The normal distribution

Challenge
A normally distributed random variable X ~ N(µ, σ 2) has interquartile
range q.
a Show that σ = 0.742q, where the coefficient of q is correct to 3 s.f.
b Explain why it is not possible to write µ in terms of q only.

3.6 Approximating a binomial distribution

Consider the binomial random variable X ~ B(n, p). Links
The cumulative binomial tables in the
It can be difficult to calculate probabilities for X formulae booklet only go up to n = 50.
when n is large. In certain circumstances you can ← Year 1, Chapter 6
use a normal distribution to approximate a
binomial distribution.

X ∼ B(50, 0.48)
Y ∼ N(24, 12.48)

0 5 10 15 20 25 30 35 40 45

You need to understand the conditions under which this approximation is valid, and learn the
relationship between the values of n and p in B(n, p) and the values of µ and σ in the normal
approximation N(µ, σ2).
■ If n is large and p is close to 0.5, then the binomial Hint
The approximation is only
distribution X ~ B(n, p) can be approximated by
valid when p is close is to 0.5
the normal distribution N(µ, σ2) where
because the normal distribution is
● µ = np
_________ symmetrical.
● σ = √​​  np(1 − p) ​​ 

Example 11
A biased coin has P(Head) = 0.53. The coin is tossed 100 times and the number of heads, X, is
recorded.
a Write down a binomial model for X.
b Explain why X can be approximated with a normal distribution, Y ~ N(µ, σ2).
c Find the values of µ and σ in this approximation.

a X ~ B(100, 0.53)
b The distribution can be approximated with a normal
distribution since n is large and p is close to 0.5. Use μ = np
c μ = 100 × 0.53 = 53
______________________ ________
σ = √​​  100 × 0.53 × (1 − 
    0.53) ​​ = 4.99 (3 s.f.) Use σ = √​​  np(1 − p) ​​  

53

M03_EDSM2_SB_AL_83282_U03_037-064.indd 53 16/06/2017 12:30

 Chapter 3

The binomial distribution is a discrete distribution but the normal distribution is continuous.
■ If you are using a normal approximation to a binomial distribution, you need to apply a
continuity correction when calculating probabilities.
The diagrams show X ~ B(14, 0.5) being approximated by Y ~ N(7, 1.872):

Y ∼ N(7, 1.872) Y ∼ N(7, 1.872) Y ∼ N(7, 1.872)

0 1 2 3 4 5 6 7 8 9 1011121314 0 1 2 3 4 5 6 7 8 9 1011121314 0 1 2 3 4 5 6 7 8 9 1011121314

P(X > 5) ≈ P(Y > 5.5) P(X = 9) ≈ P(8.5 < Y < 9.5) P(X < 3) ≈ P(Y < 3.5)

Example 12
The binomial random variable X ~ B(150, 0.48) is approximated by the normal random variable
Y ~ N(72, 6.122).
Use this approximation to find: Watch
out Remember to apply the continuity
correction. You are interested in values of the
a P(X < 70) b P(80 < X < 90)
discrete random variable X that are less than or
a P(X < 70) ≈ P(Y < 70.5) = 0.4032 (4 d.p.) equal to 70, so you need to consider values less
than 70.5 for the continuous random variable Y.
b P(80 < X < 90) ≈ P(79.5 < Y < 89.5)
= 0.9979 – 0.8898 For values of X less than 90 consider values of Y
= 0.1081 (4 d.p.) less than 89.5.

Example 13
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.

Let X = the number of bulbs producing yellow flowers Define a suitable binomial random variable.
in a sample of 80.
Use your calculator to find the exact
Then X ~ B(80, 0.55)
probability using a binomial
P(X = 50) = ​(​  
50)
80
​ ​0.​55​​  50​  0.​45​​  30​ = 0.0365 distribution. ← Year 1, Chapter 6
X can be approximated by the normal distribution
Use μ = np
Y ~ N(μ, σ2), where μ = 80 × 0.55
_____________________
= 44
_____
σ = √​​  80
   × 0.55 × (1 – 0.55) ​​ = √​​  19.8 ​​   (3 s.f.)
Write down the normal approximation.
Y ~ N(44, 19.8)
P(X = 50) ≈  P(49.5 < Y < 50.5) To estimate the probability that X takes a
= 0.9280 – 0.8918 = 0.0362 (4 d.p.) single value, apply a continuity correction
0.0365 − 0.0362
Percentage error = __________________
  
​      ​  × 100 = 0.82% by considering values half a unit below and
0.0365 half a unit above.

54

M03_EDSM2_SB_AL_83282_U03_037-064.indd 54 16/06/2017 12:30

 The normal distribution

Exercise 3F
1 For each of the following binomial random variables, X:
i state, with reasons, whether X can be approximated by a normal distribution.
ii if appropriate, write down the normal approximation to X in the form N(μ, σ2), giving the
values of μ and σ.
a X ~ B(120, 0.6) b X ~ B(20, 0.5) c X ~ B(250, 0.52)
d X ~ B(300, 0.85) e X ~ B(400, 0.48) f X ~ B(1000, 0.58)

2 The random variable X ~ B(150, 0.45). Use a suitable approximation to estimate:

a P(X < 60) b P(X > 75) c P(65 < X < 80)

3 The random variable X ~ B(200, 0.53). Use a suitable approximation to estimate:

a P(X < 90) b P(100 < X < 110) c P(X = 105)

4 The random variable X ~ B(100, 0.6). Use a suitable approximation to estimate:

a P(X > 58) b P(60 < X < 72) c P(X = 70)

5 A fair coin is tossed 70 times. Use a suitable approximation to estimate the probability of
obtaining more than 45 heads.
50
6 The probability of a roulette ball landing on red when the wheel is spun is ____
​​    ​​ 
.
101
On one day in a casino, the wheel is spun 1200 times.
Estimate the probability that the ball lands on red in at least half of these spins.

E 7 a Write down two conditions under which the normal distribution may be used as an
approximation to the binomial distribution. (2 marks)
A company sells orchids of which 45% produce pink flowers.
A random sample of 20 orchids is taken and X produce pink flowers.
b Find P(X = 10). (1 mark)
A second random sample of 240 orchids is taken.
c Using a suitable approximation, find the probability that fewer than 110 orchids
produce pink flowers. (3 marks)
d The probability that at least q orchids produce pink flowers is 0.2. Find q. (3 marks)

E 8 A drill bit manufacturer claims that 52% of its bits last longer than 40 hours.
A random sample of 30 bits is taken and X last longer than 40 hours.
a Find P(X < 17). (1 mark)
A second random sample of 600 drill bits is taken.
b Using a suitable approximation, find the probability that between 300 and 350 bits
last longer than 40 hours. (3 marks)

E/P 9 A particular breakfast cereal has prizes in 56% of the boxes. A random sample of 100 boxes
is taken.
a Find the exact value of the probability that exactly 55 boxes contain a prize. (1 mark)
b Find the percentage error when using a normal approximation to calculate the
probability that exactly 55 boxes contain prizes. (4 marks)

55

M03_EDSM2_SB_AL_83282_U03_037-064.indd 55 16/06/2017 12:30

 Chapter 3

3.7 Hypothesis testing with the normal distribution

You can test hypotheses about the mean of a normally distributed random variable by looking at the
mean of a sample taken from the whole population.
■ For a random sample of size n taken from a Hint
If you took lots of different random samples
random variable X ~ N(μ, σ2), the sample
¯,  is normally distributed with of size n from the population, their means would
mean, ​ X ​
be normally distributed.
¯ ~ N​​(μ, ___
​ X ​ ​  n ​ )​​.
σ2

You can use the distribution of the sample mean to determine whether the mean from one particular
​​  ¯,  is statistically significant.
sample, x ​​

Example 14
A certain company sells fruit juice in cartons. The amount of juice in a carton has a normal
distribution with a standard deviation of 3 ml.
The company claims that the mean amount of juice per carton, μ, is 60 ml. A trading inspector has
received complaints that the company is overstating the mean amount of juice per carton and he
wishes to investigate this complaint. The trading inspector takes a random sample of 16 cartons
and finds that the mean amount of juice per carton is 59.1 ml.
Using a 5% level of significance, and stating your hypotheses clearly, test whether or not there is
evidence to uphold this complaint.

H0 : μ = 60 The null hypothesis, H0, is that the population

mean is equal to the claimed value.
H1 : μ < 60
Let X represent the amount of juice in a The inspector is investigating whether the
carton and assume H0, so that X ∼ N(60, 32). population mean is less than 60, so this is a one-

( 16 )
32
¯ ∼ N​​ 60, ___ tailed test.
​​ X ​​ ​   ​  ​​  ¯ ∼ N(60, 0.752)
​​ or X ​​

Write out the population distribution assuming

that H0 is true.
N(60, 0.752)
Watch
out Your test statistic will be the
¯.  This will have the same mean
sample mean, ​​ X ​​
as X, but you need to divide the variance by the
​3​​  2​
sample size. The new variance is ___
​​    ​​ so the new

√ 
___ 16
​3​​  2​
standard deviation is ​​ ___
​    ​ ​​  = 0.75.
59.1 60 16
¯ < 59.1) = 0.1151
P(​​ X ​​
¯ < x ​​
Use your calculator to find P(​​ X ​​ ​​  ¯) .
0.1151 > 0.05 so there is insufficient evidence
to reject H0 and conclude that the mean
Compare P(​​ X ​​¯ < 59.1) with the significance level
amount of juice in the whole population is less
of the test. The probability of obtaining this value
than 60 ml.
​​  ¯ is greater than 5%, so you do not reject the
of x ​​
null hypothesis. Make sure your conclusion refers
to the context given in the problem.

56

M03_EDSM2_SB_AL_83282_U03_037-064.indd 56 16/06/2017 12:30