5123 Business Statistics

Assessing Risk and the Normal distribution

Case Study Small Business:
Can cupcakes run a car?
Jan has recently started her own small business making
and selling cupcakes.
The business has started well and she is considering
purchasing a small company car to use for the general
running around necessary for the business, and as a
source of advertising.
She has read in the NRMA newsletter that the average
cost of running a small car is $150 per week based on
driving 15,000 km a year (about the distance she thinks
she will drive).
The cost is assumed to be Normally distributed with a
standard deviation of $10.

2
Case Study Small Business:
Can cupcakes run a car?
The income from the business fluctuates from week to week and

Jan is concerned that in some weeks the business may make a loss
if the cost of the car and her other costs are too high.

Jan thinks in the weeks the car costs more than $163 she will

make a loss. How many weeks in a year will this occur?

She also thinks that in the weeks she has to pay less than $145 for

the car she will have a little bit extra to either put back into the
business or to put away for a holiday. How many weeks in a year
will this happen?

Is the idea financially viable?

3
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

4
 Why we use a statistical model
Ultimately we want to infer about a population.
Too costly to consider all elements of population
Instead we obtain a random sample: should be representative sample

From a random sample of 1000 households, 600 spend above

$175 per week on groceries.
Suggests P(household spends more than $175 pw on groceries) = 0.6

However, if we took another random sample of 1000

households, would we get exactly the same values?

We use a statistical model in order to infer more generally about

the population, not just give a result for a particular sample.
The statistical model for continuous variables is the Normal distribution
5
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

6
 Normal Distribution
Bell-shape
All have the characteristic shape:
unimodal, symmetric, bell-shaped

What distinguishes one from another?

Where is it
centred?

How are the

data spread?

The particular Normal distribution is determined by the:

MEAN (measure of the centre of the distribution); and
STANDARD DEVIATION (measure of the spread of the
data)
7
 Normal Distribution (ctd)
Used for a continuous variable
We usually define the variable (measure of interest) to be X
E.g., Let X represent the weekly grocery expenditure of households

X ~ N(, )
X is Normally distributed, with mean & standard deviation
[Some books use X ~ N(, 2) where 2 is representing the
variance]

E.g., X ~ N(20, 5)
X is Normally distributed, with mean 20 and standard deviation 5

The values of and generate different Normal distributions 8

Normal Curves
Same mean but different standard deviation

Larger standard deviation flatter-looking distribution
Smaller standard deviation more peaked distribution
(since more data are concentrated near mean)
9
 Normal Curves (ctd)
Different mean and different standard deviation

Location of peak reflects mean

Roundness of peak reflects varying standard deviations
(smaller standard deviation then higher peak: more data
close to mean) 10
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

11
The Empirical Rule

Applicable when the distribution is Normal.

Tells us approximately what proportions of

measurements fall within certain ranges.
Approximately 68% of measurements lie in the range
between ( - 1) and ( + 1).
Approximately 95% of measurements lie in the range
between ( - 2) and ( + 2).
Approximately 99.7% of measurements lie in the range
between ( - 3) and ( + 3).
12
The Empirical Rule (ctd)

- 3 - 2 - 1 + 1 + 2 + 3

13
The Empirical Rule (ctd)
If X ~ N(20, 5)

- 3 - 2 - 1 + 1 + 2 + 3
5 10 15 20 25 30
14
35
Lecture Example 1
Using the Empirical Rule
A bank is interested in the time taken for its tellers to
cash money orders. A sample reveals the time is
Normally distributed with a mean of 60 seconds and a
standard deviation of 10 seconds.

Using the empirical rule we know approximately:

Within 50 to 70 seconds, 68% of money orders will
be cashed
Within 40 to 80 seconds, 95% of money orders will
be cashed
Within to seconds, 99.7% (almost all) of
money orders will be cashed
15
 Lecture Example 1 (ctd)
Let X represent time taken for tellers to cash money orders
If X ~ N(60, 10)

- 3 - 2 - 1 + 1 + 2 + 3
30 40 50 60 70
16
80 90
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

17
Normal Distribution calculating
probabilities
Probabilities are areas under a Normal curve
Let X represent the weekly grocery expenditure of households
What is the probability of a household spending less than $150 per week on groceries?

P(X < 150)

90 100 110 120 130 140 150 160 170 180

How do we calculate these probabilities?

18
Standard Normal Distribution
Each variable X that follows a Normal distribution
has a unique and .
To simplify the calculation of probabilities, Normal
distributions can be converted to a standard
Normal distribution using the following
transformation.
X
Z

X ~ N ( , ) Z ~ N (0,1)
Z ~ N(0, 1) is known as the standard Normal
distribution.
19
Standard Normal Distribution (ctd)

-3 -2 -1 +1 +2 +3 X
-3 -2 -1 0 1 2 3 Z

A negative Z value indicates that X is less than the mean.

A positive Z value indicates that X is greater than the mean.
The Z value tells you how many standard deviations your
measurement is away from the mean.
20
Total area and Symmetry
Total area under the standard Normal curve is 1.
So 0.5 above zero and 0.5 below zero (due to symmetry)

-4 -3 -2 -1 0 1 2 3 4
-4 -3 -2 -1 0 1 2 3 4

P(Z > 0) = P(Z < 0) =

21
Using Complements

-4 -3 -2 -1
00 1 z 2 3 4 -4 -3 -2 -1 00 z
1 2 3 4

P(Z < z) P(Z > z) = 1 P(Z < z)

E.g. P(Z < 1) E.g. P(Z > 1) = 1 P(Z < 1)

22
Using Complements (ctd)

-4 -3 -z
-2 -1
00 1 2 3 4 -4 -3 -z
-2 -1 00 1 2 3 4

P(Z < -z) P(Z > -z) = 1 P(Z < -z)

E.g. P(Z < -2) E.g. P(Z > -2) = 1 P(Z < -2)

23
 Aside: subtracting areas to find
required area
Recall a School problem of having a round cement pond in the middle of a
rectangular grassed area and being required to determine the area of grass
which needs to be mown?

Shaded area EQUALS Area of Area of

MINUS circle
rectangle

24
Same principle: subtracting areas to
find required area

-4 -3 -2 z1
-1 00 z2
1 2 3 4

The shaded area (probability) can be found by

finding the area to the left of z2
and subtracting
the area to the left of z1

P(z1 < Z < z2) = P(Z < z2) P(Z < z1)
E.g. P(-2 < Z < 1) = P(Z < 1) P(Z < -2) 25
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

26
 Using Excel (Formulas tab, fx, Statistical)

P(Z < 1.23)

= NORM.S.DIST(1.23, true)
= 0.8907

-4 -3 -2 -1 00 11.232 3 4

P(Z > 2.14)

= 1 P(Z < 2.14)
= 1 NORM.S.DIST(2.14, true)
= 1 0.9838
= 0.0162 -4 -3 -2 -1 00 1 2.14
2 3 4
27
Using Excel (Formulas tab, fx,
Statistical)
P(Z < -1.23)
= NORM.S.DIST(-1.23, true)
= 0.1093

-4 -3 -2 -1
-1.23 00 1 2 3 4

P(Z > -2.14)

= 1 P(Z < -2.14)
1 NORM.S.DIST(-2.14, true)
= 1 0.0162
= 0.9838
-4 -2.14
-3 -2 -1
0
0 1 2 3 4 28
 Using Excel (Formulas tab, fx,
Statistical)
P(-1.96 < Z < 1.96)
= P(Z < 1.96) P(Z < -1.96)
NORM.S.DIST(1.96, true)
NORM.S.DIST(-1.96, true)
= 0.9750 0.0250
= 0.95 -4 -1.96
-3 -2 -1
00 1 1.96
2 3 4

29
Lecture Example 2
Grocery expenditure at Aldi
A random sample of 1000 households in Canberra
revealed weekly household grocery expenditures
were Normally distributed with a mean of $135 and
a standard deviation of $14.30.

Find the probability that the amount spent by a

randomly selected household is:
a) less than $150
b) more than $110
c) between $110 and $150
30
 Lecture Example 2 (ctd)
Let X represent weekly household grocery
expenditure in Canberra
Then X is Normally distributed with =$135 and
=$14.30
In notation form: X~N(135, 14.30)

We are interested in:

a) P(X < $150)
b) P(X > $110)
c) P($110 < X < $150)
31
 Lecture Example 2: a) P(X < $150)
X is Normally distributed with =$135 and
=$14.30

We are interested in P(X < $150)

X
Transform 150
into
X Zby135
using
Z 1.0490
14.30
-4 -3 -2 -1
00 11.0490
2 3 4

P(X < 150) = P(Z < 1.0490)

P(Z < 1.0490) = NORM.S.DIST(1.0490, true)
= 0.8529
32
Lecture Example 2: b) P(X > $110)
X is Normally distributed with =$135 and
=$14.30
We are interested in P(X > $110)
Transform X into Z by using
X 110 135
Z 1.7483
14.30

P(X > 110) = P(Z > -1.7483)

-1 00
-4 -1.7483
-3 -2 1 2 3 4
P(Z > -1.7483) = 1 P(Z < -1.7483)
= 1 NORM.S.DIST(-1.7483, true)
= 1 0.0402 = 0.9598

33
Lecture Example 2: c) P($110 < X
<$150)
X is Normally distributed with =$135 and
=$14.30
We are interested in P($110 < X < $150)
Transform X into Z by using
X 110 135
Z 1.7483
14.30
X 150 135
Z 1.0490
14.30

P(110 < X < 150) = P(-1.7483 < Z < 1.0490) -4-1.7483

-3 -2 -10 0 1.0490
1 2 3 4

P(-1.7483 < Z < 1.0490)

= P(Z < 1.0490) P(Z < -1.7483)
= NORM.S.DIST(1.0490, true) NORM.S.DIST(-1.7483, true)
= 0.8529 0.0402 = 0.8127 34
Lecture Example 3
Small Business: Can cupcakes run a car?
Jan thinks in the weeks the car costs more than $163
she will make a loss. How many weeks in a year will this
occur?

The cost of running a small car is assumed to be

Normally distributed with an average cost of $150 per
week and a standard deviation of $10.

Let X be the cost of running a small car

Then X is Normally distributed with =$150 and =$10
In notation form: X~N(150, 10)
35
Lecture Example 3 (ctd)
X is Normally distributed with =$150 and =$10
We are interested in P(X > $163)
Transform X into Z by using
X 163 150
Z 1.3
10
P(X > 163) = P(Z > 1.3)
P(Z > 1.3) = 1 P(Z < 1.3)
= 1 NORM.S.DIST(1.3, true) -4 -3 -2 -1 00 1 1.32 3 4

= 1 0.9032 = 0.0968
Jan is expected to make a loss in 5 (0.0968 52)
weeks of the year.
36
Lecture Example 3 (ctd)
Jan also thinks that in the weeks she has to pay less
than $145 for the car she will have a little bit extra to
either put back into the business or to put away for a
holiday. How many weeks in a year will this happen?

The cost of running a small car is assumed to be

Normally distributed with an average cost of $150 per
week and a standard deviation of $10.

Let X be the cost of running a small car

Then X is Normally distributed with =$150 and =$10
In notation form: X~N(150, 10) 37
 Lecture Example 3 (ctd)

X is Normally distributed with =$150 and =$10

We are interested in P(X < $145)
Transform X into Z by using

X 145 150
Z 0.5
10

-1 00
P(X < 145) = P(Z < -0.5) -4 -3 -0.5
-2 1 2 3 4

P(Z < -0.5) = NORM.S.DIST(-0.5, true)

= 0.3085

Jan is expected to have a little bit extra in 16 (0.3085 52)

weeks of the year. 38
Lecture Example 3 (ctd)

Is the idea financially viable?

Yes !
The number of weeks Jan is expected to make a
profit is approx. 3 times more than the number
of weeks she is expected to make a loss.

39
Lecture Exercise 1
Website downloads
As an e-business did not want customers to be dissuaded from
their website due to long download times, the time to download
a graphic from its website was assessed.

A random sample of 500 times to download were obtained.

The data were found to be Normally distributed with mean 5.3
seconds and standard deviation 1.2 seconds.

Find the probability that the time to download the graphic is:
a) less than 3.5 seconds
b) less than 7 seconds
c) more than 4.5 seconds
d) more than 6 seconds
e) between 6 seconds and 7 seconds 40
 Key messages: this week you will
learn
Why we use a statistical model to estimate risk
And what the model is for continuous variables
The properties of the Normal distribution
Key notation used for the Normal distribution and
more widely in the unit
The Empirical Rule
The standard Normal distribution and how to use it
to calculate probabilities
Using Excel to calculate the probabilities associated
with the standard Normal distribution

41
Further Reading

Basic Business Statistics 4, by Berenson et al.

Chapter 6

Please note:
The book is a reference only; its the lecture content
which dictates what you read in the book.

42
Next Week
Sampling distributions and Confidence Intervals

43