7 Normal Distribution

Candidates should be able to:

Understand the use of normal distribution to model a continuous random
variable.

Use normal distribution function tables accurately.

Solve problems using a normal distribution.

Find a relationship between x, μ, and σ given the value of P (X > x)or its
equivalent.

Recall conditions under which normal distribution can be used as an

approximation to the binomial distribution.

Solve problems using the normal approximation, with a continuity correction.

Definitions
Normal It is a Continuous Probability Distribution graphed as a Gaussian
Distribution Curve, also known as a bell curve.

It is used whenever a discrete distribution is approximated by a

Continuity
continuous one. Adjustments are made at the start and end of the
Correction
given limits.

7.1 Normal Distribution

7.1.1 Characteristics of a bell curve
The curve is symmetrical about the vertical axis, centered at the mean.

7 Normal Distribution 1
 The curve approaches the horizontal axis asymptotically on either side of the
vertical axis.

The area under the curve for a given interval is equal to the probability of a
normal random variable for that interval.

The total area under the curve is 1.

The probability density function for normal random variable X is given by:
2
f(x)= σ 1 − 21 ( x−σ μ )
2π
​
e​
​ ​

, −∞ < x < ∞ 
Description of a normal random variable:

X ∼ N(μ, σ 2 )

The function Φ(z)is used to find the area under the normal distribution curve.
The area given by the function represents the probability of a value smaller
than z .

7 Normal Distribution 2
 7.2 Normal Distribution Characteristics

7.2.1 Standard Normal Random Variable

Every normal random variable X is to be converted to a standard normal random
variable Z in the following way, in order to refer to the standard table for
probabilities:
x− μ
z= σ ​

Here normal random variable X has a mean μwith a standard deviation σ . xis a
particular value of X transformed into z by the above equation.

7.2.2 Significance of inequality sign:

The following are the most common cases encountered when solving
questions with a normal distribution.

CASE 1:

The shaded region shows the area to the left of the z value which is obtained
through the normal distribution probability table.

7 Normal Distribution 3
 CASE 2:

Since the total area under the curve is 1 and the area in the probability table is
given to the left of the z value, for this case we could subtract the probability of
z < a from 1.

7 Normal Distribution 4
 CASE 3:

The z value is negative and and the inequality used is ‘less than (<)’. Note that the
curve is symmetrical about the vertical axis so the shaded region shown on the
left side has the same area on the right side and could be written as P(z > a).

7 Normal Distribution 5
 Case 4:

In between case where the shaded region lies between two z values.

7 Normal Distribution 6
 CASE 5:

Since the value of z is negative it cannot be found directly from the normal
distribution function table. Using symmetry clearly the area to the right of the z 
value is the same as the area to the left of the given z value.

7 Normal Distribution 7
 Important Note
If z values are given, then find the area.
If the area is given, then find the z values

7 Normal Distribution 8
 Skill Check 1

The random variable Z is such that Z ∼ N(0, 1). Find the probability of
P (−1.4 ≤ Z ≤ 1)
Solution

P (−1.4 ≤ Z ≤ 1)= P (Z ≤ 1) − P (Z ≤ −1.4)

= P (Z ≤ 1) −(1 − P (Z ≤ 1.4)) 
= P (Z ≤ 1) − 1 + P (Z ≤ 1.4)
= Φz (1) − 1 + Φz (1.4)
​ ​

= 0.8413 − 1 + 0.9192
= 0.7605
= 0.761, correct to 3 decimal places

7 Normal Distribution 9
 Skill Check 2

Skilled operators make a particular component for an engine. The time taken
to make this component is modeled by the normal distribution with mean μ
minutes and standard deviation σ minutes.

Tony believes that the company is allowing too long for the job and invites
them to time him. They find that only 10% of the components take him over
90 minutes to make, and that 20% take him less than 70 minutes.

Estimate the mean and standard deviation of the time Tony takes.

Solution

This question is framed differently i.e., you have to infer the mean and
standard deviation using the areas provided.
10% of the components take him over 90 minutes to make so the shaded
area should be 0.1as shown in the figure.

To get the left side area: Φ(z) = 1 − 0.1 = 0.9

Finding the z - value from the table for this particular area gives us z =
1.2815
Returning to the problem you know that,
90 90

7 Normal Distribution 10
 90−μ 90−μ
z= σ
​ ⇒ σ
​= 1.2815 (1)

The second piece of information, that 20% of components took Tony

under 70 minutes, is also shown.

Since the point is left to the mean, the z value would be negative. We
know that

P (z ≤ −a) = P (z ≥ a) = (1 − P (z ≤ a))
So, Φ(−z) = 1 − 0.2 = 0.8

The second equation would be

70−μ
σ
​ = −0.842 (2)

Solving both equations (1) and (2) simultaneously gives us the mean and
the standard deviation.

σ = 9.418 = 9.42(3 s.f.)

μ = 77.930 = 77.9(3 s.f.)

7 Normal Distribution 11
 7.3 Normal Approximation
7.3.1 Normal Approximation to Binomial Random
Variable
In case of large number of trials within a binomial process, the binomial formula to
calculate probabilities becomes ineffective or tedious and therefore a normal
approximation can be used to estimate the probabilities.

Step 1: Satisfying the condition

Condition when a normal distribution could be used as an approximation for the

binomial B(n, p)is:

n is large enough that np> 5and nq > 5, where n = number of trials, p =
probability of success and q = 1 − p= probability of failure.

The higher the number of trials, the better will be the approximation.

Step 2: New parameters

The parameters for the normal distribution are:

Mean: μ = np
Variance: σ 2 = npq = np(1 − p)
denoted by N(np, npq).
Step 3: Continuity Correction

Whenever a Continuous Probability Distribution (Normal) is applied on a discrete

probability distribution (Binomial), a continuity correction of ±0.5must be made.
If the region we are interested in is in the form of inequalities then we should see
whether they are inclusive (≤ or ≥) or exclusive (< or >).

For example 20 ≤ X ≤ 30becomes 19.5 ≤ X < 30.5i.e., P (X ≤ 30)

could also be written as P (X < 31)and then continuity correction is subtracted,
whereas 20 < X < 30becomes 20.5 ≤ X < 29.5.

7 Normal Distribution 12
 correction
P (X > a) ⟶ ​ P (X > a + 0.5)
correction
P (X < a) ⟶ ​ P (X < a − 0.5)
correction
P (X ≥ a) ⟶ ​ P (X > a − 0.5)
correction
P (X ≤ a) ⟶ ​
P (X > a + 0.5)

7 Normal Distribution 13
 Skill Check 3

It is given that 40%of the population supports Party X. 150 members of the
population are selected at random. Use a suitable approximation to find the
probability that more than 55 out of the 150 support party X.

Solution

p = 40% = 0.4is the probability of success

n = 150
np=150 × 0.4=60 & nq=150 × 0.6 = 90.
Both npand nqare greater than 5so condiiton the for approximation is
satisfied.
We need to find P (X > 55)but continuity correction is needed.
+0.5
P (X > 55) → P (X ≥ 56) ⟶ P (X ≥ 55.5)
​

Mean =np=60;

Standard Deviation = npq= 150 × 0.4 × 0.6=6

​ ​

Let S be the number of people supporting party X.

P (S > 55.5)=P (Z > z = 55.5−60
6
) ​

=P (Z > −0.75)

=P (z < 0.75)
=Φz (0.75) = 0.7734 ≈ 0.773(from the table)
​

7 Normal Distribution 14
 Skill Check 4

9709/51/June 2022/Q5/c
The lengths of the leaves of another type are also modelled by a normal
distribution. A scientist measures the lengths of a random sample of 500
leaves of this type and finds that 46 are less than 3 cm long and 95 are more
than 8 cm long.
In a random sample of 2000 leaves of the second type, how many would the
scientist expect to find with lengths more than 1 standard deviation from the
mean?

Solution

Answer: 635
Explanation:

If we have a distance of ONE standard deviation from the mean that

means we will have the values of Z as
Z = − 1 and Z = 1
P (Z < − 1) + P (Z > 1) = 0.3174
Now, to calculate the expected value we multiply the probability by the
sample size which is 2000.

= 2000 × 0.3174 = 635

Points to Note
The Normal distribution with mean μand standard deviation σ is denoted by
N(μ, σ 2 ).
This may be given in the standardised form by using the transformation:
x− μ
z= σ

​

7 Normal Distribution 15
 In the standardized form N(0, 1), the mean is 0, and the standard deviation and
variance are both 1.

The standard normal curve is given by:

1 − 21 z 2
Φ(z) = 2π
​
e ​
​

The area to the left of the value z , represents the probability of a value less
than z , denoted by Φ(z)and is read from the normal distribution function
tables.

It is often helpful to know that in a normal distribution, roughly,

68% of the values lie within ±1 standard deviation of the mean

95% of the values lie within ±2 standard deviations of the mean

99.75% of the values lie within ±3 standard deviations of the mean.

Using the symmetry of the normal curve we see that the area to the left of the
negative z value is the same as the area to the right of the positive z value.

The normal distribution may be used to approximate suitable discrete

distributions but continuity corrections are then required.

The Binomial distribution B(n, p) may be approximated by N(np, npq),

provided n is large and p is not close to 0 or 1, so that np > 5and nq > 5

7 Normal Distribution 16