A – level Statistics

The Normal distribution

The normal distribution is one of the most important distributions in statistics. The height of
people in a village or the salaries of civil servants can be modelled by the normal distribution.

The graph of y= f(x) is shown below

The normal distribution has the following properties

 Its bell shaped

 Its symmetrical about the mean µ
 The mean = the mode = median
 The area under the curve is equal to 1
 The horizontal axis is asymptotic to the curve that is the curve does not touch the
horizontal axis
 It ranges from - ∞ to +∞
 About 68% of the distribution lies within 1 standard deviation of the mean.
If random variable X follows a normal distribution it represented by notation
X ̴N(µ, ơ 2)
The value of the standard deviation ơ determines the shape of the distribution
The smaller the value of ơ the more clustered are the values of X around the mean µ, the
greater the value of ơ the more spaced are the values of X around the mean.

The probabilities are determined by the area under the curve

The standard normal distribution

A continuous random variable that follows a normal distribution with mean equal to 0 and
standard deviation equal to 1 is said to have a standard normal distribution denoted by
Z ̴ N(0,1)

Any normal variable can be mapped onto the standard normal by the equation
Z =( X-µ)/ơ

eg If X ̴ N(50,12 2) then Z=( X – 50)/12

Areas under the normal curve can be read off from the standard normal tables .
The cumulative distribution function of the standard normal
distribution
P(Z≤ z ) = Ф (z)

If X ̴N(µ,ơ2 ) then P(X ≤a) = P[ Z ≤ (a - µ )/ơ ] = Ф[(a- µ)/ơ ]

and P( a ≤ X ≤ b) = P[(a - µ)/ơ ≤ Z ≤ (b - µ)/ơ]

Note
P (Z ≤ a) = Ф (a)

a
P (Z ≥ a) = 1 – Ф (a)

P (Z ≤ - a) = 1 – Ф (a)
 -a

P (Z ≥ -a) = Ф (a) [ using the symmetrical properties of the normal distribution]

-a
P( a≤ Z ≤ b) = Ф(b) – Ф(a)

a b
P(IzI < a) = 2Ф(a) – 1
P(IzI > a)= 2[1 – Ф(a)]
Below are the cumulative standard normal tables
Examples on how to use the standard normal tables

Let Z ̴ N (0, 1)
Find
a) P (Z≤ 1.96)
b) P (Z≤ - 0.843)
c) P (Z≥ 2.15)
d) P (Z≥ -1.247)
e) P (1.06≤ Z ≤ 2.72)
Solutions
a) P(Z ≤ 1.96) = Ф(1.96)
In the tables locate a value of z equal to 1.9 then move across the row to a column under 6,
the entry is .9750

1.96

so P(Z ≤ 1.96) = 0.9750

b) P(Z ≤ - 0.843) = 1- Ф(0.843)
In the tables locate a value of z equal to 0.8 then move across the row to a column under 4 ,
the entry is .7995 then move across to the right hand column under 3 ,the entry is 8 so add
8 to the decimal fraction 7795 to get 8003
 -0.843
So P(Z ≤ - 0.843) = 1 – 0.8003
= 0.1997
c) P(Z ≥ 2.158) = 1 – Ф(2.158)
= 1 – 0.9845
= 0.0155
d) P(Z ≥ -1.842) = Ф( - 1.842) = 0.9672
e) P(1.06 ≤ Z ≤ 2.72) = Ф(2.72)- Ф(1.06)
= 0.9967 – 0.8554
= 0.1413
f) P( -0.639 ≤ Z ≤ 1.465) = Ф(1.465) – Ф( -0.639)
= Ф(1.465) – [1 – Ф(0.639)]
= Ф(1.465) + Ф(0.639) – 1
= 0.9286 + 0.7386 – 1
= 0.6672
Exercise 1
1.Given that Z ̴ N(0,1) Find
a) P(Z ≤ 1.162)
b) P(Z ≤ - 0.911)
c) P(Z ≥ 2.057)
d) P(Z ≥ - 1.836)
e) P( 0.864 ≤ Z ≤ 1.398)
2. Given that Z ̴ N(0,1) Find
a) P( -1.23 ≤ Z ≤ 1.23)
b) P( - 2.345 ≤ Z ≤ - 1.228)
c) P( IzI ≤ 1.281)
d) P(IzI ≥ 2.25)
e) P( Z ≤ - 1.883 or Z ≥ 2.512)

Mapping any normal variable onto the standard normal

Let X be a normal random variable distributed as follows
X ̴ N(µ,ơ2) then

Examples
1. Let X ̴ N( 30, 9) find
a) P( X < 35)
b) P( X > 26)
c) P( 25 <X< 30)
d) P(Ix -30I < 5)
solution
a) P( X< 35) = P[ Z < (35-30)/3]
= P(Z< 1.667)
= Ф(1.667)
= 0.9522
b) P(X > 28) =P[Z > (26 – 30)/3]
= P(Z > - 1.333)
= Ф(1.333)
= 0.9087
 c) P(25< X< 30) = P[(25-30)/3 < Z < (30 – 30)/3]
= P( -1.667 < Z < 0)
= Ф(0) – Ф( -1.667)
= Ф(0) + Ф(1.667) – 1
= 0.5000 + 0.9522 – 1
= 0.4522
d) P(Ix -30I < 5) = P(Ix – 30I/3 < 5/3)
= P(IzI < 1.667)
= 2Ф(1.667) – 1
= 2(0.9522) -1
= 0.9044
2. The masses m of pigs delivered to a certain abattoir by a farmer are normally distributed
with mean 120kg and standard deviation 15 kg ,a delivery of 200 pigs is made to the
abattoir in one month. Find the expected number of pigs with a mass,
a) greater than 132kg
b) between 100kg and 115kg
Solution
a) P(m > 132) = P[Z > (132 – 120)/15]
= P(Z > 0.8)
= 1 – Ф(0.8)
= 1 – 0.7881
= 0.2119
The expected number of pigs with a mass greater than 132kg is equal to
0.2119 x 200 = 42.38
Approximately 42 pigs
b) P(100 < m < 115) = P[(100 – 120)/15 < Z < (115 – 120)/15]
= P(-0.667 < Z < - 0.333)
 = Ф(-0.333) – Ф(- 0.667)
= [1 – Ф(0.333)] – [1 – Ф(0.667)]
= (1 – 0.6304) – (1 – 0.7477)
=0.1173
The expected number of pigs with a mass between 100kg and 115 kg is equal to
0.1173 x 200 = 23.46
Approximately 23 pigs.
Exercise 2
1. X ̴ N( 10,3) find
a) P(X < 8.5)
b) P(X > 11.2)
c) P( 9.4 < X < 10.8)
d) P(Ix-10I < 4)
e) P(Ix-10I > 2.8)
2.It takes on average 35 minutes with a standard deviation of 6 minutes for mechanics at a
certain workshop to fit a spare part on a vehicle, assuming the fitting times are normally
distributed find the probability that fitting the spare part will take
a) more than 38 minutes
b) less than half an hour
c) between 28 minutes and 34 minutes
3. A Zupco bus has 66 passengers on a church trip.the ages of the passengers are normally
distributed with mean of 48 years and standard deviation of 12 years.
How many passengers approximately are :
a) between 35 and 60 years old
b) older than 55 years
c) younger than 40 years.
Finding Z when given Ф(z). Using the tables in reverse
Given that Ф(z) = 0.9738 find z
We will use the standard normal tables in reverse. In the tables locate the entry .9738 or
the one nearer to .9738 but less than .9738.
The entry .9738 corresponds to the z value 1.94 therefore if
Ф(z) = 0.9738, z = 1.94
Given that Ф(z) = 0.8692 find z
In the tables we don’t have the entry .8692 but the closet entry is .8686 corresponding to
the z value 1.12, now the difference between the decimal parts of .8686 and .8692 is 6
move across the row with .8686 to the right hand columns, the difference of 6 is under the
column with a 3 at the top so add the digit 3 in front of the z value 1.12 to get the value
1.123.
If Ф(z) = 0.8692 then z = 1.123

Example 1
Find the value of a such that
a) P(Z < a) = 0.7459
b) P(Z < a) = 0.2315
c) P(Z > a) = 0.8876
d) P(Z > a) = 0.1432
Solutions
a) P(Z < a) = 0.7459
Ф(a) = 0.7459

a
a = Ф-1(0.7459)
from the graph we can see that a has a positive value and from the tables z = 0.662
if P(Z < a) =0.7459 then z = 0.662

b) P( Z < a) = 0.2315

Ф(a) = 0.2315

-a

From the graph we can tell that a is negative

Ф( -a) = 1 – Ф(a)

= 1 – 0.2315

= 0.7685

-a = Ф-1(0.7685)

= 0.734

a = - 0.734

c) P(Z>a)= 0.8876

-a

Ф( - a)=0.8876
-a = Ф-1(0.8876)

= 1.214

a= -1.214

d) P(Z > a)= 0.1432

1 – Ф(a) = 0.1432

Ф(a) = 1- 0.1432

a = Ф-1(0.8568)

= 1.066

Exercise 3

Find the value of a given that

a) P (Z< a) = 0.8584

b) P (Z< a) = 0.3476

c) P (Z> a) = 0.9215

d) P (Z> a) = 0.1874

e) P (IzI < a) = 0.7556

Example 2

The masses of upper 6 students at a college are normally distributed with mean 64 kg and
standard deviation 5 kg

Given that 90% of the students have a mass greater than m kg, find the value of m

Solution

Let X be the mass of a student

X ̴N(64,52)
P(X > m) = 0.90

P[Z> (m – 64)/5] = 0.90 -z

z = (m-64)/5
z is negative
Ф(-z) = 0.90
-z= Ф-1(0.90)
z = - 1.282µ
(m-64)/5 = -1.282
m = 5(-1.282) + 64
m = 57.59 kg

Example 3
The height X of trees in a rainforest are normally distributed with mean µ and variance ơ 2 ,
given that P(X >30) = 0.3 and P(X > 25.4) = 0.65 calculate to 1 decimal place the values of µ
and ơ.
P(X > 30) = 0.3
P[Z > (30-µ)/ơ] = 0.3
1 –Ф[(30- µ)/ơ]= 0.3
Ф[(30-µ)/ơ]= 0.70
30-µ/ơ = Ф-1(0.70)
30-µ = 0.524ơ ………………………………….eqn 1
P(X > 25.4) = 0.65
P[Z>(25.4 - µ)/ơ] =0.65
25.4 - µ/ơ = -Ф -1(0.65)
 25.4 - µ = - 0.385ơ ……………………………..eqn 2
Solving the two equations simultaneously
µ= 27.3 and ơ = 5.1 to 1 dp.

Exercise 4
a) Random variable X is normally distributed with mean µ and variance 25, given that
P(X < 27.5) = 0.8686 calculate the value of µ
b) If X is normally distributed as follws X ̴N(µ,ơ2), given that P(X > 58.37) = 0.02 and that
P(X<40.85) = 0.01 show that µ = 50.16 and find the value of ơ
c) Random variable X is normally distributed with mean 120 and variance 64 find
i) the lower and upper quartiles
ii) the 35th percentile.

The normal approximation to the binomial distribution

The normal distribution can be used to approximate probabilities of the binomial distribution
but only under certain conditions.
If random variable X is binomially distributed as follws X ̴Bin(n,p) and n is large (n≥30) and p is
not very small such that np >5 and nq > 5 then X can be approximated by the random variable Y
which follows a normal distribution with mean µ = np and variance ơ2 = npq
X ̴Bin(n,p) ≈ Y ̴ N(np,npq)
Since the binomial distribution is discrete and the normal distribution is continuous an
allowance should be made using the ½ continuity correction

The approximations are as follows.

If X ̴ Bin(n,p) is approximated by Y ̴ N( np,npq)
P(X< 5) ≈ P(Y < 4½)
P(X ≤ 5) ≈ P(Y < 5½)
P(X >5) ≈ P(Y>5½)
P(X ≥ 5) ≈ P(Y > 4½)
P(3≤ X < 5 ) ≈ P( 2½ < Y< 4½)
P(X = 5) ≈ P( 4½ < Y< 5½)

Example
Random variable X follows a binomial distribution as follows X ̴Bin(225;0.8), using suitable
approximation find
a) P(X > 190)
b) P( 170 ≤ X < 175)
Solution
np = 180 >5 , nq = 45 >5 so we can use the normal approximation.
µ =np = 180 , ơ2= npq = 36

a) P( X>190) ≈ P(Y > 190½)

≈P[Z> (190½ - 180)/6]
≈P(Z > 1.75)
≈1- Ф(1.75)
≈1-0.9599
≈ 0.0401

b) P( 170 ≤ X < 175) ≈ P(169½ < Y < 174½)

≈ P[(169.5- 180)/6 < Z< (174.5-180)/6]
≈ P(-1.75<Z<-0.9167)
≈ Ф(-0.9167) – Ф(-1.75)
≈ Ф(1.75) – Ф(0.9167)
≈ 0.9599 – 0.8204
 ≈ 0.1395

Exercise 6
1. A fair coin is tossed 100 times , using suitable approximation find the probability of getting at
least 60 heads.
2. 15% of items produced by a certain factory are defective, in a day the factory produces 1000
items. Find the probability that on a randomly selected day the factory produces
a) less than 140 defective items
b) at least 165 defective items
c) between 145 and 170 (inclusive) defective items