3) M A2 The Normal Distribution

Prior knowledge check
1) The probability that a one month 3) The probability that a plate made
old Labrador puppy weighs under using a particular production process is
2kg is 0.735. Two puppies are faulty is given as 0.16. A sample of 20
plates is taken. Find:
chosen at random from
different litters. Find:
a) !(#$%ℎ '()*+ 2-.) a) The probability that exactly two
plates are faulty
b) !(*012%34 $(* '()*+ 2-.) b) The probability that no more than
three plates are faulty
2) 5~7(20,0.4). Find:
a) ! 5 = 6
b) ! 5 ≥ 8
c) !(3 ≤ 5 ≤ 10)
The normal distribution
Let 5 = ('C#*+ $D ℎ*1)E Fℎ*( 8 2$G(E 1+* %ℎ+$F(
You need to understand the shape
and basic ideas behind the normal !(# = %)
distribution
In year 12 you will have

encountered discrete random
variables 0 1 2 3 4 5 6 7 8
So although we haven not included a y-axis, the idea is that

à The key numerical point was that ! can only take certain values, and each will be assigned a
these can only take certain probability (the height of its bar)
values, determined by the
context !(# = %)
0 1 2 3 4 5 6 7 8
3A
The distribution
You need to understand the shape and of heights as a
basic ideas behind the normal Histogram
distribution
The normal distribution focuses on

continuous random variables
130 140 150 160 170 180
These can take any numerical value (eg If we reduce the classwidths used…
Height), and are modelled using a curve
instead
à For a quantity such as height, the

majority of people will be close to the
average
à As we move further away from the 130 140 150 160 170 180
average height, there are fewer
people (ie – there would be a lower The shape of the Histogram becomes smoother. If
probability of randomly selecting one we continue indefinitely, we get what is known as
at that height) the normal distribution curve…
3A
175cm
You need to understand the shape and
basic ideas behind the normal
distribution
For example, if we decided to model the

heights of adult males in a city…
à Imagine that the mean height was

175cm, and the standard deviation
was 12cm…
à The curve will be split into equal

halves either side of 175cm '
140 150 160 170 180 190 200 210
à The area under the curve is used to

represent probabilities, with the total
area underneath being equal to 1
3A
distribution
à The normal distribution has 2

parameters, the mean (H) and variance
(I !), so the standard deviation is I
à It is symmetrical, so
Mean = Median = Mode
à It has asymptotes at each end

H−I H H+I
à The total area beneath it is equal to 1
à It has points of inflexion at H + I and

H−I
3A
distribution
à Approximately 68% of the data is So what would

within one standard deviation of the this area be?
mean
à Approximately 95% of the data is

within 2 standard deviations of the
mean
163$% 175$% 187$% '

à Almost all (99.7%) of the data is For example, using the data from before, the height
within 3 standard deviations of the of adult males is being investigated, and we find the
mean mean is 175cm and the standard deviation is 12cm
à This means the area on the diagram above would

be 0.68 (ie – the probability of a randomly chosen
male falling into this region would be 68%)
3A
distribution
The diameters of a rivet produced by a

particular machine, 5 mm, is modelled as
5~L(8, 0.2!). Find: Area = Area =
0.5 0.5
8%% %
a) ! 5 > 8
b) !(7.6 < 5 < 8.4)
c) !(5 > 7.8)
à You should always start with a quick

sketch of the normal distribution,
with the mean indicated on it…
3A
distribution

5~L(8, 0.2!). Find:
7.6%% 8%% 8.4%% %

a) ! 5 > 8 = 0.5
Mark on the limits we are given in part b, and
b) !(7.6 < 5 < 8.4)
compare it to the standard deviation
c) !(5 > 7.8)

3A
distribution

5~L(8, 0.2!). Find:
7.8%% 8%% %
a) ! 5 > 8 = 0.5
Mark on the value we are given in part c)
b) !(7.6 < 5 < 8.4) = 0.95
c) !(5 > 7.8)

3A
distribution
Note that the formula used to generate

the curve to the right is…
1 "!# !
!
Φ " = ( $% !
2&'
This means the area under

the graph, up to point 0
3A
à Now, on your calculator, press
menu
( = 30
You need to be able to use the +=4
normal cumulative distribution
function on your calculator to find
probabilities à Now press 7 for distributions
Given that 5~L(30, 4!), find: 30 33
a) ! 5 < 33 à Now press 2 for the normal cumulative

distribution
b) !(5 ≥ 24)
c) !(33.5 < 5 < 38.2) à Now input the lower limit. If there is
no lower limit, choose a value at least 5
d) !(5 < 27 $+ 5 > 32) standard deviations below the mean (so
in this case any value less than or equal
to 10 would be fine)
Even though we will use your

calculator, we should still sketch à Now input the upper limit (in this
the situation… case 33), the standard deviation and
the mean
à Press equals and the calculator will tell you the area!
3B
( = 30
probabilities
Given that 5~L(30, 4!), find: 24 30
à Input the lower limit (in this case

a) ! 5 < 33 24 – it does not matter whether the
inequality also includes 24 or not)
b) !(5 ≥ 24)
c) !(33.5 < 5 < 38.2)
d) !(5 < 27 $+ 5 > 32) à Now input the upper limit (in this
case there is none, so any value
above 50), the standard deviation
and the mean
calculator, we should still sketch
the situation…
à Press equals and the calculator will tell you the area
3B
( = 30
probabilities
Given that 5~L(30, 4!), find: 30 33.5 38.2
a) ! 5 < 33 à Now input both limits,

the standard deviation and
b) !(5 ≥ 24) the mean (it does not
matter whether the
c) !(33.5 < 5 < 38.2) inequality includes the
value or not)
d) !(5 < 27 $+ 5 > 32)
à Press equals and the calculator will tell you the area
the situation…
3B
( = 30
normal cumulative distribution In this case, we can find the
function on your calculator to find area between 27 and 32, and
probabilities then subtract it from 1…
Given that 5~L(30, 4!), find: 27 30 32
a) ! 5 < 33
à Now input the limits,
b) !(5 ≥ 24)
mean and standard
c) !(33.5 < 5 < 38.2) deviation
d) !(5 < 27 $+ 5 > 32)

the situation…
3B
You need to be able to use the Use your calculator as before. The lower limit should be
normal cumulative distribution , the upper limit should be at least (5 times the
function on your calculator to find standard deviation above the mean).
probabilities
The mean is and the standard deviation is …
An IQ test is applied to a population of

adults. The scores, X, on the test are
found to be normally distributed with
!~#(100, 15! ). Adults scoring more than
140 on the test are classified as ‘genius’.
a) Find the probability that an adult

chosen at random achieves a ‘genius’
classification
b) Twenty adults take the test. Find

the probability that two or more
are classified as ‘genius’
3B
For this part you will need to use what you learned in year 12
You need to be able to use the (the binomial distribution)
function on your calculator to find There are 20 adults in the sample, and the probability of a
probabilities genius is 0.00383
An IQ test is applied to a population of

adults. The scores, X, on the test are
found to be normally distributed with
!~#(100, 15! ). Adults scoring more than
140 on the test are classified as ‘genius’.
a) Find the probability that an adult

chosen at random achieves a ‘genius’
classification
= 0.00383
b) Twenty adults take the test. Find

the probability that two or more
are classified as ‘genius’
3B
à Now, on your calculator, press ( = 20
menu
You also need to be able to use the +=3
inverse function on you calculator,
to find values that will give specific
Area = 0.75
probabilities
à Now press 7 for distributions
Given that !~# 20, 3! , find, to two

20 -
decimal places, the values of , such
that:
à Now press 3 for the Inverse
normal distribution
a) - ! < , = 0.75
b) - ! > , = 0.4
c) - 16 < ! < , = 0.3 à Now input the probability we
want, as well as the standard
deviation and mean
Draw a sketch to get an idea of where

the value should be à Press equals and the calculator
à You can use the area to estimate the will tell you the value which has this
value we are looking for… area below it (as in, below the value,
to the left of it)
3C
( = 20
probabilities
Area =
0.4

20 -
decimal places, the values of , such
that:
à Be careful here! The calculator

a) - ! < , = 0.75 bases calculations on the area below
the value given. So in this case, we
b) - ! > , = 0.4
need to put in the probability as 0.6.
c) - 16 < ! < , = 0.3
à After putting the mean and standard deviation in as well, the

Draw a sketch to get an idea of where calculator will tell you the value…
the value should be
à If the area above a value is 0.4, that
value must be above the mean
3C
( = 20
to find values that will give specific Area =
probabilities 0.3
Area =
0.0912
decimal places, the values of , such 16 - 20
that:
à The total area ‘below’ value " will
be equal to 0.3 plus the area below
a) - ! < , = 0.75 16. We can find the area below 16
b) - ! > , = 0.4 using the function on your calculator
we used in section 3B
c) - 16 < ! < , = 0.3
à It is 0.0912…
Draw a sketch to get an idea of where

the value should be à So put in an area of 0.3912, as well
as the mean and standard deviation, in
order to find the value "
3C
( = 20
You also need to be able to use the + = 1.5
Area = 0.6
probabilities
Plates made using a particular

20 .
manufacturing process have a diameter,
5 67, which can be modelled using a
normal distribution, 5~#(20, 1.5! ). Use your calculator as before to find the value 8.
à The mean is 20, the standard deviation is 1.5, and

a) Given that 60% of plates are less
the area/probability is 0.6
than 8 67, find the value of 8.
b) Find the interquartile range of the

plate diameters
3C
( = 20
You also need to be able to use the Area = + = 1.5
inverse function on you calculator, 0.25
Area =
to find values that will give specific 0.25
probabilities
Area = Area =
0.25 0.25
10 20 /0
normal distribution, 5~#(20, 1.5! ). à Firstly, remember that for normally distributed
data, Mean = Median = Mode, so the value of 20 is
also the Median
à Secondly, remember that the quartiles and the
= 20.3867 median split the data into 4 equal parts, with the
same quantity of data in each
plate diameters
à Therefore, the area below the lower quartile (and
all the other areas), will be 0.25
3C
( = 20
probabilities
Area =
0.25
normal distribution, 5~#(20, 1.5! ).

= 20.3867
plate diameters
3C
( = 20
probabilities
Area =
0.25
manufacturing process have a diameter, 18.988 20 21.011
normal distribution, 5~#(20, 1.5! ). à So the interquartile range will be the difference
between the two values we calculated
(note that we could also find the difference between
the Median and the LQ, and just double that!)
= 20.3867
plate diameters
3C
( = 25
You need to be able to use the
standard normal distribution. If
+=3
you need to work out an unknown Let %~2(25, 3! )
mean or standard deviation, you
will need to use this…
A set of normally distributed data can The coding will change all 25 31
be coded so it fits the standard normal values in the same way, such
distribution that the mean becomes 0 and
the standard deviation
becomes 1
(=0
à The standard normal distribution has
a mean of 0 and a standard deviation of +=1
1
Then Z~2(0, 1! )
All values maintain their relative positions
à So a value of 31 in the original

distribution will become a value of 2 in
the standard distribution (2 standard 0 2
deviations above the mean)
Note that the 2 diagrams are NOT TO SCALE!
3D
%−(
5=
+

you need to work out an unknown
The original set
of values
The coding used is shown, and would be
applied to all values in the original data The mean
set…
*−,
)=
'
The The standard

standardised deviation
set of values
3D
%−(
5=
+

Shoe size Frequency Coded data Frequency
(!) (:)
standard normal distribution. If 6 1 -2.21 1
you need to work out an unknown 7 3 -1.47 3
will need to use this… 8 6 -0.73 6
9 10 0 10
10 6 0.73 6
For example, imagine we had this table
of data on shoe sizes (!) 11 3 1.47 3
12 1 2.21 1
We could calculate the mean and We could calculate the mean and standard
standard deviation for this set of data deviation for this new set of data
%−(
5=
+
à Now, we code the data using the

formula we saw, to create a new data Notice that the numbers in the new data
set, Z (the frequencies stay the same) set are also symmetrical about 0…
3D
%−(
5=
+

You could calculate -(! < 53) using what you have learnt so
You need to be able to use the far this chapter, along with your calculator
à This question is asking you to write an equivalent
probability, based on the standardised normal
distribution
à The letter Φ is the Greek capital phi
The random variable !~#(50, 4! ). Write
in terms of Φ(<) for some value <:
a) -(! < 53)
3D
%−(
5=
+

You need to be able to use the ( = 50
standard normal distribution. If +=4
mean or standard deviation, you Original

in terms of Φ(<) for some value <: 50 53
a) -(! < 53)

(=0
= Φ(0.75)
+=1
If we compare the diagram for the
original distribution, and the new one, Standardised
the probability of getting below 53 on
the first is the same as the probability
of getting below 0.75 on the second
à You can also think of it is ’53 is 0.75 0 0.75
standard deviations above the mean’
3D
%−(
5=
+


in terms of Φ(<) for some value <:
b) -(! ≥ 55)
3D
%−(
5=
+

You might be asked to find a value for z

which corresponds to a particular
probability. You are given some standard
values in the formula booklet… This is saying that the value with an
probability of 0.3 above it is 0.5244
à Be careful here, these values (=0

correspond to the region above a
value in the standard normal +=1
distribution
Standardised
Area = 0.3
3D
%−(
5=
+

>~#(127, 16! )
A = 127
standard normal distribution. If à The 95th percentile B = 16
you need to work out an unknown has 5% of the data above
mean or standard deviation, you it, so there is a value
will need to use this… which has this property… Area =
0.05
127 ?
The systolic blood pressure (pressure
when the heart beats) of an adult
population, > 77?@, is modelled as a
normal distribution with mean 127 and
:~#(0, 1! )
standard deviation 16.
A=0
à We will need to start
A medical researcher wants to study by finding the equivalent
B=1
adults with blood pressures higher than
the 95th percentile. Find the minimum value on the standardised
blood pressure needed for an adult to distribution
Area =
be included in her survey 0.05
0 ?
à Start by drawing a sketch for the
original distribution, and the
standardised one…
3D
%−(
5=
+

>~#(127, 16! )
A = 127
standard normal distribution. If B = 16
will need to use this… Area =
0.05
127 153
?
à Use the table from before to find
the value which has C = 0.05
:~#(0, 1! )
A=0
B=1
Area =
0.05
0 1.6449
?
à Now we need to use the formula in
reverse. We have the value of < and
want the corresponding value of 8
3D
%−(
5=
+

!~#(A, 3! )
We now know that the point 20
A =?
You need to be able to find on the original distribution is
0.8416 standard deviations above
unknown values of S, T, or both B=3
the mean…
à Use the formula above to find

The random variable !~#(A, 3! ). Given what the mean is… Area =
that - ! > 20 = 0.2, find the value of A. 0.2
? 20
à As in the previous section, start by
drawing 2 sketches, one of the
distribution for !, and another for :~#(0, 1! )
the standard distribution
A=0
B=1
à Use the table to find the
corresponding value of <…
Area =
0.2
0 ?
3E
%−(
5=
+

!~#(50, ?! )
We now know that the point 46
A = 50
You need to be able to find on the original distribution is
0.7998 standard deviations below
unknown values of S, T, or both B =?
the mean…
à Use the formula above to find

A machine makes metal sheets with what the standard deviation is…
width ! 67, modelled as a normal
distribution such that !~#(50, B ! ).
50
a) Given that - ! < 46 = 0.2119, find

the value of B. :~#(0, 1! )
A=0
à Draw the sketches… B=1
à You can use the inverse function on

your calculator (see section 3C) to
find the missing value on the
standard distribution ? 0
à The probability is 0.2119, the mean is

0 and the standard deviation is 1
3E
%−(
5=
+

You need to be able to find
unknown values of S, T, or both
A machine makes metal sheets with

width ! 67, modelled as a normal
distribution such that !~#(50, B ! ).
a) Given that - ! < 46 = 0.2119, find

the value of B.
b) Find the 90th percentile of the

widths
à You can do this using the inverse

function on your calculator
à The 90th percentile is the value with

90% of the data below it
3E
%−(
5=
+

!~#(A, B ! )
Lower tail A =?
unknown values of S, T, or both B =?
The random variable !~#(A, B ! ). Given

that - ! > 35 = 0.025 and - ! < 15 =
0.1469, find the values of A and B.
4
à Start by drawing the two sketches…

:~#(0, 1! )
à Find the two values for < in the Upper tail A=0
standard distribution diagram. You B=1
can use the table of values for the
upper tail, and the inverse function
on your calculator for the lower tail… Area = Area =
0.1469 0.025
à Then use the formula above with ? 0 ?

each of these…
3E
%−(
5=
+

unknown values of S, T, or both
The random variable !~#(A, B ! ). Given

that - ! > 35 = 0.025 and - ! < 15 =
0.1469, find the values of A and B.
à Now we have 2 equations, we can

solve them simultaneously
3E
%−(
5=
+

You need to be able to approximate a
binomial distribution
Let’s compare the two quickly…
Binomial distribution Normal distribution

à Discrete data à Continuous data
No. heads when flipping a Height of the flipped coin

coin 7 times
You can use the processes you have learnt for the normal distribution to approximate a binomial
distribution under 2 conditions
à The value of F (the number of trials) must be ‘large’ (this will mean the distribution is smoother)
à The probability of success must be close to 0.5 (if it is not, then the distribution will not be
symmetrical)
3F
%−( Using normal distribution to à F must be large
5= approximate a binomial distribution à C must be close to 0.5
+
The normal distribution ( = 78
You need to be able to approximate a Let !~G(100,0.45)

Starting with a Binomial distribution,

5~7((, U)
à We are going to use a normal

distribution to approximate it (for
example, the binomial distribution
tables you are given only go up to
( = 50
à So we are going to use a normal

distribution 5~L(H, I)
à We need to be able to get from (

and U to the values of H and I
3F
%~9(7, 8)
%~2((, +)
8 H H8
binomial distribution 1 10 10
2 10 20
3 10 30
The expected value of a set of data is
essentially its mean. 4 10 40
5 10 50
Imagine we had a normal unbiased 6 6 10 60

sided dice. If we rolled it 60 times we
# = 60 ' () = 210
would expect to get 10 of each
number.
Let 0 be the number we get when

rolling the dice, D be the frequency of
that number, and ( be the total
number of trials
3F
%~9(7, 8)
%~2((, +)
8 C(8) 8×C(8)
binomial distribution 1 1J
6
1J
6
2 1J 2J
6 6
The expected value of a set of data is 3 1J 3J
essentially its mean. 6 6
4 1J 4J
6 6
5 1J 5J
However, rather than doing a trial and 6 6
using that to calculate the mean (ie its 6 1J 6J
6 6
expected value), you can calculate it
using the probabilities of each event
happening…
à Doing this calculates the expected

value straight away, without needing
to divide by 0
3F
%~9(7, 8)
%~2((, +)

What you have just seen is showing

that the following two calculations are
equivalent…
∑"
= ∑ "×0(")
.
Multiply each outcome by its
Add all the outcomes up, and divide by probability, and then add them all
how many there are in total up after
(note that in the example we used H8, but So really it is ∑(8×C(8))
the idea is the same – add up all the
values! The fx just sped things up a bit)
3F
∑0 %~9(7, 8)
= ∑ 0×U(0)
( %~2((, +)
8 8! H H8 !
binomial distribution 1 1 10 10
2 4 10 40
3 9 10 90
It does not matter what set of values
we start with. 4 16 10 160
5 25 10 250
à For example, if the starting values 6 36 10 360

were all 0 ! instead of 0…
3F
∑0 %~9(7, 8)
= ∑ 0×U(0)
( %~2((, +)
8 8! C(8) 8 ! ×C(8)
binomial distribution 1 1 1J
6
1J
6
2 4 1J 4J
6 6
It does not matter what set of values 3 9 1J 9J
we start with. 6 6
4 16 1J 16J
6 6
5 25 1J 25J
à Like the previous example, we can 6 6
find the expected value of 0 ! by 6 36 1J 36J
6 6
using the original probabilities…
3F
∑0 ∑ 0! %~9(7, 8)
= ∑ 0×U(0) = ∑ 0 !×U(0)
( ( %~2((, +)

What you have just seen is showing

that the following two calculations are
equivalent…
∑ "$
= ∑ " $ ×0(")
.
Multiply each squared outcome by
Add all the squared outcomes up, and its probability, and then add them
divide by how many there are in total all up after
So really it is ∑(8 ! ×C(8))
3F
∑0 ∑ 0! %~9(7, 8)
= ∑ 0×U(0) = ∑ 0 !×U(0)
( ( %~2((, +)
!
You need to be able to approximate a !
∑ 0! ∑0
I = −
binomial distribution ( (
Can you remember the formula for the

variance of a data set from year 12?
à So we need to focus on rewriting

this part…
X(0 !×U(0))
3F
%~9(7, 8)
%~2((, +)
You need to be able to approximate a Imagine we are considering the number of heads when
binomial distribution tossing a biased coin 5 times. Let 8 = FL7MNO PH ℎN,RS,
and - ?N,R = 0.4
Can you remember the formula for the à Lets put together a table…
# 8(#) # ! ×8(#)
5 5
0 0
0.4" 0.6 #
0! ×
0
0.4" 0.6 #
à So we need to focus on rewriting
this part… 1
5
0.4$ 0.6 %
1! ×
5
0.4$ 0.6 %
1 1
5 5
X(0 !×U(0)) 2 2
0.4! 0.6 &
2! ×
2
0.4! 0.6 &
- 5 5
Remember ( how 3 0.4& 0.6 !
3! × 0.4& 0.6 !
U / 1to
− Ufind
-./the 3 3
=àX 0 !×
calculation for
0 a an event that is 5 5
/01 4 0.4% 0.6 $
4! × 0.4% 0.6 $
binomially distributed? 4 4
5 5
5 0.4# 0.6 "
5! × 0.4# 0.6 "
( , -., 5 5
! 5=+ = U 1−U
+ F '
8!× C 1−C ()'
8
So we are doing this calculation for each value of 8, and then adding them up…
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)
You need to be able to approximate a You hopefully also remember the following from
binomial distribution year 12 (Binomial expansion powerpoint)
7 7!
=
Can you remember the formula for the ; 7 − ; ! ;!
à So now we need to rewrite this…
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)


3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)
(
(F − 1)! C ')$ 1 − C ()'
You need to be able to approximate a FC U 8×
F − 8 ! (8 − 1)!
binomial distribution '+$


-
(! U / 1 − U -./
=X 0 !×
( − 0 ! 0!
/01
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)

()$
(F − 1)! C , 1 − C (),)$
FC U (W + 1)×
F − W − 1 ! (W)!
,+"
7 7!
=
; 7 − ; ! ;! What if we let # = # − 1
and 3 = 4?
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)

()$ ()$
F−1 , (),)$ F−1 , (),)$
FC U W× C 1−C +U C 1−C
W W
,+" ,+"
!
2
U 0.4, 0.6!),
W
,+"
When 4 = 0 When 4 = 1 When 4 = 2
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)

()$ ()$
F−1 , (),)$ F−1 , (),)$
FC U W× C 1−C +U C 1−C
W W
,+" ,+"
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)

()$
This term is saying to multiply all the possible values of 4 by
FC U W×C W + 1 their probabilities, and then add the answers up
,+"
à What does that give us?
à It gives us the expected value of 4
à And how do we calculate the expected value if we know the

number of trials and the probability of success?
= FC F − 1 C + 1
à Multiply the number of trials by the probability of success!
In this case, the probability of success is U, and

the number of trials is (( − 1)
So the expected value of 4 is ( − 1 U
3F
F * ()* %~9(7, 8)
- !=O = C 1−C
O
%~2((, +)
The normal distribution I=

( = 78
(U(1 − U)
!
You need to be able to approximate a !
∑ 0! ∑0
I = −
binomial distribution ( (
If you remember (quite a while back),

we were finding an expression for the
variance
The expression we just found is that:
(U ( − 1 U + 1
à A reminder that we were finding

this so we could use it to replace the
expected value of 0 !, which we can
now do!
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)

A biased coin has ! Z*1) = 0.53. The

coin is tossed 100 times and the
number of heads, X, is recorded.
a) Write down a binomial model for 5
b) Explain why 5 can be approximated

using a normal distribution
c) Find the values of H and I in this

approximation
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)
You need to be able to approximate a à Something to be careful is that the normal

binomial distribution distribution is continuous and the binomial
distribution is discrete
The binomial random variable à You need to apply a ‘continuity correction’

5~7(150,0.48) is approximated by the
normal random variable [~L(72, 6.12!).
a) Use this approximation to find

!(5 ≤ 70)
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)

The binomial random variable

5~7(150,0.48) is approximated by the
normal random variable [~L(72, 6.12!).
a) Use this approximation to find

!(5 ≤ 70)
b) Also use the approximation to find

!(80 ≤ 5 < 90)
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)

For a particular type of flower bulb,

55% will produce yellow flowers. A
random sample of 80 bulbs is planted.
Calculate the percentage error

incurred when using a normal
approximation to estimate the
probability that there are exactly 50
yellow flowers.
à First, calculate the actual

probability using the binomial
distribution
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)



yellow flowers.
! [ = 50 = 0.0365
à Now convert the problem to a
normal distribution…
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)



yellow flowers.
! [ = 50 = 0.0365
( = 44 + = 4.45 …
3F
%~9(7, 8)
%~2((, +)

( = 78
(U(1 − U)



yellow flowers.
! [ = 50 = 0.0365
( = 44 + = 4.45 …
3F
You need to be able to do hypothesis
testing using the normal distribution
Imagine we had a set of data as follows:
1 2 3 4 4 5 5 6
6 6 7 7 8 8 9 10 11
For this data, the information is as

follows:
!~#(6, 2.655 )
3G
You need to be able to do hypothesis à Imagine we took samples from this data, all of
testing using the normal distribution size 2 (ie – choosing 2 values at random)
à For example:
Imagine we had a set of data as follows:
Value 1 Value 2 Mean
4 5 4.5
1 2 3 4 4 5 5 6
3 9 6
6 6 7 7 8 8 9 10 11
6 5 5.5
%~2(6, 2.65! ) 2 8 5
7 10 8.5
6 8 7
à Let’s now calculate the mean and variance of this
data set…
Sample mean =
Variance (of the sample means) =
Standard Deviation (of the sample means) =
3G
testing using the normal distribution 4 5 4.5
3 9 6
Imagine we had a set of data as follows: 6 5 5.5
2 8 5
1 2 3 4 4 5 5 6 7 10 8.5
6 6 7 7 8 8 9 10 11 6 8 7
%~2(6, 2.65! )
?
%~2(6.08, 1.33! )
Population range:
Sample size 2 range:
Sample size 10 range:
3G
testing using the normal distribution 4 5 4.5
3 9 6
Imagine we had a set of data as follows: 6 5 5.5
2 8 5
1 2 3 4 4 5 5 6 7 10 8.5
6 6 7 7 8 8 9 10 11 6 8 7
%~2(6, 2.65! ) à The key idea to take from this is that with a larger
sample size, the range of possible values is lower
?
%~2(6.08, 1.33! )
à Hence, the variance is also going to be lower as there
is a smaller range for the values to exist in
Population range: 1-11
à When dealing with a sample, the variance is
Sample size 2 range: 1.5-10.5 estimated by dividing the population variance by the
size of the sample
Sample size 10 range: 4.2-7.8 à That way, a larger sample will have a smaller
estimated value for the variance…
3G
When hypothesis testing with a

population, you use values as follows:
%~2((, + ! )
With a sample, we use the following

relationship instead:
+!
?
%~2 (,
7
3G
%~2((, + ! )

+!
?
%~2 (,
7
1) Write your Hypotheses

testing using the normal distribution ?" : A = 60 ?$ : A < 60
A company sells fruit juice in cartons. The
amount of juice in a carton has a normal
distribution with a standard deviation of 3ml.
The company claims that the mean amount of

juice per carton, ", is 60ml. A trading
inspector has received complaints that the 2) Convert the population distribution to the sample
company is overstating the amount of juice distribution
per carton and wishes to investigate this
complaint.
The inspector takes a random sample of 16

cartons and finds that the mean amount of
juice per carton is 59.1ml.
Using a 5% significance level, and stating your 3) Now, using the sample distribution, we need to find
hypotheses clearly, test whether or not there the probability that a random sample of 16 gives a mean
is sufficient evidence to uphold the equal to or lower than 59.1…
complaints.
3G
%~2((, + ! )

+!
?
%~2 (,
7

A company sells fruit juice in cartons. The 3) Now, using the sample distribution, we need to find
amount of juice in a carton has a normal the probability that a random sample of 16 gives a mean
distribution with a standard deviation of 3ml. equal to or lower than 59.1…

inspector has received complaints that the
company is overstating the amount of juice
complaint.
Area
=?
cartons and finds that the mean amount of 59.1 60
Use your calculator as you have seen in prior sections
Using a 5% significance level, and stating your to calculate the probability of being below 59.1
hypotheses clearly, test whether or not there
is sufficient evidence to uphold the
complaints.
3G
%~2((, + ! )

+!
?
%~2 (,
7

A company sells fruit juice in cartons. The

amount of juice in a carton has a normal
distribution with a standard deviation of 3ml.

inspector has received complaints that the
company is overstating the amount of juice
complaint.

cartons and finds that the mean amount of
Using a 5% significance level, and stating your

hypotheses clearly, test whether or not there
is sufficient evidence to uphold the
complaints.
3G
%~2((, + ! )

+!
?
%~2 (,
7

It is also possible to use the

standardised normal distribution when
using a sample
Population Sample
Variance
Variance
+!
%~2((, + ! ) ?
%~2 (,
7
Converting values to Converting values to
the standard normal Square the standard normal Square
distribution root the distribution root the
variance variance
Standard
deviation Standard
deviation
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
1) Write your Hypotheses

A machine produces bolts of diameter # à Remember that the à The alternative hypothesis is
null hypothesis is the where we consider the
where # has a normal distribution with mean default position ie) statement in the question. In
0.580cm and standard deviation 0.015cm. Nothing has changed this case, the claim is that the
mean has changed (could be
This machine is serviced and after the service higher or lower)
a random sample of 50 bolts from the next
production is taken to see if the mean 2) Convert the population distribution to the sample
diameter of the bolts has changed from distribution (using 5 for diameter)
0.580cm. The distribution of the bolts after
the service is still normal with a standard
deviation of 0.015cm.
a) Find, at the 1% level, the critical region

for this test, stating your hypotheses
clearly
3) Now, using the sample distribution, we need to find
à Remember that the critical region is the the regions that have a probability of 1% or less (ie
region where the null hypothesis would be have an total area of 0.01)
rejected…
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
0.015!
?" : A = 0.580 ?$ : A ≠ 0.580 \ ~L 0.580,
]
You need to be able to do hypothesis 50
testing using the normal distribution 3) Now, using the sample distribution, we need to find
the regions that have a probability of 1% or less (ie
A machine produces bolts of diameter # have an total area of 0.01)
where # has a normal distribution with mean
0.580cm and standard deviation 0.015cm. Specific
distribution
This machine is serviced and after the service
production is taken to see if the mean
diameter of the bolts has changed from Area = Area =
0.580cm. The distribution of the bolts after 0.005 0.005
deviation of 0.015cm. ? 0.580 ?
Standard
a) Find, at the 1% level, the critical region distribution
clearly
à Remember that the critical region is the Area = Area =

region where the null hypothesis would be 0.005 0.005
rejected…
? 0 ?
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
0.015!
?" : A = 0.580 ?$ : A ≠ 0.580 \ ~L 0.580,
]
distribution
deviation of 0.015cm. ? 0.580 0.5854
?
Standard
clearly

rejected…
−2.5758 0 2.5758
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
0.015!
?" : A = 0.580 ?$ : A ≠ 0.580 \ ~L 0.580,
]
distribution
deviation of 0.015cm. ?
0.5745 0.580 0.5854
Standard
clearly

rejected…
−2.5758 0 2.5758
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
0.015!
?" : A = 0.580 ?$ : A ≠ 0.580 \ ~L 0.580,
]
distribution
deviation of 0.015cm. 0.5745 0.580 0.5854
Standard
clearly

rejected…
−2.5758 0 2.5758
3G
%~2((, + ! ) %−(
5= +

+! 7
?
%~2 (,
7
?" : A = 0.580 ?$ : A ≠ 0.580

A machine produces bolts of diameter #

0.580cm and standard deviation 0.015cm.

diameter of the bolts has changed from
0.580cm. The distribution of the bolts after
deviation of 0.015cm.
a) Find, at the 1% level, the critical region

clearly
[ ≤ 0.575 or 5
5 [ ≥ 0.585
b) The mean diameter of a sample of 50

bolts is found to be 0.587mm. Comment
on this in light of the critical region
3G

3) M A2 The Normal Distribution

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

3) M A2 The Normal Distribution

Uploaded by

Copyright:

Available Formats

Prior knowledge check

In year 12 you will have

So although we haven not included a y-axis, the idea is that

The normal distribution focuses on

à For a quantity such as height, the

For example, if we decided to model the

à Imagine that the mean height was

à The curve will be split into equal

à The area under the curve is used to

à The normal distribution has 2

à It has asymptotes at each end

à The total area beneath it is equal to 1

à It has points of inflexion at H + I and

à Approximately 68% of the data is So what would

à Approximately 95% of the data is

163$% 175$% 187$% '

à This means the area on the diagram above would

The diameters of a rivet produced by a

à You should always start with a quick

The diameters of a rivet produced by a

7.6%% 8%% 8.4%% %

à You should always start with a quick

The diameters of a rivet produced by a

à You should always start with a quick

Note that the formula used to generate

This means the area under

Given that 5~L(30, 4!), find: 30 33

a) ! 5 < 33 à Now press 2 for the normal cumulative

Even though we will use your

Given that 5~L(30, 4!), find: 24 30

à Input the lower limit (in this case

Given that 5~L(30, 4!), find: 30 33.5 38.2

a) ! 5 < 33 à Now input both limits,

Given that 5~L(30, 4!), find: 27 30 32

d) !(5 < 27 $+ 5 > 32)

Even though we will use your

An IQ test is applied to a population of

a) Find the probability that an adult

b) Twenty adults take the test. Find

An IQ test is applied to a population of

a) Find the probability that an adult

b) Twenty adults take the test. Find

Given that !~# 20, 3! , find, to two

Draw a sketch to get an idea of where

Given that !~# 20, 3! , find, to two

à Be careful here! The calculator

à After putting the mean and standard deviation in as well, the

Draw a sketch to get an idea of where

Plates made using a particular

à The mean is 20, the standard deviation is 1.5, and

b) Find the interquartile range of the

a) Given that 60% of plates are less

à So a value of 31 in the original

The normal distribution

The The standard

The normal distribution

à Now, we code the data using the

The normal distribution

a) -(! < 53)

The normal distribution

The random variable !~#(50, 4! ). Write

a) -(! < 53)

à You can also think of it is ’53 is 0.75 0 0.75