
to 3 s-f.
=O
7407
Var(Z) = Var(3X— 4) = 9Var(X) = 9 x 1.56 1024
= 14.04 = PERO Disan

Answers
 1 1 1
—2X) = 7 — 2E() = 7 - (2 x (-1.78))
(ii) E(Y) = E(7 Q4 a) ED) = [1 x a [3 x z|+ [5 x Al
= 10.56 1 1 1
ee xe]
Var(¥)=Var(7 — 2X)= (-2)VarlX) = 4 x 7.23 +[7 =xg]+[9OK at
xe] +[1
= 28.9 to 3 sf. =6
(iii) E(Z) = E(7 + 2X) = 7 + 2E(X) ; ;
= 7 +(2 x (-1.78)) - b) ED) =[1 x 2]+[9 x g]+ [25 xs]
= 3.44 to 3 sf. : ;
Var(Z) =Var(7 + 2X) = (+2)°Var(X) = 4 x 7.23 + [49 x 2] +[81 xg] +[121 xg]
ee as = 143 _ 47.7t03sf.
Q4 a) E(X)=[10 x 0.2] + [12 x 0.3] + [13 x 0.1] 3
+ [15 x 0.3] + [16 x 0.1] = 13 Var(D) = 47.7— 6 = 11.7 to3.s4.
E(X2) = [100 x 0.2] + [144 x 0.3] + [169 x 0.1] Aevihigass
+ (225 x 0.3] + (256 < 0.1] = 173.2 Q5. a) = EO) =
Var(X)= 173.2 — 132= 4.2 (6 =a+ iy =—1
b) E(Y) = E(26 — mX) = 26 — mE(X) = 26 - 13m g? = Val) ae
So 0 =26-13m > m=2. (8-1+1¥-1_ 63
Var(¥) = Var(26 — mX) = (-m)*Var(X) = 4 x 4.2 =F etiatot eee
= 16.8 b) Substitute your values for the mean and variance
Cc). EQSEBxH=e) =3EX) c= 6.x 13)-e=39-e¢ into the inequality and rearrange — then it’s just a
So 30 =39-e>c¢=9 simple probability question.
Var(Z) = Var(3X — c) = 9Var(X) = 9 x 4.2 = 37.8 i P(X — |< 0) = P(X - 4.5\< 55)
= P(|\X-4.5|< 2.291)

Exercise 2.4 — The discrete uniform Hi the modulumefiy< 4:5 ss less tham zy Vint
Cieues : means that the size (i.e. ignoring any minus signs)
distribution of X— 4.5 is less than 2.291. For example, when
X = 3, X-4.5 =-1.5. The modulus of this is 1.5,
which is less than 2.291.
|X— 4.5|< 2.291 is satisfied by
X = 3, 4, 5 and 6.
So P(X
— |< 0) = PLX= 3, 4, 5, 6)
Seat) ae here ear
ips oyee ete
ee ee eA ae) 2A Q6 a) Any single digit number means that Y could take
> 2 = iw 2) any value between 0 and 9 inclusive.
oy r Gee
pak = S$" = OE2 Seaao ee45
o? =Var(Y)=Cae
1
4 &= (SOHRI
T2 bo 8=ecore
77 = 8.25

EY)= 24°Cee
= + = 135 b) P(\¥—n|
>0)=P(\Y-4.5| > ¥8.25)
= P(|Y—4.5| > 2.87
ek Saal (5—12+ 1f—1 45 : | : ae
Vi aoe ee = |Y—4.5| > 2.87 is satisfied by Y= 0, 1, 8 and 9,
qe so P(|Y — | > o) = P(Y=0, 1, 8,9)
Siehae lee eee

Pizimeh hades DDL Lie When, Te |05) c) P(|\¥—p| > 20)= P(\Y—4.5|
>278.25)
T0 | 10} 10| 10] 10|10|10| 10 | 10 | 10 = P(|¥-4.5|
>5.74)
Be 91 S26 : There are no values
that Y can take such
E(2)= 25" = US = 155 that|¥— 4.5]
>5.74, 50
PY —al > 20)=0.
Varig)
JO mes ee oP
OS A
ety Dae=1 _
99 | Q7 ate ~
ee
233 DEs ieee

Be 2 2 ile
= 6, a Vata
So the standard deviation is the square root of the 12
variance, i.e. /8.25
= 2.87 to 3s.f. ._ 7 a1 a eee
7 12 asbees
b) EC) SECS 27 = VEZ) Ok) =
Var(X) = Var(2 — 2Z) = (-2)*Var(Z) =4 x 4=16

Answers
Review Exercise — Chapter 3 = 0)=P(R <0)=011
Q1 a) All the probabilities have to add up to 1. PRS = PRS 1) PR = OVS 0520 = 0.4
So 0.5+k+k+3k= Oot okay Wes 5k
= 0.5, R= PR = 2) PR = S105 = 05
hese Onl
So the probability distribution ofR is:
b) P(Y <2) =P(Y=0)+P(¥=1)=0.5+0.1 =06.
r 0 | 2
Q2 PW = 0.2)=PW=0.2) = 0.2 PUR") Ol Ora lmOn5
PW < 0.3) = PW = 0.2) + PW =03)=0.4
POSRS)=.05
P(W < 0.4) = P(W = 0.2) + P(W = 0.3)
+ P(W = 0.4) = 0.7 There are 5 possible outcomes, and the probability of
P(W < 0.5) = P(W = 0.2) + P(W = 0.3) each of them is k, sok=1+5=0.2.
+ P(W = 0.4) + P(W = 0.5) = Mean of
X =on 4 =),
So the cumulative distribution function of W is : (Ae Ois let tk geil
Naa ON i a ae

Q3
cater [erfaTe
a) P(X<1)=P(X¥=1)=
a) As always, the probabilities have to add up to 1:
sosbvnc pp
ae (ore
5 of a
54 A aes
PX <2) = P(X¥=2)+ PiX= )=se+H= b) E(x) =(1 x 2)+ (2x4)
P(X
s 3) = P(X = 3) + PX = 2) + P(X¥=1)
Baer dis os +(3 xg)
+(4x 54)
5s aE STs25 145 24359 490.. Sia)
57
P(X s 4) = P(X = 4) + P(X =3) 24 cr2dee
+ P(X = 2) + P(X = 1)
orem, 6 14 31, 16 E(X’) = (1? x 2)+ (2? x 4)
BS 2 be AT 2S. By 25
+ (3°
3)+(# x33)
AS OE 27ia BONES 1 DONE 53
Se
WISETO
BSS 5 a5
Olin 251 gen 5 os 22
S55 24 ieee
So the cumulative FST function is:
Var(x) = EC) — [e(ap = 33— (12)
_ 424 — 361_ 63
64 ~+~«64
E(2X — 1) = 2E(X)—1
19, cons

Var(2X — 1) a= 2Nat(X) =A xX oe
036316

eagle Ses es Q7 a) EX)=(1 x 0.1) + 2 x 0.2) + GB x 0.25)

oe es + (4 x 0:2)+(5 x 0.1) + (6 x 0:15)
= 3.45
b) PUX< 1)=PX=1)=3 b) Var(X) = E(x?) — (E(X))?
a E(X?) = (1 x 0.1) + (4 x 0.2) + 9 x 0.25)
+ (16 x 0.2) + (25 x 0.1) + G6 x 0.15) = 14.25
In general, P(X< x) = So Var(X) = 14.25 — 3.45? = 2.3475
So the cumulative distribution function is:

Answers
 Q8_ To find the cumulative distribution function, you need eZ a) All the probabilities must add up to 1, so
to find P(X< x). This is the probability that the highest 2k+3k+k+k=1, i.e. 7k =1, and so k=
score rolled is no more than x. So all 3 dice must roll [1 mark]
a score no more than x. The individual probability of
one dice rolling a score of no more than x is - — so b) P(X< 0) = P(X= 0) =4 [1 mark]
the probability iseall 3 dice roll a score no more P(X< 1) = P(X= 0) + P(X= 1) = 2 [1 mark]
than x is (z) = (since the probabilities of the
316
6 P(X< 2)= P(X = 0) + P(X= 1)
individual dice rolls are ia [1 mark]
+ P(X =)e=2
SO) Fs) (G78)= Sl,
DpSey Oy 8 P(X< 3)= PX= 0) 4+ P(X= 1)
316
6
Now to find the probability distribution, you need to + P(X =2)+P(X=3)=1
work out P(X= x) for each x. [1 mark]

P(X = 1) = PX ieDifae -D10

3

bed . 49, pies So the cumulative distribution function is as in

P(X = 2) = P(X s 2)— P(X < Te ee the following table:
1)= 576 - 576 = 276
wre x 0 1 al ae -
: S3
ee
a AD
3
% 1
P(X = 3) = P(X < 3) -P(X s 2) = a7— - a7@ P(X <x) % %
= 976
P(X = 4) = P< 4)- P< 3)= 4-
3 3
P(x?) ance)
3 = SF
ar woes
P(X¥=5)=P(X<5)-P(¥<4)= e 3
2-- 2 = Ot61 a [1 mark]
asset a using part a).)
P(X = 6) = P(X < 6) —
So
— the distribution is: a)

wap 5
AQ. |) 3Z |) iL
a6 316 21GH 2061 216.) 216 [1 mark]

b) Mean = ae = 4.5 [1 mark]

: TO ya _ 99
Exam-Style Questions — Chapter 3 Variance = ess [1 mark] = 1

a) The ee of getting 3 heads is: = 8.25 [1 mark]

5Teageabel
X=5 x =3 3 mark] c) P(X < 4.5) = P(X= 0) + P(X=1)
The probablyof sete 2 heads is: + P(X= 2) + P(X = 3) + P(X= 4) 7 mark]
3X- 5x 4 x 3-3 a (multiply by 3 because =0.5 [1 mark]
any of the three coins could be the tail — the a) PX=1)=a, PX = 2) =2a, P(X =3)=3a. *
order in which the heads and the tail occur isn’t Therefore the total probability is 3a + 2a + a = 6a.
important). [7 mark] This must equal 1, so a = z. [1 mark]
Similarly the probability of getting 1 head is: a a e(2 xe)
2
b) E(X) =(1 Xe) x)3
Led § lee
A ERBGA GER =seats [1 mark]
And the probability of getting no heads is:
tlie le eel a 5 [1 mark]
Do 7S Deeg
So the pera of 1 or no heads c) E(X?) = Var(X) + [E(X)P
2 aar = ant3 mark] = 2 + (5)= 2 5 a
: [1 mark]
cae the ee glistDungD of X is:
= 34 = 6 [1 mark]
d) E(3X+ 4) = 3E(X)+4=3 xf+4= 11
b) You need the probability that ¥>10p [1 mark] [1 mark]

This is just P(X= 20p) = 2 [1 mark] Var(3X+ 4) = 3’Var(X) = 9 x a [1 mark]

=5 [1 mark]

180 Answers
 E(X) = (0 x 0.4) +(1 x 0.3)
a)
+ (2X 0.2) + (3 x 0.1)
Chapter 4: Correlation and
=0+03+40.4+4 0.3 [1 mark] Regression
=1 [1 mark]
b) E(6X + 8) = 6E(X¥)+8=6+48 [1 mark] 1. Correlation
= 14 [1 mark]
C) The formula for variance is
Exercise 1.1 — Scatter diagrams and
Mant x)= EC) — [Ey correlation
So first work out E(X?): QI a) 25
E(X*) = (0? x 0.4) + (1? x 0.3) oOo

+(2° x 0.2) (3° x,0.1)

NS
Ae
= 0.3 + 0.8 + 0.9 [1 mark] visits
Cinema
=2 [1 mark] vores
Seti

Then complete the formula by using your answer to Distance from centre, km
part a): ‘
b) Negative correlation
Var(X) = EX’) — [E(X)P = 2 — (1?) [1 mark]
Q2 a) ~30
=1 [1 mark]
d) Var(aX+ b) = a’Var(X)
Var(> — 3X)
= (-3)Var(X) = 9Var(X) [1 mark]
= ox
QO 2 & & & 10 12
=9 [1 mark] Length (cm)
a) E(X) = [0 x a] + [1 x 5d] + [2 x 5] + [3 x 0.2] b) Positive correlation
=7b+0.6 [1 mark] c) The circumference of 3.5 cm
$0 1.3 =7b+0.6 [1 mark] => b=0.1 [1 mark] or the length of 11.9 cm.
All the probabilities must add up to 1 so:
a+0.5+0.1+0.2=1
=> a=0.2 [1 mark]
[1 mark]
Exercise 1.2 — S,,, S,, and S,,
Q1 a) Add rows to the table for x”, y? and xy. And
b) EGS x+ 3) = 73E) +3 [1 mark] include a column showing the total of each row.
= (79 x 1.3)+3=4 [1 mark] actetal Gat! OralPace seeee 7See ae oe
12 O i)
C) E(X2) = [0? x 0.2] + [12 x 0.5] + [22 x 0.1] |x2|
1121 ||36 | |81 ||16 ||64| a | 25 | 347
+ [32 x 0.2] = 2.7 [1 mark] 1600
Var(X) = Bt*) — (E(X))? [7 mark] iy [26478 |162)20715221
6017385]
—1.32=1.01 [1 mark].
SO >) = 45, yp = 92, Doe S847,
d) Var(4fia 1 Var(X) [1 mark]
dv = 1600 and >)xy = 738.
=1.01 [1 mark]
Sy ae A5 x 92
b) Sa ey 2 2 738 5

1 pe B aw —§
= 146.57 (to 2 d.p.)
; a) _ 45°
= 57.71 (to 2 dp.)
Sa ae
eae
apres = 1600 —
92?
-

= 390.86 (to 2 d.p.)

Ea 29 x- 109
Q2 a) S,=Yiy z 589
= 43.2

S =x nes 167 — 97
22 =-1.2
swy Soe le 2031 — 102 = 345.2

Answers
 Q3 a) Make atable including p’, q* and pq.
pee xD ye 206x 50" Include a final column showing the
total of each row.
¢ Os 2062 _
xX fafa tefistte
7e pee arabe
gts. he
5, i pane (29) = 326 at = 76 lp?|169|81|225|100|64|121|
9
r25149| 4 |16/| 64[ 1 |
Q3_ S, n= ee = 1880 — 1156114,
pq - 4" 2X14 = 305 350

7 OTD 115?
=) pS 208s 6
oY = 172, >)pq = 350
= 178.83 (to 2 d.p.)

ie og — 4
S,,= yay = 2762 — 1142
14 _=596 e 2
S = pt — 22) = 1100 — 2 = 42
PP

Exercise 1.3 — Product moment

2

s = ig
99
2D = 177 — 32 = 44
correlation coefficient
c) This means:
Q1 SF
DAY
te
BIS OCS.
ee ee a ea
—

Cae (Qua) ee eee

2 v SopSeq

WS = — 18
SS OR 04.19 io Cosisioanio
fi
V¥42 x 44 teases
= 343.7333...
Q4 a) First find >)/ = 43 and dw = 30.4.
S,= yy“OD
— 2 saad ZZ
473 752
—£2- = 98
oa
Then:

This means:
I
S. = Sipser = 181,75 - Bx304
oe Sey
¥ SxxSyy
= 18.35
=50 ey fig. 1y = 258 — 432 = 26.875
= — 0.2 72 (to
(to33sig.
sig. fig.) 5,= yp
343.7333... x 98
=
Q2 a) Make atable including x’, y° and xy. Include a (uM _ 198.2 304° = 12.68
final column showing the total of each row.
Sap = 4 w

b) This means:
Siw
i

18.35
= (26.875
1839. eesipttig:
X 12.68 2 0 994ttouO 23 Steteo
SS eae = 91148 — 1403 x 519 c) This value for r is very close to 1, which shows
n 8 that there is a very strong positive correlation
= 128.375 between the length ofa leaf (/) and its width (w).
The longer a leaf is (i.e. the higher the value of J),
S.=7 DxOre 246395 — OS
—“= 1403?
the wider it tends to be (i.e. the higher the value
of w).
= 343.875

S,=Lee
Dy 2 5192 _= 262.875
= 33933 — 312
b) This means:
——
Se
Sia Se
2 128.375 = ADT G08 sig.fig.)
V¥343.875 X 262.875

Ee
Answers
 Q5_ Make atable including a2, s? and as.
b) This is fairly close to -1, and so there is a strong
Include a final column showing the
negative correlation between a patient's weight
total of each row. (w) and their kidney function (k). The higher a
patient's weight is (i.e. the higher the value of w),
3249 | 4225 |8836 7744
the lower their kidney function tends to be (i.e.
79.21 | 23.04] 29.16] 7.84 the lower the value of 4).
507.3| 312 [507.6 |246.4

So D)a = 650, )s = 54.2, Dia? = 48 608, Exercise 1.4 — Coded data

dis” = 364.52; as. 33732:5 Q1 Create a new table showing the coded data values u
and Eh
v. extra rows for w2, v2 and wv.
SS Sas Lads 2370752puss? |
=-181.9444.., eo
a
Spee’ aoe)Y = 48608 — war by
= 1663.5555... Luv
So (sy = 364.52 242° So )ju=27, vy =29) dv = 169,
va = 1 y= (el
= 38.1155... Then:
This means:
= uv muda ae DTD
6
iG = ae

V Sages

= 0,5
— 181.9444...
Vvl663.5559.5 30 lllooe.
2.

= — 0.723 (to3 sig.fig.)

S.=yiuv 4) =1649 ar = 47s
This is fairly close to -1, and so there is a fairly strong Svw = iv 2 Ouvy = 175
aa 29°
6
negative correlation between the age of a patient (a)
=BA4O38 330
and their score on the memory test (s). The older a
patient is (i.e. the higher the value of a), the lower

of
This means:
their score on the memory test tends to be (i.e. the
lower the value of s). —

Suis
Q6 a) Make atable including w”, k* and wk.
Include a final column showing the Se S00 eB dip)
A/a eS 42 0930 8
total of each row.
Because wu and v are related by linear transformations
[ow [Meet [e742] 56!|p3) 79 [esa | 64)|71s |80
88 |
to x and y, this means that the PMCC between x and y
|| 79 | 62 |23 [70] 48 [91 | a4|68 |37| must also be 0.012 (to 3 d.p.)
46.24 Create a new table showing the coded data values u
[wk|521.4|606.8|220.8|589.3[379.2 |491.4]537.6|482.8 |325.6 and v. Include extra rows for wv’, v and uv.
So Dw = 675, ik = 58.3, hw =51 975, | 1 48
om TOM Sm om Mom 0
ee = 421.69, wk = 4154.9 a 81 | 121] 64 | 356
See Dds"
ae eee oe x 58.3

= 217.6

Sa" =
Oe. AS ars 6757
9 So Diu = 48, v= 120; or = 6,
div? = 1982, Suv = 759
= 1350
Then:
Se (Qa a9 16912283
n 9
5,= Yiw mud = 759 — 48 X120 _ 39
= 44.0355...
aa are 48" = 68
This means: 2

LS 5, = yy — QF ~ 1982 120° = 182

Vv SHSEE

= — 0.892 (to 3sig. fig.)

~ 77350
x44.0355...

Answers
 This means: This means:
San
i

V Saga

= = = 0.35 i(to3 sig.tig. a SiS ee 0.151(to 3 sig.fig.)

68 x 182 io? Seay ¥227.875 X 560 :
Because u and v are related by linear transformations Because wu and v are related by linear transformations
to x and y, this means that the PMCC between x and y to w ands, this means that the PMCC between w and
must also be 0.351 (to 3 sig.fig.) s must also be -0.151 (to 3 sig.fig.)
Q3 Create a new table showing the coded data values u
and v. Include extra rows for 12, v2 and uv. Q5 Your coding needs to make the numbers easier to
ro[1[2]3]+ls[*[7 [#21 use. So use the coding: u = x — 4000, and v = 55
Setarpasarbe De DaPa ae Then create a new table showing the coded data
fw [oo [a[4 {9 |16|25|36|49|64|20 values u and v. Include extra rows for wu’, v? and uv.
[ [1024]625]441]269]49[r44]64|16|4 [2656 Le [olnya|
Fao [25[a2[sr[20]oo[8[2s[16]29 Olan]
Peta fetse [2fas
So: i= 36, ey =O, ou = XO",
pepe eferd fetaba
ov = 2656, » wy = 298
Then:
Par|8 [40 86[a5[rao]a2[349|
So Siu = 83, v = 25, pie = 1275, Dov = 143,
oe Dat) 2 — 36128
=914 yw = 349
Then:
S,,= Dw i —-S=
Su ne
= 204— 362
32 _= 60
2 aoe 2m = 349 —83X25
S= >= Quy = 2656 — 48
= 3.1666...
= 835.5555...
SE 4F (Quy e075 Oe
This means: = 126.8333...
2

S.= yy = 143 — 23° = 38.8333...

r = iS

See
vw

care
r= aia 5.oso tari ie:
JOO 835.5555. G20 Ma 2ete NE.) This means:
Because wu and v are related by linear transformations a ‘Sigs

to s and t, this means that the PMCC between s and t 7 J Sinvy

must also be -0.956 (to3 sig.fig.) = 3.1666... = 0.0451 (to 3 sig.fig.)

126.8333... X 38.8333...
Q4 Create a new table showing the coded data values u So the correlation coefficient for x and y
and v. Include extra rows for w?, v? and uv. is 0.0451 (to 3 sig. fig.).:
ree] 4 [ofl 7[a] 9 [wf
Pe[afasfone [ars [30[m3[20 2. Linear Regression
Exercise 2.1 — Explanatory and response
variables
Q1 Explanatory variable: the time spent practising the
So Div = 101, Div = 120, }i? = 1503, piano each week
yYv = 25,60) yw = 1461 Response variable: the number of mistakes made in a
Then: test at the end of the week
It is the amount of practice that would determine the
_ 101 ee 45
performance in the test, not the other way around.
Q2 Explanatory variable: the age of a second-hand car
Response variable: the value of a second-hand car
It is the age of the car that would affect its value, not
the other way around.

Answers
 Q3 Explanatory variable: the population of a town Call the equation of the regression liney = a + bx.
Response variable: the number of phone calls made Then the gradient of the regression line is b,
in a town in a week where:
It is the population that would affect the number of
calls, not the other way around. b=ee
S. = Saao —0.44457...

Q4 Explanatory variable: the amount of sunlight falling = -0.445 (to 3 sig. fig.)
on a plant in an experiment And the intercept of the regression line is a,
Response variable: the growth rate of a plant in an where:
experiment
It is the amount of sunlight that would affect the
a=>y 7 —br = trae ee
growth rate, not the other way around. = Be — (— 0.44457...) x 330
228
(Or you could say that the amount of sunlight can be
directly controlled, as this is an experiment.) =wit pantse ant fig.).
So the equation of the regression line of
yonx is: y= 31.2 —0.445x
_ Exercise 2.2 — Regression lines C) The value ofb tells you that the temperature
Q1 Call the equation of the regression liney = a + bx. decreases by approximately 0.445 °C for every
Then the gradient of the regression line is b, where: extra degree of latitude (since when x increases
E eSll 95 4oy by 1, y decreases by 0.445).
b= 5 = 739-4 = ':82855.. = 1.83 (to3 sig. fig.) The value of a suggests that on the equator (when
And the intercept of the regression line is a, where: x = 0) the mean annual temperature is about
dy = pet Sd 2E
a2=y—br=
d) (i) When x = 19, the residual is:
-us — (1.82855...) x e = 24 —(31.16006... — 0.44457... x 19)
= 1.29 (to 3 sig. fig.)
= ee(to 3 sig. fig.).
(ii) When x = 41, the residual is:
So the equation of the regression line of e = 10—(31.16006... — 0.44457... x 41)
NOM MIS2 0) — 223 2a Oon = -2.93 (to3 sig. fig.)
Q2 a) Q3 a) You first need to find S,, and S.,.

eeDee 92 x 264
ee
= 167.2

pe?
= 2
= ey.ce (072 92? 725.6
OF 10) 20 308 e240 50) COR 70 If the equation of the regression line is P = a + be,
b) Start by working out the four summations >)x, then the gradient of the regression line is b,
yD, ee and >ay: where:
It’s best to draw a table. == Be
Se —905.6
1672>.=O

Te pepey
a[|
41345

4 = 0.741 (to3 sig. fig.)

96fa And the intercept of the regression line is a,

where:
Pepe
2oe pose
324| 300 |550| 468
al

So }ix = 350, Diy = 156, ix? = 15 696, = 204 _ (0.741134...) x

Dixy = 3928
= Oe (to 3 sig. fig.).
Then:
So the equation of the regression line of
og Deen 6 sOr aise Poneis: P=19.6
+ 0.741e
n
b) The value ofb tells you that a graduate
ates? employee's salary will generally increase by £741
# 2
Se Quast = 15696 — 320" = 3446 per year (since when eincreases by 1, P increases
by 0.741, which is equivalent to £741, as the
salaries are given in thousands of pounds).
The value ofa tells you that a newly employed
graduate (when e = 0) typically earns a salary of
£19 600.

Answers
Exercise 2.3 — Interpolation and - q=-0.9 + 0.1p, x — => 20
and so 10y—3 =—0.9 + 0.1 x
extrapolation
Now rearrange this into the formy = a + bx.
Ql a) y is the response variable (since this is the
regression line ofy on x). So 10y = -0.9 + 0.05x — (0.05 x 20) + 3
j.e. 10y = 0.05x + 1.1
b) (i) 1.67 +0.107 x 5 = 2.205 Ory = 0.005x + 0.11
(ii) 1.67 + 0.107 x 20 = 3.81
3 y = 17.4 -0.78x, so t—45 = 17.4 — 0.78(w— 60)
Q2 a) 103 — 4.57 x 4 = 84.72 2 Rearrange this to give:
This is interpolation (since 4 is between 2 and
t= 17.4 —0.78w + (0.78 x 60) + 45,
15). This estimate should be reliable. or t = -0.78w + 109.2
b) 103 —4.57 x 20=11.6
This is extrapolation (since 20 is greater than the Q4 a) Sy LLY — 618 28 x 124
10
largest observed value of x, which was 15).
This estimate may not be reliable. = 270.8
CM 103457 <7 = 71-01
This is interpolation (since 7 is between 2 and
—(xy = 140
S,.= Dx 26° =6 61.6
—4
15). This estimate should be reliable. Call the equation of the regression inay =a-+bx.
Q3 a) VA 20, =139 Then the gradient of the regression line is b,
This is interpolation (since 20 is between 17 and where:
35, which are the values ofx between which data
was collected). This estimate should be reliable. oe_ Spall
oe2708:
Gs = 4.396103..
b) 1.4x50+7=77 = 4.396 (to4 sig. fig.)
This is extrapolation (since 50 is greater than the And the intercept of the regression line is a,
largest value ofx for which data was collected). where:
This estimate may not be reliable.
Q4 a) 58.8 — 2.47 x 7 = 41.51 — so the volunteer a=y—br= uy — ue
would be predicted to have approximately 42
spots. This is interpolation (since 7 is between 2 _ 124.— (4.396103...) x 28
and 22, which are the values ofx between which == Saree = ae ae fig.).
data was collected). This estimate should be
So the equation of the regression line of
reliable.
yonx is: y= 0.09091 + 4.396x
b) 58.8 —2.47 x 0 = 58.8 — so the volunteer would
be predicted to have approximately 59 spots. b) v-7= 0.09091 + 4.396 x (75 — 2)
This is extrapolation (since 0 is less than 2, which So v =-1.70 + 0.0440s
was the smallest value of d for which data was
collected). This estimate may not be reliable.
Q5 a) Start by working out the four summations >)x,
DW ee and oe
c) Using the formula for d = 30 is extrapolation, It's best to draw a table.
since 30 is greater than 22, the largest value of
d for which data was collected. The model isn’t
valid for d = 30, since you can’t have a negative
number of spots. But this doesn’t mean that the
regression equation is wrong.

Exercise 2.4 — Linear regression with $6 ))7'=28, iy =0, Dix = 188 ay 75

coded data Then:
Q1 Substitute the expressions definingp and q into the
equation of the regression line for g on p. This gives 5, = La — 2X2? Se
you an equation involving justx and y. = 2975
q = 40
+ 2p, soy— 50 = 40 + 2(x - 7)
Now rearrange this so it is in the form y = a + bx, Su= oe = (xy = 188 — 28 = 57.333...
where a and b are constants.
y=40 4+ 2x-14 +50 = 76 + 2x Call the equation of the regression liney = a + bx.
So the equation of the regression line ofy on x is Then the gradient of the regression line is b,
y= 76 + 2x. where:

_ Sy =275 _
b= <= 573337 = 4.796511...
= -4.80 (to3 sig. fig.)

Answers
 And the intercept of the regression line is a, c) Explanatory: a person’s disposable income
where: Response: a person’s spending on luxuries
d) Explanatory: the number of cups of tea drunk per
= §0 ~ (4.796511...) x 28
day
Response: the number of trips to the loo per day
= 22.4 (to 3sig. fig.). e) Explanatory: the number of festival tickets sold
So the equation of the regression line of Response: the number of pairs of Wellington
yon iS y= 22.4 — 4.80% boots bought
b) Substitute expressions for x and y into your Q3 a) cil)=Sy 68e za aot = 46.504
regression equation: t— 29 = 22.4 — 4.8035
This means that: t= 51.4 — 0.480s (ies = 570455 — ted 10") = 183.24
b) beeSee ena
esas = oes...
Review Exercise — Chapter 4 = 3.94 (to3 sig. fig.).
Q1 a) 4547
40 x
35 x c) a=w—br, wherew= uu = 109.9
30 x
25 x x x x
and7 = 2 =e 4
20 x
15
x
10
Soa = 109.9 — 3.9403... x 51.74
5

0 7 : : - 7
x

- x
= — 93.9711...— 94.0
(to3sig. fig.).
50 100 150 200 250 300.
d) The equation of the regression line is:
b) First you need to find these values: w = 3.94r — 94.0
dix = 1880, Dy = 247, x’ = 410400, e) When r= 60, the regression line gives an estimate
for w of: w = 3.94 x 60 — 94.0 = 142.4 ¢
> Y = 6899 and >)xy = 40600.
f) This estimate might not be very reliable because
Then it uses an r-value from outside the range of the
Sits a
= De Se OO
1880 x 247
ah
original data. It is extrapolation.
You'll be given the equations for finding a regression line
= —5836 — but you still need to know how to use them, otherwise
the formula booklet will just be a blur.
S
u (ux)es =_ 410 400 io
1880’ And you need to practise USING them of course...
= 56960 Q4 Rearrange the expressions P = r—5 and QO= 8Ww to get
formulas for r and w. These arer =P +5
Wy
Dy r= a ; cit iden
LIF eB 6899 —
Da7F
OM = 798.1 and w = 80.
Substitute these expressions for r and w into the
This means: regression line’s equation from Question 3.
a= De Wi= S94 03s xr 93-97 Iles,
Vv Sa So and so 8Q = 3.9403... x (P + 5) — 93.9711...
OD
= F56560
x7581 0.866 (to 3sig.fig.) Now rearrange to get an equation of the form
Q=a+bP:
c) The PMCC tells you that there is a strong negative SO 309403e. x Pi GA9403m 5 SI 93,9 /Alilnes)
correlation between drink volume and alcohol i.e. O=0.493P — 9.28.
concentration — cocktails with smaller volumes
tend to have higher concentrations of alcohol.
Don't panic about that PMCC equation — you need
Exam-Style Questions — Chapter 4
1 a) 0.90 7»
to know how to USE it, but they give you the formula in 0.80 x x
the exam, so you don't need to REMEMBER tt. 0.70 x

Explanatory: the annual number of sunny days

0.60
Q2 a) 0.50 x x
Response: the annual number of volleyball- 0.40
related injuries 0.30
0.20
b) Explanatory: the annual number of rainy days 0.10 x

Response: the annual number of Monopoly-

0.00 + r T r r Vx
0 4 6 8 10

related injuries [2 marks for all points plotted correctly, or 1

mark if at least 3 points are plotted correctly.]

Answers
 b) You need to work out these sums: Then the gradient b is given by:
dix = 36, Hy = 3.94, be oe
Se
— 0.29015
0.1226
25 6ees.
ex = 204, = 246076, ay — 17.66 = 2,367 (to 3d.p.).
[1 mark]
Then:
And the intercept a is given by:
Sy = Dixy —Ld 3.04
= 17.66 ex at BO
= = 0.07
Al == may
¥ Para
_ 67.49
Ser ee 2) 3006824 26.12
8
Ot) 36F
= 0.709 (to 3d.p.).
[1 mark]
(24676 = aa = 0.52715
2 2

=> (29) So the regression line has equation:

y = 2.367x + 0.709
[3 marks available — 1 for each correct term] [1 mark]
This means: To plot the line, find two points that the line
eae BA. = 007 passes through. A regression line always passes
V Say V42 X 0.52715 through (x,y), which here is (3.27, 8.44). Then
= — 0.015 (to 3d.p.). [1 mark] putx = 3 (say) to find that the line also passes
c) This very small value for the correlation through (3, 7.81).
coefficient tells you that there appears to be only Now plot these points (in circles) on your scatter
a very weak linear relationship between the two diagram, and draw the regression line through
variables (or perhaps no linear relationship at all) them [1 mark for plotting the line correctly].
Remember, you get marks for method as well as correct
[1 mark].
answers, so take it step by step and show all your
Dunia 2 workings.
8.8
Cc) y= 2367 x 37154305709
2 8.6
&
‘ as
Residual for d) (ii)
= 8.17 (tos-signtig.),
» 8.4 Residual for d) (i)
(or 8.16 if b = 2.36663... used) [7 mark]
E 8.2 a
Ss
a Regression line (part b))
This should be reliable, since we are using
interpolation within the range of x for which
7.8
x we have data [7 mark].
7.6
aa
x (ii) y = 2.367 x 3.88 + 0.709 = 9.89 (to 3 sig.
3 3.1 oz 3.3 3.4 3.5)
20-metre time fig.)
[1 mark]
[2 marks for all points plotted correctly, or 1
mark if at least 3 points are plotted correctly.] This could be unreliable, since we are
b) It’s best to make a table like this one first: extrapolating beyond the range of the data
[1 mark].
po) A 2|S
Bisealteiy
heels a5) 4
x 3.09 d) (i) residual = 8.25 ~ (2.367 x 3.32 + 0.709)
= -0.317 (3 sig. fig.),
(or -0.316 if b = 2.36663... used)
[1 mark for calculation, 1 mark for plotting
50.) x = 26.12) > y=.67.49, residual correctly]
dX = 85.4044, and >) xy = 220.645. (ii) residual = 8.59 — (2.367 x 3.27 + 0.709)
= 0.141 (3 sig. fig.),
[2 marks for at least three correct totals,
(or 0.142 if b = 2.36663... used)
or 1 mark if one total found correctly.]
[1 mark for calculation, 1 mark for plotting
Then: residual correctly]
Se 70.645.
wa = 0.29015 a) Put the values into the PMCC formula:
[1 mark] sae Wadd 12666
S.= 85.4044 — oie 201226 VSxSy 310880 x 788.95
= 0.809 (to3sig. fig.).
[1 mark]
[1 mark for correctly substituting the values into
the PMCC formula, and 1 mark for the correct
final answer.]

Answers
b) There is a strong positive correlation between
the miles cycled in the morning and calories
consumed for lunch. Generally, the further they
cycled, the more they ate [7 mark].

C) 0.809 (to 3 sig. fig.) [1 mark]

Remember — the PMCC won't be affectedifyou
multiply a set of data values by the same number, so
changing the data from miles to km doesn't change it.
a) Ate = 12E5) OS TO Ws 2O 2 Be
hee (O°G028-— 1285) 9331624 [1 mark for at least 4 points plotted correctly,
Atx= [47, or 2 marks if all points plotted correctly.]
y = 211.599 + (9.602 x 14.7) = 352.7484 b) Start by working out the summations
[1 mark for each value of y correctly Sox = 115 [1 mark] and Diy = 544 [1 mark].
calculated]
Then:
b) Using the equation:
‘Residual = Observed y — Estimated y’:
Ate = i225:
§,=Day
—22D?
Residual = 332.5 — 331.624 = 0.876 [1 mark]
= 6446 — aot [1 mark]
Atx = 14.7: Residual = 352.1 — 352.7484 = —1374 [1 mark]
= —0.6484 [71 mark]

a) Find the summations >)p = 65 [1 mark] and }\q ey

Ss1975 —
aac
2 = 321.875
[1mark]
= 0 [71 mark].
Then:
C) The gradient of the regression line is b, where:
e od 65 x 0 _ Sy _ —1374 7
b= 5 = 397-975 [1 mark] = 4.26873...
hee >)Pa 5 843
10
[1 mark] = —4.27 (to3 sig. fig.) [7 mark]
And the intercept of the regression line is a,
= 843 [1 mark]
where:
_ (Spy as 65? a=yo
ox = ——
a
aae ee te 10) n

= 148.5 [1 mark] = 244 — (— 4.26873...) x LP 1 mark]

S,,= iG te — $548 586 = 02 = 129 (to 3 sig. fig.) [7 mark].

d) The regression line must pass through the point

= 548 586 [7 mark] (x, ¥) = (14.375, 68). It must also pass through
Se the point (0, 129).
b) a
VS ppSaq 1407
Bae OSS
ek ee 238 [1 mark] 120
¥148.5 x 548586
100
= 0.0934 (to 3 sig.fig.) [7 mark]
80 c xm

Cc) 0.0934 (to3 sig. fig.) [1 mark] 607-4

Becausep and q are related by linear 40 2 H

transformations to d and #, this means that the 20 #

PMCC between d and t must be the same as that x

between p and q [7 mark]. ° 5 IO 1s 20 25 30)

[1 mark if the line passes through one correct
d) The PMCC between d andt is very low [7 mark],
which suggests that the length of a run the point, or 2 marks if line is drawn correctly.]
day before a race has little impact upon the e) Brand E appears to be overpriced [7 mark] — it
performance of these athletes [7 mark]. is much higher than the line, whereas the other
points lie very close to it [7 mark].

Answers
 e) The &(z) table doesn’t contain negative values of
Chapter 5: The Normal z, $O use a ibe to aesan equivalent area:
Distribution
2. The Standard Normal S1a74i0' 01.14

Distribution, Z So, P(Z<-1.14)=P(Z> 1.14)

= ]1-P(Z<1.14) = 1-0.8729 =0.1271
Exercise 2.1 — The standard normal
f) The &(z) table doesn’t contain negative values of
distribution z, so use a sketch to show an equivalent area:
Q1 a) Using the ®(z) table, 6(1.87) = 0.9693 f(z) f(z)

b) Using the ®(z) table, (0.39) = 0.6517

c) Using the ®(z) table, &(0.99) = 0.8389
d) Using the ®(z) table, (3.15) = 0.9992 =I Z One? (05m ae
Q2 a) P(Z> 2.48) =1-P(Z< 2.48) So, P(Z < -2.06) = P(Z = 2.06)
= 1-0.9934 = 0.0066 =1=P(Z < 2.06) = 1 — 0.9803 = 0.0197
b) P(Z> 0.85) = 1-P(Z< 0.85) We haven't drawn diagrams for the answers to Q4 below,
= 1 —058023.= 051977 but it often makes it easier to see how to tackle a question if
©) P(Z> 1.23) =1-P(Z< 1.23) you do a quick sketch.
= 1-—0.8907 = 0.1093 Q4 ae RUInG4 Zeon IZ ae) ee P(Z < 1.34)
d) P(Z>0.14)=1-P(Z<0.14) = 0.9854 — 0.9099 = 0.0755
= 1-0.5557 = 0.4443 b)) ROW6 <7 1292) = BAe?) P(Z < 0.76)
Q3 a) Usea sketch to find an area you can look up in = 0.9726 — 0.7764 = 0.1962
the table: c) PEI45=27201 )aRG—017)— P(Z < -1.45)
.- A = P(Z< ON) = RZ=e )
Remember, P(Z< —z) = P(Z= z).
= P(Z< 0.17)— [1 — P(Z< 1.45)]
= 0.5675 — (1 — 0.9265) = 0.4940
=3.35 0
d) P(-2.14<Z< 1.65) = P(Z< 1.65) — P(Z< -2.14)
SO, F(Z 335) — IZ
=<3:3) 109996
= P(Z< 1.65) — P(Z2 2.14)
b) Use asketch to find an area you can look up in = P(Z< 1.65) — [1 — P(Z< 2.14)]
the table: = 0.9505 — (1 — 0.9838) = 0.9343
‘% K.
e) P(-1.66 < Z< 1.66) = P(Z< 1.66) — P(Z < —1.66)
= P(Z< 1.66) — P(Z = 1.66)
= P(Z< 1.66) — [1 — P(Z< 1.66)]
=0.240 00.24 = 0.9515 = (1 — 0.9515) = 0.9030

So, P(Z > —0.24)= P(Z

< 0.24) = 0.5948 f) P(-0.34 <Z< 0.34) = P(Z< 0.34) — P(Z < -0.34)
= P(Z < 0.34) — P(Z= 0.34)
c) Use asketch to find an area you can look up in = P(Z< 0.34) — [1 —P(Z << 0.34)]
the table: = 0.6331 — (1 — 0.6331) = 0.2662
. (2) AN.
g) P(-3.25 < Z<-2.48) = P(2.48 < Z < 3.25)
= P(Z < 3.25) — P(Z < 2.48)
= 0.9994 — 0.9934 = 0.0060
-1.210 01.21 h) PC1.11<2< O17) =PO017 =Z2 1.14)
So, P(Z> -1.21) P(Z < 1.21) = 0.8869 =P(Z<1.11)—P(Z< 0.17)
= 0.8665 — 0.5675 = 0.2990
d) The ®(z) table doesn’t contain negative values of
There are different ways of answering some of these, so don't
z, $0 use a sketch to show an equivalent area: worry ifyou used a different method to get the same answer.
ne NN

-0.629 00.62

So, P(Z < -0.62)= P(Z

>0.62)
= 1-P(Z< 0.62)=1—-0.7324= 0.2676

Ee
Answers
Exercise 2.2 — Using the tables in b) P(0<Z<z) =0.4854
reverse = P(Z < z) = 0.4854 + P(Z < 0)
= 0.4854 + 0.5 = 0.9854
Q1 a) Using the ®(z) table, ®(z) = 0.8577 for z = 1.07
®(z) = 0.9854 for z = 2.18
b) Using the ®(z) table, ®(z) = 0.8264 for z = 0.94 Q5 a) Tala Zea) nD
c) P(Z<z) = 0.3783 = P(Z>-z) = 0.3783 SHO<6g<4 <7) S024
= P(Zs -z) = 0.6217 > z=0.31 >z=-0.31 = P(Z < -2) = 0.2422 + 0.5 = 0.7422
d) P(Z<z)=0.004 > P(Z>~-z) = 0.004 ®(z) = 0.7422 tells you that -z = 0.65 = z =-0.65
=> P(Z< -z) = 0.996 > ~=2.65 > z =-2.65 b) PzZ<Z<0)=0.1443
e) PZ > 2) = 0.758 P(Z <—2) = 0:758 = P(0<Z<-z) =0.1443
= —z = 0.70 = z=-0.70 = P(Z < -z) = 0.1443 + 0.5 = 0.6443
®(z) = 0.6443 tells you that -z = 0.37 = z = -0.37
f) P(Z>z)=0.9441 = P(Z<-z) = 0.9441
2s al a) P(1.5 <Z<z)=0.0406
26
= P(Z <z)—P(IZs 1.5) = 0.0406
2) R(Z> Zz) = 04801 = PZ =z) = 0.5199
= P(Z < z) = 0.0406 + P(Z < 1.5)
= 4 = (0.05 = PZ <z) = 0.0406 + 0.9332 = 0.9738
h) P(Z>z) =0.0951 = P(Z < z) = 0.9049 ®(z) = 0.9738 for z = 1.94
eo | b) POM8= Zz) = 10,0691
Q2 a) Using the percentage-points table, p = 0.005 for | => P(Z< z) - P(Z< 0.58) = 0.0691
Z=)2:5758 = P(Z< z) = 0.0691 + P(Z< 0.58)
b) Using the percentage-points table, p = 0.2 for => PZ <z) = 0.0691 + 0.7190 = 0.7881
z= 0.8416 (z) = 0.7881 for z = 0.80
C) A) On (Zz) = 3 = 244 C) ae1.3 <Z <z)= 0.871
P(GA<2z)—= PZ s—1-3)= 0.871
dR Az) = 0S RZ)
= 015 = 2 —1.0364
P(Z
< z) = 0.871
+ P(Z < -1.3)
©) ela Z4 7) le—alaZe)0a PZ <z)=0:871 +l —PZ<13))
— 420 ol— 2 —— 126 Ilo ate(Z<z)= 0.871 + (1 — 0.9032) = 0.9678
=>
Pre 2 Oo PZ = =z) = 0.15 (z) = 0.9678 for z= 1.85
= 7 = 1.0364 > z=-1.0364 d) P(-0.54 < Z<z)=0.667
gs) P\ZSz)= 0.6 => PZ <2) = 0:6 => PiZe =z) =0.4 P(Z < z)— P(Z < -0.54)= 0.667
7 0223) = 0.2539 P(Z< z)= 0.667 + P(Z< -0.54)
P(Z
( < z))== 0.667 + [1 —-P(Z< 0.54)]
h) RZ 2) — 0199 — RZ <7) 099
=> PZ( <z)= 0.667 + (1 — 0.7054)= 0.9616
iZiee2) 0.0 23.63) — 7 =| 283263
®(z) = 0.9616 for z = 1.77
Q3 a) P(-z < Z<z) =0.599, so the remaining area is
e) PGaZ = 027) 05458
1 — 0.599 = 0.401. Using symmetry,
— (AO) Az Ona4 8
P(Z =z) = 0.401 = 2 = 0.2005
= P(Z < z) = P(Z < 0.27) — 0.5458
= 4<— 4) = OS = 2 = One
= P(Z < z) = 0.6064 — 0.5458 = 0.0606
An alternative way of doing this is to divide O.599 by
But you can't look up O.0606 in the table, so there's
2 to get PCO < Z<z) =O2995. Then add O.5 to get
still some work to do...
AE < B) = OVS = P(Z = -z) = 0.0606 = PZ < -z) = 0.9394
b) Piz <Z<z)=0.9426, so the remaining area is ®(z) = 0.9394 tells you that -z = 1.55 =>z=-1.55
1 — 0.9426 = 0.0574. Using symmetry, P(g < Z<-1.25) = 0.0949
RiGee) 0)05745 2 = 010237 et
> ie -1.25)— P(Z< z) = 0.0949
=a) 10 Ole | 2) P(Z < z) = PZ < -1.25) — 0.0949
c) Pi-z<Z<z)=(0.4, so the remaining area is P(Z < z)= PZ 1.25) — 0.0949
1—0.4=0.6. Using symmetry, P(Z < z) =[1 —P(Z< 1.25)] — 0.0949
P(Z =z) = 0.6 +2 =0.3. Using the P(Z< z)= (1 — 0.8944)— 0.0949 = 0.0107
percentage-points table, p = 0.3 for z = 0.5244 = Pg = -z)= 0.0107 = P\Z < =z) = 0.9893
a
d) P(-z < Z<z) = 0.98, so the remaining area is ®(z) = 0.9893 tells you that —z = 2.30 > z= -2.30
1 — 0.98 = 0.02. Using symmetry,
P(Z > z) = 0.02 + 2 =0.01. Using the
percentage-points table, p = 0.01 for z = 2.3263
Q4 a) P(O < Z< z) = 0.3869
= P(Z < z) = 0.3869 + P(Z< 0)
= 0.3869 + 0.5 = 0.8869
@(z) = 0.8869 for z = 1.21

Answers
3. Normal Distributions and Q7 a) P(240 < X< 280)
- p/240 260 280
— 260
Z-Tables ieemip Le
ee ee
Exercise 3.1 — Converting to the P( Z < 1.33)— P(Z< -1.33)
Z distribution P(Z<133)-PZ21.35),
PZ
P( <1.33) =11 = PZ =1.33)
QI a) PIX<50)=P(Z< wt = P(Z < 2) =0.9772 = 9 x PIZ< 1.33) —1
= 2 x 0.9082 — 1 = 0.8164
b) PX< 43) = P(Z < ae
b) P(232 <X< 288)
=P(Z< 0.6)= 0.7257 " pilee 00 288— 260
= P/z> OG
PY > 28) 2A = P( iS teats )
Q2 a)
Xr> 28)= P(Z> AEA)=PP(Z
> 1.63)
= Peron = ee)
= 1—P(Z< 1.63) = 1— 0.9484 = 0.0516 = P(Z < 1.87) — P(Z< -1.87)

b) P(¥> 25) =P(z> x4

= P(Z < 1.87) — P(Z= 1.87)
(> 25)= P(Z>227A) =PZ>0.41) =PG=1.87)— [fl RZ een
= 1-P(Z< 0.41) = 1— 0.6591 = 0.3409 = 20 P(A= 1o7)—|
= 2 x 0.9693 —1 = 0.9386
107
a) P(¥> 107) = P{z > 107 =—120"
120 _ pz s -2.06)
Q3
aaaera ae Q8 a) PL <a) = 0.9938 = P(Z < 4579) - 0.9938
= PZ 2206).= 0.9803 Using the (2) table, &(@) = 0.9938 for z = 2.50.
b) P(X> 115)=P(Z 115120
Z > LS =120/ ~ pz5 -0.79) So, £570 = 2.5 > a=2.5x4+70=80
V40 5
= P(Z < 0.79)= 0.7852 b) PLX< b) = 0.7734 = P(Z < 279) = 0.7734
Q4 a) Px <15)=P(Z =bol
b—1/)-
15 P(Z<-0.67) Using the ®(z) table, ®(z) = 0.7734 for z = 0.75.
= P(Z> 0.67) = 1 —-(Z< 0.67) So, 279 = 0.75 > b= 0.75 x4+70=73
= 1-0.7486= 0.2514
Q9 a) P(X> m) = 0.0102 = P(Z > wo = 0.0102
b) P(X < 12) = P(Z <4=1)- BIZ= =167) m—95\ _
=> P(Z = i) = 0.9898
SIZ 167) S | WAS a7)
= 1—0.9525 = 0.0475
Using the ®(z) table, B(z) = 0.9898 for z = 2.32.
0 50)
63 —— 50
7 < 63 So, me = 232 => m= 2x 545 = 106.6
Q5 a) P62 =X < 63) = p(22=250
=PO4<27< 2.6) b) PLX > ) = 0.2296 = P(Z > Lay = 0.2296
= P(Z < 2.6) — P(Z< 0.4) = P(z7<12) = 0.7704
= 0.9953 — 0.6554 = 0.3399
Using the ®(z) table, ®(z) = 0.7704 for z = 0.74.
b) P(57 <X<66) = p(32=50 < 7 < 66=50)
So,‘— = 0.74
> t=0.74x5+95 = 98.7
oleae aan)
= P(Z< 3.2) —P(Z< 1.4) Q10 a) PX <¢) = 0.1251 = P(Z<£ awe) 01251
= 0.9993 — 0.9192 = 0.0801
= P(Z = A8
Ti)—c\ _
= 0.1251
Q6 a) P(0.45 < X < 0.55)
— p(0.45 — 0.6 O55 OG Switch c and 48 around to get —z.
i P( 0.2 aa Or )
48 —c
= P-0.75 < Z<-0.25) = P(Z S25
<< SS) ) 0.8749
S1HOLS <4 ()7/5)) Using the ®(z) table, ®(z) = 0.8749 for z= 1.15.
=P\(Z.10Y/5) — Zs O22)
= 0.7734 — 0.5987 = 0.1747 So,48 —€ = 1.15 c= 48-115 x 10 =365
b) O53) <ACz (0) Exe}) b) PUX<d)= 0.0096 = P(Z< d—T0 48) = 0.0096
_=Pl2
p/0.53 5
— 0.6 0.58
— 0.6 ) = P(Z> 48 4) = 0.0096
= PCE0.35 < Z <-0.1) ~ 48 adh )= 0.9904
= P(Z sao
=P ONee 4035)
= P(Z< 0.35) — P(Z< 0.1) Using the ®(z) table, &(z) = 0.9904 for z = 2.34.
= 0.6368 — 0.5398 = 0.097
So, 4 =2.34>d=48-234x10=24.6

Answers
Q11 a) PUY > w) = oe
0.9177 = P(Z > ¥— 23) - 0.9177 c) P(6<L< 11) =P(2584 <7<U— 84)
al
| tO 7 aS 1.10272 0.84)
Using the ®() table, ®(z) = 0.9177 forz = 1.39. P(Z< 0.84)— P(Z< -1.10)
= P(Z < 0.84)— P(Z= 1.10)
So,BAW = 139 > w=73 -1.39 x 6 = 64.66 P(Z< 0.84)— [1 — P(Z< 1.10)]
b) PUY >) = 0.6664 = P(Z > £73) = 0.6664 nen n ME ee
Vie ead 6 Q2 a) Let F ~ N(36, 6?) represent the length of time
oF P(Z Se )= 0.6664 taken to replace red blood cells, in days.
Using the ®(z) table, ®(z) = 0.6664 for z = 0.43. P(T< 28) = P(Z<2838)
ViMeet
50) = CiKe = 0.43 =>k= 73 -0.43 x 6 = 70.42 = P(Z < -1.33)= P(Z> 1.33)= 1 — P(Z< 1.33)
= 1 — 0.9082 = 0.0918
Q12 P(.18-a<X<
18+ a) =0.899
b) Let b = the number of days taken by Bella.
= p(Bas—18 <7 < oewe
18 —a—18 18
Then, P(T'> b) = 0.063.
a
= Pe a 2 < wr) = 0.899 = IOV =— 0.063
= P(Z > 22)
The remaining area = 1 — 0.899 = 0.101. = P(Z aUre 0.9370
So, P(Z
(
= 05)
4.) =
O:
0.101 + 2 = 0.0505. 4
= Using the ®(z) table, ®(z) = 0.9370 for z = 1.53.
= P(Z< 24.)= 0.9495 So, 2=38
36 = 1.53
pee the ®(z) table, ®(z) = 0.9495 forz = 1.64. => b=1.53 x 6 436 = 45.18 days
So, ae
4 Ke7.642995 1.645004 = 0.82 Don't forget to say days’ in your answer — you need to
answer the question in the context it was asked.
Q13 P(170 <X <2) =0:377
170 = 170 Peng Q3 a) Let 7 ~ N(51, 2.1’) represent the ‘personal best’
= P( AO <Z< AO )= 0.377 time taken to run 400 m in seconds.
= P(0<Z< $510) - 0377 ee)
= P(Z> -0.81) = P(Z< 0.81) = 0.7910
= P(z< tle
ay) - P(Z < 0) = 0.377 So, 79.1% are slower than Gary.
PZ = ai = 0377 = 05 = 0.877 b) Let a= the time to beat.
a Then, P(T< a)= 0.2.
Using the ®(z) table, ®(z) = 0.877 for
z = 1.16.
P(Z
< 4") = 0.2
So, 0 = 1.16 = t= 1.16 x 40 + 170 = 216.4 zi, zi)=
Q14 P(107.6 < X< v) = 0.1677
rz > 2h=4) = 0.2
107.6 Using the percentage-points table,
= eae— 98 4 yot08\
ae )= 0.1677
p = 0.2 forz = 0.8416
= P(0.64 < Z< = ie28) = 0.1677 So,
254; = 0.8416
— 98
= P(Z < T=") — PZ< 0.64) = 0.1677 = a=51-2.1 x 0.8416 = 49.2 s (to 3 s.f.)
= P(Z Sec 1 8) = 0.1677 + 0.7389 = 0.9066 Q4 a) Let Z ~ N(300, 50?) represent the lifetime of a
Using the ®(z) table, ®(z) = 0.9066 for z = 1.32. battery in hours. a
200 —
So, X98 = 1.32 3 y= 1.32 x15 +98= 117.8 <200) = P(Z < “aa==)
P(L
=a Da P4sS DSi S4s 2)
° : ° ; = | — 019772 — 0,0228
Exercise 3.2 — The normal distribution
f . ° °
in real-life situations 80 — 300)
b) PUL > 380) = P(z> 380 300

QI a) LetZ ~ N(8.4, 3.12) represent the length of a = (GO) eA 16)

worm in cm. = 1-0.9452 = 0.0548
P@=<95)= P(Z <9564) c) Assuming that the lifetimes of the batteries are
j independent, the probability that both last at least
Bae O20) S20 380 hours = 0.0548 x 0.0548 = 0.0030 (to 2 s.f.).
P(L > 10) =P ery 1064ae P(Z > 0.52) If two things
Zi are independent,
0 it means iMyou can multip!Ply
ea eet 2 3.1 le their probabilities together.
= 1-—P(Z< 0.52) =1—0.6985= 0.3015

Answers
 d) P(160<L<h) = 0.9746 c) P(X> 528) =0.1292
160 — 300
> Reg a h — 300 5 0.9 746 eS p(Z> a
100
= 0.1292
h — 300 a Pp(Z < eal = 0.8708
ea 2:8 G2 ee a 0.9746
Using the ®(z) table, ®(z) = 0.8708 forz = 1.13.
=a p(z< 2= 300 ‘ P(Z<-2.8)= 0.9746 528 —
eles
rene7 < b=300Ms P(Z >2.8) = 0.9746 100
=> p= 528-1.13 x 100=415
seppieaaonyiin it (Z < 2.8)] = 0.9746
d) P(X< 11.06) = 0.0322
= P(Za he aT )= 0.9746
+(1 —0.9974) Thi) Oley poy
=>Pp(Z<d ee = 0.0322
=> P(Z< h =300)- 0.9772
Using the ®(z) table, ®(z) = 0.9772 for z = 2.0.
adPp(Z= ro = 0.0322
So,h— 300 2 n=? x 50 + 300
= 400 =P(z< a = 0.9678
Q5 a) Let M~ N(60, 3’) represent the mass of an egg in Using the ®(z) table, ®(z) = 0.9678 for z = 1.85.
grams. 11,06 —= 185
Oa
P(M > 60 — m) = 0.9525
== 1-05 x0:4 1-06 —sies
= P(Z >o0 60) = 019525
= P(Z > =r) = 0.9525 e) P(X> 1.52) =0.9938

= P(Z = =) = 0.9525 = P(Z> 52 Wy = 0.9938

For")
p—1. =) ca
Using the ®(z) table, ®(z) = 0.9525 for z = 1.67. = P(Z< ou 1 0.9938
So, 2
3 = (6/17
= ldo7 63 = SO Using the ®(z) table, ®(z) = 0.9938 for z = 2.50.
= 5 grams to the nearest gram. pu — 1.52
=)
0.02
b) Let c = the maximum mass of an egg in one of —> leo OLO2e4 alo theo,
farmer Elizabeth's sponge cakes.
Then, P(M< c) = 0.1. Start with a sketch showing what you know about X.
So, P(Z < 80) 2081
=e pes =on
Using the percentage-points table, p = 0.1 for 6.45 H
IN
20.17
B= ||PAINS
= =< = L286 So, P(X< 20.17) = 0.95 + 0.025 = 0.975
=S @S C60 = 3 5% 1 QRS

= 56.1552 = 56.2, to 3 s.f.

= P(z< 37
Sa)
=#) - 0.975
So, the maximum mass is 56.2 grams. Using the ®(z) table, ®(z) = 0.975 for z = 1.96.
20a ee
2a = 1.96

Exercise 3.3 — Finding the mean = p = 20.17 — 1.96 x 3.5 = 13.31

and standard deviation of a normal There are different ways you could go about this question.
distribution For example, you also know that P(X > 20.17) = 0.025, so
you can look up p = 0.025 in the percentage-points table.
QI a) P(X <23)=0.9332 > P(Z< ee
6
)= 0.9332 Or, using the symmetry of the graph, you know that \1 is
exactly in the middle of 6.45 and 20.17, so you can simply
Using the ®(z) table, &(z) = 0.9332 forz = 1.50. find the average of these 2 values.
23—
6
Pea155p=23-15x6=14
a) P(X< 53)=0.8944
b) P(X <57)=0.9970 > P(Z< ane8 *)= 0.9970 = P(Z< =
ooo = P(Z<22)= 0.8944
Using the ®(z) table, ®(z) = 0.9970 for z = 2.75. Using the ®(z) table, (2) = 0.8944 forz = 1.25.
Of bh
=./) — p= eo XO 5
ead => 2=1.250=5+125=4

Answers
 b) PY< 528) = 0.7734 Subtracting equation 1 from equation 2 gives:
= P(Z < 228 =510) _ P(Z < 8) = 0.7734 36 —30-- (1) = 2.60- 1.460
=O oo 5
Using the ®(z) table, ®(z) = O:7734 tor z = 0175. Putting o = 5 into equation 1 gives:
= 18-075 30=18:0.75 =24 w=30—=1.4x5 = 23
SO =28 aie! ora. 5
C) P(X> 24) = 0.0367 b) P(X< 4) = 0.9332 = P/ et= | ZA 93380
= p(Z > 2412) x P(z>Z2)= 0.0367 ®(z) = 0.9332 for
z= 1.50
= P(Z <7) 7 = 0.9633 4— wu
See es 1.5 > 4-= . 1)
1.50 (equation
Using the ®(z) table, ®(z) = 019633 forz= 1.79.
P(X < 4.3)= 0.9987 => P(Z = Enea
3 = 0.9987
aTee oo a= = 1.79 = 3.91 (to3 sf.) ®(z) = 0.9987 for z = 3.00
d) P(X< 0.95) = 0.3085 = =F
43 —p
= 3 > 4.3 - 1 = 30 (equation 2)
= P(z< 0.95 —0.98) = P(z< — 9.03) Subtracting equation 1 from equation 2 gives:
4.3-4--(-) =30-1.56
= 0.3085 = 055 = SoS
sor= 0)
= P(Z> 2.03) = 0.3085 Putting o = 0.2 into equation 1 gives:
= 41 sc0.2ee 7
= P(Zs 2.03) = 0.6915 SOIL = 3.7 ance =O2
Using the ®(z) table, &(z) = 0.6915 forz=0.5.
C) P(X< 20) = 0.7881 = P(Z< sar =) = 0.7881
0.03
Se Cath = 0.5 > 0 = 0.03 + 0.5 = 0.06 ®(z) = 0.7881 forz = 0.80

e) P(X> 4.85) = 0.8365 a0 FP = 0.8 = 20-1 = 0.80 (equation 1)

= p(z > 483=9.6) = P(Z>— 0. 2.22) P(X< 14) = 0.0548 => P(Z nal ae a) = 0.0548
= 0.8365 = P(zZ>F——14) = 0.0548
= P(Z<O23) = 0.8365
SAS pay _= 0.9452
Using the ®(z) table, ®(z) = 0.8365 for z = 0.98.
®(z) = 0.9452 forz= 1.60
=> 222 = 0.98 3.0 =0.75 = 0.98
= 0.765 (to 3 sf.) =F el = 1.6 > —14 = 1.60 (equation 2)
Adding equations 1 and 2 gives:
Q4 Start with a sketch showing what you know about X. 20-14-w+p=0.80+ 1.60
=>6=240>0=2.5
Putting o = 2.5 into equation 2 gives:
p=1.6x2.5+14=18
Sot= Sanda=25
61 68 75
d) POC<1690)) = 0297015
SOME 75) = Onl5 = P(z< P2=#) - 0.9713
= P(z
> B88) — p(z>Z) = 0.15
oO
O©¢) = 019713 forz= 1.90
Using the percentage-points table, 696 —
0.15 forz = 1.0364.
joes
aaa = 1.9 = 696 — = 1.90 (equation 1)

= 2=1.03640= 7 = 1.0364 =6:75 (to:3 sf.) P(X< 592) = 0.2420

se P) = 0.9192 => P(Z a
592 — pw )= 0.2420
Q5 a) P(X< 30)= 0.9192 = P(Z< oO
oe
®(z) = 0.9192 for z= 1.40 = P(Z ye = 0.2420
30 i
aut
= 1.4 > 30-p= 1.40 (equation 1) p= 592)
=> P(Z oe | 0.7580
P(X< 36)= 0.9953 => P(Z < gor a) = 0.9953 O(z) = 01758 tor z— 0:70
&(z) = 0.9953 for z = 2.60 =
yw — 592 = 0.7 > p — 592 = 0.70 (equation 2)
=> a P = 2.6 > 36 —p= 2.60 (equation 2) Adding equations 1 and 2 gives:
696 — 592 -w+p=1.90+ 0.70
= 042.60 =o —40
Putting o = 40 into equation 2 gives:
pw = 0.7 x 40 + 592 = 620
So 1.= 620 and o = 40

Answers
 e) P(X> 33) =0.1056 a) Let H=height incm. Then H ~ N(175, oe
= P(Z> 33 = BNE
)= 0.1056 P(H> 170) = 0.8

Don! )
—p(z>
= p(z> WO —2)=08
=> P(Z = a) = 0.8944
5 =0.8 > P(Z2 5)
= P(Z< 2) 5 = 0.2
&(z) = 0.8944 for z= 1.25
oe = 1.25 = 33 —p = 1.250 (equation 1) p= 0.2forz = 0.8416 = 2 = 0.8416
>0=5+
0.8416 = 5.94 cm (to 3sf.)
BOGS 21) = 019599
b) P(171 <H<179)
=P Ze ZI = =) = 0.9599 ens 179
—175
= P(Z <
p—21
)= 0.9599 =2( aac 52 ae)
a
= PE067 <=Z < 0.67)
(2) = 0.9599 for
z= 1.75 = P(Z < 0.67) — P(Z < -0.67)
= P(Z < 0.67) — [1 — P(Z< 0.67)]
e ae = 1.75 > p—21 = 1.750 (equation 2) = 2 x 0.7486 - 1
Adding equations 1 and 2 gives: = 0.4972
33 =21-—p+p=1.250+ 1.750 Let R = rainfall in cm. Then R ~ N(u, 07).
sS12=305>0=4
Putting o = 4 into equation 2 gives: P(R < 4) = 0.102 > P(Z< Le 20.102
w=1.75x4+21=28
So = 28 anda
=4 i — =0.102 > Piz < — = 0.898
f) P(X> 66) = 0.3632 ®(z) = 0.898 for
z= 1.27
=P(Z> nSa E) = 0.3632 > uae = 1.27 = p—4 = 1.276 (equation 1)
=> P(Z =
66 —
oe = 0.6368 P(R > 7) = 0.648 = P(z > 2—#) = 0.648
®(z) = 0.6368 forz = 0.35
66 —
= P(z<4#=*) = 0.648
= 0.35 = 66 — p= 0.350 (equation 1)
o ®(z) = 0.648 for z = 0.38
P(X< 48) = 0.3446
= eo = 0.38 > p-7 = 0.380 (equation 2)
48 — uw
=> P(Z z o )= 0.3446 Subtracting equation 2 from equation 1 gives:
p— 48 b= p—4 = (7) = 1.270 —0.386
=> P(Z > o )= 0.3446
= 3 =O0.0%e SoS 3.37 (to 3 Se)
u— 48
=> P(Z = ae = 0.6554 Putting o = 3.37 cm into equation 1 gives:
i= Ix B37 A =8.2 Brom (tolsisa)
®(z) = 0.6554 for z = 0.40 So 1 = 8.28 cm and o = 3.37 cm (to 3 sf.)
e aes = 0.4 = 1 - 48 = 0.40 (equation 2)
Adding equations 1 and 2 gives: Review Exercise — Chapter 5
66 — 48-11 + p = 0.350 + 0.40
Using the ®(z) table:
= 18=0./50 >0=24
Putting o = 24 into equation 2 gives: a) P(Z< 0.84) = 0.7995
p= 0.4 x 24+ 48 = 57.6 b) P(Z < 2.95) = 0.9984
So , = 57.6 and o = 24
c) P(Z> 0.68) = 1 —P(Z< 0.68) = 1 —0.7517
Q6 a) Let V= volume of vinegar in ml. = 0.2483
Then V ~ N(u, 52).
d) PZ2 155)] 1 = PZ <1 55)
DOUG
PWV= 506) =07/ 19 Pp(Za )=o.719 = 1-0.9394 = 0.0606
5
®(z) = 0.719 for z= 0.58 ©) RZ <2 0 Pie 2 10) = i Piano)
te DOGr = 1-—0.9821 = 0.0179
»
P=0.58 f) P(Z<-0.01) = P(Z> 0.01)
=.= 506—0:58 x 5 =503.1 ml = 1-P(Z< 0.01) = 1 — 0.5040 = 0.4960
b) P(V< 500) = P(Z< es g) PZS010)=1=PZ <010) =1 = 0.5398
= 0.4602
= P(Z <=0.62) = P(Z= 0:62) == PGs 0262)
= 1 -—0.7324 = 0.2676
h) P(Z< 0.64) = 0.7389
So, 26.76% of bottles contain less than 500 ml. i) P(Z>0.23)=1-P(Z< 0.23) = 1 — 0.5910
= 0.4090

Answers
 P(0.10 < Z< 0.50) = P(Z< 0.50) — P(Z< 0.10)
= 0.6915 — 0.5398 = 0.1517 d) P(47 <X<57)= p(2 20 <7 < ey
P(-0.62:s Z < 1,10).= P(Z< 1.10) — P(Z < -0.62) = P(-0.75
< Z< 1.75)
= P(Z< 1.10) — P(Z> 0.62) = P(Z< 1.75)- P(Z< 0.75)
= P(Z< 1.10) — (1 — P(Z< 0.62)) = P(Z< 1.75)- P(Z> 0.75)
= 0.8643 — (1 — 0.7324) = 0.5967 = P(Z<1.75)— [1 - P(Z< 0.75)
P{-0;99 <7 <'—0174) = P(0.74 < Z < 0.99) = 0.9599— (1 - 0.7734)= 0.7333
= P(Z< 0.99) — P(Z < 0.74) Q4 a) P(X <0) = P(z<2=2)
= 0.8389 — 0.7704 = 0.0685
= P(Z< -0.71) = P(Z> 0.71)
Q2 If P(Z < z) = 0.9131, then from the &(z) table, =1-P(Z< 0.71) =1-0.7611 = 0.2389
Zee Or
b) P(X <1) = P(Z< 42)
b) If P(Z < z) = 0.5871, then from the &(z) table,
O22. = P(Z< -0.57) = P(Z> 0.57)
C) If P(Z> z) = 0.0359, then P(Z < z) = 0.9641. =1-P(Z< 0.57) =1-0.7157 = 0.2843
From the ®(z) table, z = 1.80.
d) If P(Z > z)= 0.01, then from the percentage-points
C) P(X> 7) = P(Z > 2>2) = PZ> 0.29)
table, z= 2. 3263: = 1-P(Z< 0.29) =1-0.6141 = 0.3859
) If P(Z < z) = 0.4013, then z must be negative (and d) P(2<X <4) = P(As2<7< 42) i
so won't be in the ®(z) table). '
= P(-0.43 < Z<-0.14)
But this means P(Z= —z) = 0.4013, and so
=P(0.14<Z< 0.43)
P(Z< -z) = 1 — 0.4013 = 0.5987.
= P(Z< 0.43) — P(Z < 0.14)
Using the ®(z) table, —z = 0.25, soz =-0.25.
= 0.6664 — 0.5557
Ifyou need to draw a graph here to make it a bit
= 0.1107
clearer what's going on, then draw one.
If P(Z = z) = 0.995, then P(Z< -z) = 0.995 OS a) Here
X ~ N(80, 15).

= P(Z>-z) = 1-0.995 = 0.005 If P(X < a) = 0.99, thenP(Z 2 a — 80 == OL0)

Vv15 )
From the percentage-points table, -z = 2.5758. — 80
S07 =e. Tis
SaSa SOS (using percentage
t points)
ts).
When you've answered a question like this, always ask
yourself whether your answer looks ‘about right. Here,
Rearrange this to get a = 80 + 2.3263 x V15
you need a number that Z is very very likely to be = 89.01(to 4 s.f.)
greater than.. so your answer is going to be negative,
b) |X— 80| < b means that
Xis ‘within b’ of 80,
and it's going to be pretty big. So z = —2.5758 looks
Le. 80-b<X< 80+5.
about right.
Since 80 is the mean of X, and since a normal
8) If Pz < Z< z) = 0.5034, then the remaining area distribution is symmetrical,
= 1 — 0.5034 = 0.4966, and 0.4966 = 2 = 0.2483. P(0-b<X< .ifb) = 0.8 means that
So, P(Z < z) = 0.5034 + 0.2483 = 0.7517. P(X> 80 + b)=
This is the area between —z and z + the area to the left 80 + ”==100) | =H
PZ 2 = aS Ze oO. 1 O)eil
CfeZs
Using the ®(z) table, z = 0.68. Sow ele (using percentage points).
Or you could divide 0.5034 by 2 to get P(O< Z < z) V15
= 0.2517. Then add O.5 to get P(Z < z) = O./517. And so b = 4.964 (to 4 s.f.)
It's really important you remember that a normal
h) lia Opa aa N10) 93.97
distribution is symmetrical. You often need to use
Az) (ZO 5 Ons Oi
= (4 <2) = OSM 1A= O25)
symmetry to work out these sorts of questions.
=> P(Z < z) = 0.3917 + 0.5987 = 0.9904 Q6 If X represents the mass of an item in grams,
then X ~ N(55, 4.47).
Using the ®(z) table, z = 2.34.
Q3 a) P(X< 55) == P(Z< 25
=‘i50 ) a) PUY< 55) = P(Z< 2299) = PZ <0) =
= P(Z< 1.25) = 0.8944 b) PLX < 50) = P(Z< 2299) = Piz<-1.14)
b) P(X< 42) = P(z< 4520) = P(Z> 1.14) =1-P(Z< 1.14)
= 1-0.8729 =0.1271
= P(Z<-2) =P(Z>2)= i=RZe))
= 1-0.9772= 0.0228 C) PUY> 60) = P(Z> 29-92) = PZ> 1.14)
C) P(X> 56) = P(Z > 22528) =1-P(Z< 1.14) = 1 - 0.8729 = 0.1271
=PZe i.5)=1 —PiZ= 1.5)
= 1 — 0.9332 = 0.0668

Answers
Q7 If X represents the mass of an egg in kilograms, Exam-Style Questions — Chapter 5
then XY ~ N(1.4, 0.37).
1 Let Y represent the exam marks. Then XY~ N(50, 157).
Watch out for the units here — you need to make them the
same for both the mean and the variance. a) PIX <30)= Piz 409) 2 PiZehs3)
[1 mark] :
PUK <a) = 0.8830 = P(Z < a4) = 0.8830
= P(Z> 1.33) =1-—P(Z< 1.33) [7 mark]
&(z) = 0.8830forz= 1.19 = 1 — 0.9082 = 0.0918 [7 mark]

= wat 2119 el 19k OST 1957 b) P(x241)=P(Z= oe) = P(Z> -0.6)

[1 mark]
ome
Q8 PW< 8) = 0.8925 = P(ZZ )= 0.8925 = P(Z < 0.6) = 0.7257 [1 mark]
V10 So 0.7257 x 1000 = 726 is the expected number
B(z) = 0.8925 forz = 1.24 who passed the exam [1 mark].
8 = pt
= = 124> pS S— 1.24
x 10 c) If ais the mark needed for a distinction, then:
Vio 4
= 4.08 (to3 s.f.) PX al 30.1 = P(Z = 4— 2°) _ 0.1 [1 mark]
221 —p DON Morz— Zoe
Q9 P(X> 221) = 0.3085 =>P(Z S | = 0.3085
= 4—
29 = 1.2816 [1 mark]
SsP(Zae
< = 1=0.3085,=0,6915
=> a = 69.2 (to 3 s.f.) [1 mark]
®(z) = 0.6915 forz=0.5
a) Pi <50) = 05123
= SEL 05 > p= 221 —0.5 x 8=217
Q10 PIX< 13)
=0.6
=>P(Z eo 6ut
) = 0.123 [1 mark]

= p(z< Beth. P(Z< 4) => p(z> = )=0.123

o =0.6
a
=> P(Z=5)=04 2 Pp(Z eu aa 2461232 0877 11 mark
p = 0.4 forz = 0.2533 in) 0. B77 lore =a Ac
Vi) 0)
=> 2 = 0.2533 = o = 2 + 0.2533 =7.90 (to 3 s.f) 6 = 1.16 [7 mark]
41 = 116% 6.4.50.= 56.96 [1 markd
Q11 PY< 110) = 0.9678
110 — 108 = P(Z < a =aa 0.9678
=> P(Z <<3)
b) RAST)= p(z > A= 36.06)
= P(Z > 2.34) [1 mark]
®(z) = 0.9678 forz= 1.85 = 1—P(Z< 2.34) [1 mark]
=> 2 =| ,85.=5.6 =2 4 1.85.
Gant =1.08 (to 36h) = 1 — 0.9904 = 0.0096 [1 mark]
c) Since the distribution is symmetrical about 1,
P(X< pp. + a) = 0.9 [1 mark]
Q12 RUC <a) — Ol 765— P(Z <Bee = 0:97.83
= (z< Beat) _ p(Z<#) =09
(2) = 0.9783 for z = 2.02 [1 mark]
= PAH 2.02 = 15.2 p= 2.020 (equation 1) => P(Z =sS zi
@\ = 0.1
P(X> 14.8) = 0.1056 = P(z > 8—#) - 0.1056 p=0.1 for z= 1.2816
= P(Z=
14.8
— pw
)= 1 — 0.1056
= 0.8944 => 4 = 1.2816 [1 mark]
a 6
=> a=7.69 (to 3 s.f.) [1 mark]
®(z) = 0.8944 for z= 1.25
14.8 —
= ees = 1.25 > 14.8 —= 1.256 (equation 2)
Subtracting equation 2 from equation 1 gives:
15.2 — 14.8 — — (pt) = 2.020 — 1.250
= 04=0776 >6 = 0.51948... = 0.519 (to 3 sf.)
Putting o = 0.519... into equation 1 gives:
p= 15.2 —2.02 x 0.519... = 14.2 (to 3 sf.)
So 1 = 14.2 and o = 0.519 (to 3 s.f.)

Answers
3 a) Let X represent the lifetime of a battery in hours. For a normal distribution, mean = median,
a)
Then X ~ N(, 0°). so median = 12 inches [7 mark].
PX< 20) =0.4=5 P(Z < a E) = 0.4 [1 mark] b) Let X represent the base diameters.
20 Then X ~ N(12, o?). P(¥> 13) = 0.05 [1 mark],
= P(z>F— )=04 13 — 12 1
sO P(Z >aoe = P(Z > 5 = 0.05 [1 mark]
Using the percentage-points table,
p= OA Torz = 02533 Using the percentage-points table,
p = 0.05 for z = 1.6449
fu — 20
ar = 0.2533 [1 mark]
= . = 1.6449 [7 mark]
= 1 — 20 = 0.25336 (equation 1)
=>o0=1 + 1.6449= 0.608 (to 3 s.f.) [7 mark]
P(X< 26)=0.8 5 ieee )= 0.8 [1 mark]
26
= fi
C) P(X< 10.8) g= P(Z< (0.8aoe— 12 )
= P(Z= o P)=0.2 = P(Z< -1.97) [1 mark]
Using the percentage-points table, SHZS Ir)= 1=(4S< or
p= 02 1Oez = 0.8416 = 1 - 0.9756 = 0.0244 [7 mark]
So you would expect 0.0244 x 100 = 2 pizza
a F — 0.8416 [1 mark] bases to be discarded [1 mark].
=> 26 — 1 = 0.84160 (equation 2) d ~S P(at least 1 base too small)
Adding equations 1 and 2 gives: = 1 — P(no bases too small).
pb — jt — 20 + 26 = 0.25330 + 0.84160 [1 mark] P(base not too small)= 1 — 0.0244= 0.9756.
= 6 = 1.09496 => o = 5.47995... = 5.48 (to 3 s.f.) P(no bases too small)= 0.9756?= 0.9286
Putting o = 5.47... into equation 1 gives: [1 mark]
hi O25
3 Sao Are ate Ole? 1-4 (tO. 3S...) P(at least 1 base too small) = 1 — 0.9286 [1 mark]
So = 21.4 hours [7 mark] = 0.0714 [7 mark].
and o = 5.48 hours [7 mark] (to 3 s.f.) a) LetX represent the volume of compost in a bag.
Then X ~ N(50, 0.47).
b) From part a), X ~ N(21.4, 5.48?)
P(X< 49) = P(Z< a9
=
a 50 )
So P(X= 15) ba 15—21.4 )[1 mark]
= P(Z 23a
= P(Z < —2.5) [1 mark]
= P(Z==1.17) = PZ< 1.17) [1 mark] = P(Z> 2.5) = 1—-—P(Z< 2.5) [1 mark]
= 0.8790 [1 mark] = 1 — 0.9938 = 0.0062 [7 mark]

a) P(X> 145) = P(Z > Hehe b) P(X> 50.5) = P(Z > 282
= 20)
= P(Z> 1) [1 mark] = P(Z> 1.25) [1 mark]
= 1—P(Z< 1) [1 mark] = 1-—P(Z< 1.25) [1 mark]
= 1 — 0.8413 = 0.1587 [1 mark] = 1-—0.8944 = 0.1056 [7 mark]
So in 1000 bags, 0.1056 x 1000 [7 mark]
b) P(120 < X <j) = 0.4641 ~ 106 bags [7 mark] (approximately) would be
= p(20120 e7e ae = 0.4641 expected to contain more than 50.5 litres of
compost.
= p(o<z<L7
i200 )= 0.4641 ff mark] C) P(Y< 74) =0.1
= P(Z< See
Stee P(Z < 0) = 0.4641 = P(z< A=) = p(z<—3) =0.1
Se 3) = 0.1 [1 mark]
= P(Z< i).
use 0.4641 +0.5 Using the percentage-points table,
e ae
asi) - 0.9641 [1 mark] p =0.1 forz= 1.2816
= a = 1.2816 [1 mark]
a = 0.9641 for
z = 1.80
= 120 =o =i = lao
= 0.780 litres (to 3 s.f.) [7 mark]
= j=1.8x
25 +120 =165 [1 mark]

Answers
Addition law Dependent variable Histogram
A formula linking the probability of Another name for the response A diagram showing the frequencies
the union and the probability of the variable. with which a continuous variable falls
intersection of events A and B. in particular classes — the frequency
Discrete random variable of a class is proportional to the area
A random variable with ‘gaps’ of a bar.
between its possible values.
Bivariate data Discrete uniform distribution
Data that comes as an ordered pair of A discrete random variable has a
variables (x, y). discrete uniform distribution if all the Independent events
possible values it can take are equally If the probability of an event B
Boxplot happening doesn’t depend on
A diagram showing the median, likely.
whether or not an event A happens,
quartiles and greatest/least values of a Dispersion events A and B are independent.
data set, as well as any outliers. Measures of dispersion describe how
spread out data values are. Independent variable
Another name for the explanatory
variable.
Coding
Coding means transforming all the Interpolation
Event
readings in a data set to make the Predicting a value ofy corresponding
An event is a ‘group’ of one or more
numbers easier to work with. to a value of x within the range for
possible outcomes.
which you have data.
Complement (of an event A) Expected value
The group of all outcomes interquartile range
The expected value of a random
corresponding to event A not A measure of dispersion. It’s the
variable is the ‘expected’ mean of a
happening. difference between the upper quartile
large number of readings. It’s the sum
and the lower quartile.
Conditional probability of all possible values multiplied by
A probability is conditional if it their probability. Intersection (of events A and B)
depends on whether or not another The set of outcomes corresponding
Explanatory variable
event happens. to both event A and event B
In an experiment, the variable you
happening.
Continuous random variable can control, or the one that you think
A random variable which is measured is affecting the other.
on a continuous scale. It may take Extrapolation
any value in a given range (i.e. with Predicting a value ofy corresponding Linear regression
no ‘gaps’ between possible values). to a value of x outside the range for A method for finding the equation of
Correlation which you have data. a line of best fit on a scatter diagram.
A linear relationship between two Location
variables showing that they change Measures of location show where the
together to some extent. ‘centre’ of the data lies.
Fence
(A correlation does not necessarily
mean a causal relationship.)
If a data value lies outside a fence, Lower quartile
then it is an outlier. The value that 25% of data values in a
Cumulative distribution function data set are less than or equal to.
A function, F(x), which gives the
Frequency density
probability that a random variable, The frequency of a class divided by its
X, will be less than or equal to a class width.
particular value, x. Mean
A measure of location — it’s the sum
of a set of data values, divided by the
number of data values. ~

Glossary
Median Product moment correlation Standard normal variable, Z
A measure of location — it’s the value coefficient f# A random variable that follows a
in the middle of the data set when all A measure of the strength of the normal distribution with mean O and
the data values are in order of size. correlation between two variables. variance 1.
Mode
Stem and leaf diagram
A measure of location — it’s the most
A way of displaying data in which the
frequently occurring data value. data values are split into a ‘stem’ and
Qualitative variable
Mode ofa discrete random variable A variable that takes non-numerical a ‘leaf’. Data values with the same
The value that the random variable is values. stem are shown in the same row of
most likely to take — i.e. the one with the diagram.
Quantitative variable
the highest probability.
A variable that takes numerical
Mutually exclusive values.
Events are mutually exclusive (or just
Quartiles Tree diagram
‘exclusive’) if they have no outcomes The three quartiles Q,, Q, and Q, Tree diagrams show probabilities for
in common, and so can’t happen at divide an ordered data set into four
sequences of two or more events.
the same time. parts. Trial
A process (e.g. an experiment) with
different possible outcomes.
Normal distribution Random variable
A continuous ‘bell-shaped’ probability A variable taking different values with
distribution where the further from the different probabilities. Union (of events A and B)
mean a value is, the less likely it is to The set of outcomes corresponding
occur. Range
to either event A or event B (or both)
A measure of dispersion. It’s the
happening.
difference between the highest value
and the lowest value. Upper quartile
Outcome The value that 75% of data values in a
Regression line
One of the possible results of a trial data set are less than or equal to.
A line of best fit found using linear
or experiment.
regression.
Outlier
Residual
A freak piece of data lying a long way Variable
The difference between areal-life
from the majority of the values in a A quantity that can take a variety of
observation and what a regression
data set. values — its value is not fixed.
line predicts.
Response variable Variance
In an experiment, the variable you A measure of dispersion from the
Percentiles think is being affected. mean.
The percentiles (P,-P,,) divide an
Variance of a discrete random variable
ordered data set into 100 parts.
The ‘expected’ variance of a large
Probability distribution for a number of readings.
Sample space
discrete random variable
The set of all possible outcomes of a Venn diagram
A table showing all the possible A Venn diagram shows how a
trial.
values that the random variable can
collection of objects is split up into
take, plus the probability that it takes Scatter diagram different groups, where everything in
each value. Graph showing the two variables in a
a group has something in common.
bivariate data set plotted against each In probability, the objects are
Probability function
other. outcomes, and the groups are events.
A function that generates the
probabilities of a discrete random Skewness
variable taking each of its possible Measures of skewness tell you how
values. symmetrically distributed a data set is.
Z-tables
Product law Standard deviation Tables relating to the standard normal
A formula used to work out the A measure of dispersion calculated by variable (Z) — such as the cumulative
probability of two events both taking the square root of the variance. distribution function ®(z), and the
happening. percentage-points table.

Glossary
 A D
addition law 61-64 Gata™ 3 independent events 68-70
assumptions 1, 2 dependent variables 114 testing for independence 69
discrete random variables 80-82 independent variables 114
discrete uniform distribution interpercentile range 25, 26
B 938799
discrete variables 3
interpolation 20, 21, 24-26,
118, Aig
back-to-back stem and leaf ‘interquartile range 23-25, 35
dispersion 23-26, 31-35,
diagrams 12 intersection (of two events)
45-47, 93
bivariate data 104, 106
grouped data 24, 25, 33 56-50 Olle oo 68
box plots 28-30

C E L
events 53 linear interpolation 20, 21,
causation 113
exclusive events 63, 64 24-26
class boundaries 4-7
expected value 90-92 linear regression 114-117
class mid-points 5
of a discrete random with coded data 121, 122
class width 5
variable 90 lines of best fit 104, 105, 114
coding data
of a function of a discrete location 14-17, 19-22,
finding mean and standard
random variable 96, 97 45-47
deviation 37-39
of the discrete uniform comparing measures 22
linear regression 121, 122
distribution 99 grouped data 19-21
product moment correlation
experiments 53 lower class boundary 5-7
coefficient 111, 112
comparing distributions 45-47 explanatory variables 114 lower fence 28-30
complement (of an event) 56-58
extrapolation 118, 119 lower quartile 23-25, 43, 44
conditional probability 65-68,
72-75
continuous variables 3
-
M
converting variables to Z 135 fences 28-30 mathematical models 1, 2
correlation 104-106 frequency density 7-9 mean 14, 15, 22,31-34, 41
limitations 113 frequency tables 4, 5 of a discrete random
cumulative distribution function variable 90
discrete random variables of anormal distribution 126,
86, 87
normal distribution 128, 150
G 143-145
of grouped data 19, 38
grouped data
cumulative frequency 17, 20, of the discrete uniform
mean 19, 38
21, 24-26
median 20, 21 distribution 99
modal class 19 using coding 37-39
quartiles 24, 25 medial 16," 17, 20-22) 41
standard deviation 33, 38 mid-points of classes 5
variance 33, 38 modal class 19
grouped frequency tables 4, 5 mode 16, 17, 22, 41
modelling process 2
multiplicationlaw 65-68
H mutually exclusive events 63, 64
histograms 7-9, 126

Index
N T
negative correlation qualitative variables 3 transforming variables to Z 135
105, 106, 108 quantitative variables 3 tree diagrams 71-75
negative skew 41-44 quartile coefficient of skewness for dependent events 72-75
normal distributions 126, 127 42-44 for independent events 71, 72
cumulative distribution function quartiles 23-25, 42-44 trials 53
12350
mean 126, 143-145
shape 126, 127
standard deviation
R U
random variables union (of two events)
127, 143-145
continuous 126 56) 50,50 1) eGs
standard normal distribution
discrete 80 upper class boundary 5-7
128-134
ranger 23735 upper fence 28-30
table of percentage points
regression lines 115-117 upper quartile 23-25, 43, 44
(eo
2 et
interpreting 117
using Z-tables 128-134
residuals 115
variance 126, 127
response variables 114
risk 1
V
variables 3, 80, 104, 105, 114,
O 126
variance 31-35
observations 3
outcomes 53, 54, 56
S of a discrete random variable
sample space 54, 56 93-95
outliers 28-30
sample-space diagrams of a function of a discrete
54, 83, 84 random variable 96, 97
scatter diagrams 104-106, 108 of grouped data 33, 38
P set notation 56 of the discrete uniform
Pearson’s coefficient of skewness simplifications 2 distribution 99
42, 44 skewness 41-47 Venn diagrams 56-60
percentage-points table 132, 151 Pearson's coefficient of 42, 44
percentiles 25, 26 quartile coefficient of 42-44
PMCC 108-113
positive correlation
standard deviation 32-35
of a normal distribution
Z
Z (standard normal variable) 128
TOS 1067108 127, 143-145
Z-tables 150, 151
positive skew 41-44 of grouped data 33, 38 using them 128-134
predictions 2 using coding 37-39
probability 53-75 standard normal distribution
formula for equally likely — 128-134
outcomes 53 statistical tables 150-151
probability distributions using them 128-134
80-84, 86-88 stem and leaf diagrams
discrete uniform distribution Tagen 7
98799 back-to-back 12
probability functions S. 106-108
31,782) 87, Go S. 106-108
product law 65-68 S,, 106-108
product moment correlation
coefficient 108-113
limitations 113
with coded data 111, 112

Index
 S1 Formula Sheet
These are the formulas you'll be given in the exam, but make sure you know exactly
when you need them and how to use them.

Probability Discrete Distributions

P(A U B) = P(A) + P(B) — P(A B) For a discrete random variable X taking
values x, with probabilities P(X= x,):
P(A' B) = PIA x P(BIA)
Expectation (mean): E(X) = pf = 2x,P(X = x,)

P(A|B) = P(BJA)P(A) Variance: Var(X) = 0? = L(x, — b)?P(X = x,)

P(B|A)P(A) + P(B|A’)P(A’)
= 2x7PX Sa ae

For a function g(X): E(g(X)) = Xg(x,)P(X = x;,)

Continuous Distributions
Standard continuous distribution:

Correlation and Regression

For a set of n pairs of values (x,, y,):

s = S) = 2X(x, — ¥? = 2x? - a SP = L(y, - y? = Zy? - eyt

S,, = Bx, - ¥)y, - 7) = Exy, - EVE

The product moment correlation coefficient is:
% | Seo) a Eee ee
Sa8y
i

{(% — XH (x! = CTV oy2BeeBar)

The regression coefficient of yonxisb = On esLH =x)=)

Sis d(x, — xy
Least squares regression line ofy on x is y= a+ bx wherea= y — bx
| MES1T
Pry $1 Formula Sheet
Exam Practice Papers and
Worked Answers

Ho Exam Practice Papers and Worked Answers CD-ROM above?

li Contact CGP customer services on 0870 750 1262
and ask for an A-Level Mathematics Statistics 1 for Edexcel CD-ROM —
I % we’ll send you one right away...

www.cgpbooks.co.uk

0412 - 9870
 A-Level Maths for Edexcel S1
This student book is a comprehensive guide to success in the Statistics 1 Module
of A-Level Maths, matched exactly to the Edexcel course requirements.
Features include:
os Straightforward EMORGR ene supported by clear examples.
= Hundreds of helpful tips throughout the book. |
* Practice questions and exam-style questions for each chapter.
* Detailed, step-by-step answers to every question.
* Iwo printable practice exam papers on the CD-ROM.

And since it’s a CGP book, rest assured it covers Shed you need
as painlessly as possible.

£13.99 www.cgpbooks.co:uk .
(Retail Price) ; . , ee

P es bs
De aw &
a = = nak. . aa. 5 " ‘ “
4 sr a i . eaOe
are