MSIN0154 Statistics for Business Research

Lecture 2

Probability

Dr Yufei Huang
Contents
• Set Theory
- Definitions
- Union, Intersection, Compliment
• Probability Theory
- Definitions
- Probability rules
- Conditional probability
- Bayes theorem

 Excel

3
Brain Teaser
Q: There are three doors. Behind one of them is a prize. Monty Hall, the
host, asks you to pick a door. Suppose you pick Door #1. Monty opens
one of the other two doors (say Door #3) and show, there is nothing
behind Door #3. He gives you the choice of either sticking with your
original choice, #1, or switching to Door #2.

Should you switch?

a) Yes, you should switch
b) No, do not switch

4
 Set Theory
• Definition. A set is a collection of objects, called elements or members.

• Examples:
– R is the set of all real numbers between -∞ and ∞.
– {1,2,4,7} is the set including the numbers 1,2,4,7.
– Intervals are also sets, e.g. all numbers between 3 and 4
– Set of daily prices of the stock of HP during the last year.

6
Set Theory
• Definition. A subset is any collection of elements in a set. An empty set
is a set which contains no elements (Φ).

• Examples
– (3,4] is a subset of R.
– The price of HP’s stock on the first day of each month is a subset
of the set of its daily prices during a year.
– The empty set is a subset of any set.

7
 Set Theory: Union
• Union. Let A,B, be two sets. Their union is the
set which includes all elements which are in A or A B
in B. It is denoted by A∪B.

• Example:
A: the set of daily prices of HP’s stock between
A∪B
January and April 2014.
B: the set of HP’s daily prices between March
and May 2014.
A∪B: the set of HP’s daily prices between
January and May 2014.

8
 Set Theory: Intersection
• Intersection. Let A,B, be two sets. Their
intersection is the set which includes all A B
elements which are in A and in B. It is denoted
by A∩B. A∩B

• Example:
A: the set of daily prices of HP’s stock between
January and April 2014.
B: the set of HP’s daily prices between March
and May 2014.
A∩B: the set of HP’s prices between March and
April.

9
Set Theory: Compliment

• Compliment. Let A be a subset of a set D. D

Its complement, Aഥ, consists of all of the elements
in D apart from those in A. A ഥ
A

• Example.
D is the set of a survey results.
A is the set of men’s answers.
ഥ is the set of the women’s answers.
A

• An important set is the set of all possible subsets

of a given set, A.

• Example.
If A={1,2,3}, the set of all subsets consists of:
{ Φ ,{1},{2},{3},{1,2},{1,3},{2,3},{1,2,3}}.
 Probability Theory: Definitions
• An experiment is any process or procedure for which more than one outcome is
possible.
• Examples:
– Tossing a coin: Possible outcome: head
– Rolling a die: Possible outcome: 6
– The price of IBM’s stock tomorrow: Possible outcome: $187.81

• Sample Space is the set of all possible outcomes of an experiment.

• Examples:
– Tossing a coin: {head, tail}
– Rolling a die: {1,2,3,4,5,6}
– The price of IBM’s stock tomorrow: [0, ∞)

• An event is a subset of a sample space.

• Examples:
– The price of IBM’s stock tomorrow is between $180 and $190.
11
Probability Theory: Definitions
• Probability is a numerical measure of the chance that an event will occur.

• Probability Measure is a function P from the set of all of the events of a sample
space Ω to [0,1], which satisfies for all disjoint events
A1, A2,…, An:
- P(A) ≥ 0
- P(A1 ∪ A2 ∪… ∪An)= P(A1) + P(A2) + … + P(An).
- P(Ω) = 1

• A probability measure is a function which assigns a probability to each

outcome in the sample space.

• Note. for any event A, 0≤P(A) ≤1:

- P(A)=0, if A never happens.
- P(A)=1, if A always happens.

12
Types of Probability

• Subjective Probability

• A-priori Classical Probability

• Empirical Classical Probability

13
Subjective Probability
• Subjective Probability: Probability values are assigned subjectively by people.

• Example:
- I have a feeling that the probability that IBM’s stock price tomorrow will be
higher than $190 is 80%.
- Someone less optimistic might estimate this probability by 50%.

• Advantage: Does not involve calculations

• Disadvantage: Subjective
- Bilgin (2012) showed that people judge losses to have higher probabilities
than gains
- “It is very likely that this will happen” → Estimate the probability

14
A-priori Classical Probability
• A-priori classical probability: the probability is determined according to a certain
theory.

• Example: Random walk model predicts that the probability that IBM’s stock price
will increase tomorrow is 50%, because according to the random walk model,
the probability of an increase in the price equals to the probability of decrease.

• Advantages:
– May be used to develop theories
– Helps calculations

• Disadvantages:
– One needs a trustworthy theory

15
 Empirical Classical Probability
• Empirical classical probability:
– Probability is based on observed data.
Number of cases in which the event occured
P(event) =
The total number of cases

• Example: One monitors a very large number of trading days, e.g. 1000.
and observes, 481 times that the price increased, then Prob.=481/1000=0.481.

• Advantages:
– Does not require a theory
– Based on real data

• Disadvantages:
– Requires collection of data (time, money,…)
– Based on the assumption that probability distributions do not change.
(Is this assumption true?)
16
Exercise
In a marketing survey, 1000 consumers were asked whether they intend to
buy a new product or not, and in a follow-up survey, the same group of
consumers were asked whether they have bought this product. The data is
shown below:

Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

1. What is the probability that a person plans to purchase this product?

2. What approach did we use here?

17
Solution
Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

1. The probability that a person plans to purchase this product within a year is:

number of people who planned to purchase

P(planned to purchase) =
numberof people in the sample
200 + 100
= = 0.3
1000

2. We used the empirical classical probability approach.

18
Probability Rules
1. ഥ)=1
P(A)+P(A
2. P(A ∩B)+P(A∩B ഥ)=P(A)
3. P(A ∪ B) = P(A) + P(B) - P(A ∩ B)

• Can you prove the above rules graphically?

A B

19
 Conditional Probability
• Definition: The probability of certain events depends on the
probability of other events.

• Example:
- Fruit and veges depend on the sun.
(http://www.nzherald.co.nz/lifestyle/news/article.cfm?c_id=6&objectid=11723008 )

• We denote the probability of A given B by P(A|B).

P(A  B)
• If P(B)≠0, then P(A | B) =
P(B)

20
Exercise
In a marketing survey, 1000 consumers were asked whether they intend to
buy a new product or not, and in a follow-up survey, the same group of people
were asked whether they have bought this product. The data is shown below:

Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

• What is the probability that a consumer purchased this product given that he or
she planned to purchase?
• What is the probability that a consumer did not purchase this product given
that he or she planned to purchase?
21
Solution
Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

P(Purchased and planned) 200/1000 200

P(Purchased | Planned) = = = = 0.6667
P(Planned) 300/1000 300

P(Did not purchase and planned) 100/1000 100

P(Did not purchase | Planned) = = = = 0.3333
P(Planned) 300/1000 300

22
 Bayes Theorem
P(A | B)P(B)
• Bayes' Theorem: P(B | A) =
P(A | B)P(B) + P(A | B )P( B)
• ഥ) and P(B
Purpose: If we know P(A|B),P(B),P(A|B ഥ), then we can calculate P(B|A).

• Example: A factory manager uses quality test to identify defective products. The
probability that a product is defective is 0.03. When the product is defective, the
probability that this quality test will give a positive result is 0.90. When the
product is not defective, the probability of a positive test result is 0.02. Suppose
that the quality test has given a positive result. What is the probability that the
product is actually defective?

• Solution: D: the product is defective, 𝐷: ഥ the product is not defective

ഥ: test is negative
T: test is positive, T
P(D)=0.03, P(D ഥ )=1-0.03=0.97, P(T|D)=0.90, P(T|𝐷 ഥ )=0.02
P(T | D)P(D) 0.9 * 0.03
P(D | T) = = = 0.582
P(T | D)P(D) + P(T | D)P( D) 0.9 * 0.03 + 0.02 * 0.97
• Exercise: Can you prove Bayes Theorem using conditional probability?
Bayes Theorem in Court
http://www.theguardian.com/law/2011/oct/02/formula-justice-bayes-
theorem-miscarriage

24
Independent Events
• Definition: Two events, A and B are called independent if P(A|B)=P(A).
Namely, the outcome of B does not affect the probability of occurrence
of A.

• Example: Two coin tosses are independent if the coins are fair.

• Note: The following necessary and sufficient conditions for

independent events are equivalent:
P(B | A) = P(B) P(A | B) = P(A) P(A  B) = P(A) * P(B)

• Exercise: Can you prove the above note?

26
Exercise
In a marketing survey, 1000 consumers were asked whether they intend to
buy a new product or not, and in a follow-up survey, the same group of
consumers were asked whether they have bought this product. The data is
shown below:
Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

• Given that a consumer planned to purchase, what is the probability that he/she
finally purchased the product?
• Given that a consumer planned not to purchase, what is the probability that
he/she finally purchased the product?
• Are consumers’ willingness to purchase and their actual purchase decisions
independent or not? Why?
27
Solution
Actually Purchased
Planned to Purchase Yes No Total

Yes 200 100 300

No 250 450 700
Total 450 550 1000

• Solution.
P(Purchased and planned) 200/1000 200
P(Purchased | Planned) = = = = 0.6667
P(Planned) 300/1000 300

P(Purchase and Planned Not) 250/1000 250

P(Purchased | Planned Not) = = = = 0.3571
P(Planned Not) 700/1000 700

P(Purchased | Planned) = 0.6667 P(Purchased) = 0.45

So the two events are not independent.

28
The Birthday Problem
To start, P(two people share birthday) = 1 - P(no people share birthday).

The probability of two people sharing birthday is difficult to get, but we can get the
probability of no people sharing birthday:
• One person
This person can have any birthday. P(no)=(365/365)=1

• Two persons
P(no)=(365/365) * (364/365)=99.73%

• Three persons
P(no)=((365/365) * (364/365) * (363/365)=99.18%
……
• n persons
P(no)=((365/365) * ((365-1)/365) * ((365-2)/365)*…*((365-n+1)/365)
=365! / ((365-n)! * 365^n)=[365*364*…*(365-n+1)]/ 365^n

P(two people share birthday)=1- P(no people share birthday)

=1-[365*364*…*(365-n+1)]/ 365^n
29
The Birthday Problem

Next, We will solve the Birthday Problem using Excel.

30
MSIN0154 Statistics for Business Research

Seminar 2

Probability
Reminder: Definitions
• An experiment is any process or procedure for which more than one
outcome is possible.
• Sample Space is the set of all possible outcomes of an experiment.
• An event is a subset of a sample space.

Example: Flip a coin twice

Sample space S = {HH, HT, TH, TT}
An event E: first flip results in a Head
{HH, HT}
Reminder: Definitions
• An experiment is any process or procedure for which more than one
outcome is possible.
• Sample Space is the set of all possible outcomes of an experiment.
• An event is a subset of a sample space.

• Probability measure is a function P from the set of all of the events of a

sample space Ω to [0,1], which satisfies for all disjoint events A,A1, A2,…,
An:

1. P(A) ≥ P(Φ) = 0
2. P(A1 ∪ A2 ∪… ∪An)= P(A1) + P(A2) + … + P(An).
3. P(Ω) = 1
Reminder

According to the empirical classical approach,

Reminder
A A
Probability rules
• Theorem. P(A)+P(A ഥ)=1.
• Theorem. P(A ∩B)+P(A∩B ഥ)=P(A).
• Theorem.
P(A ∪ B) = P(A) + P(B) - P(A ∩ B).
A∩ B A ∩B
• Definition. Conditional probability is the
probability of the occurrence of an event
A, given information about the occurrence
of an event B. P(A  B)
If P(B)≠0 then P(A | B) = P(B) A B

• Bayes' Theorem.
P(A | B)P(B) A ∩B
P(B | A) =
P(A | B)P(B) + P(A | B)P( B)
• Definition. Two events, A and B are called A B
independent if P(A|B)=P(A).
A∪B
Exercise 1
The ratings of Windows 8.1 on
Amazon.co.uk and Amazon.com
are given in the figures.

1.What is the probability that a

customer gave Windows 8.1
one star on Amazon.co.uk?

2. What is the probability that a rating is equal to or

more than 4 stars?
Exercise 1
The ratings of Windows 8.1 on
Amazon.co.uk and Amazon.com
are given in the figures.

3. Is the probability that

customers gave two
stars or less the same on
the UK and US websites?
Exercise 2

• If you flip a coin 3 times, what is the probability

that you will get Heads on at least one of the 3
flips?
Exercise 3 (Aczel, 2.11)
• According to an article in Fortune, institutional investors recently changed
the proportions of their portfolios toward public sector funds. The article
implies that 8% of investors studied invest in public sector funds and 6%
in corporate funds. Assume that 2% invest in both kinds. If an investor is
chosen at random, what is the probability that this investor has either
public or corporate funds?
Exercise 4
200 people were interviewed regarding their intention to buy your product and
whether or not they are financially able to do so. The following results were
obtained:
Intention to buy
Can afford Yes No Total
doing so
Yes 40 20 60
No 80 60 140
Total 120 80 200

What is the probability that:

1. a customer is financially able and has a desire to buy?
2. a customer has a desire to buy?
3. a customer has an intention to buy given that she has the financial ability to pay?
4. Are customers intentions about buying the product independent of their ability to
afford it?
Exercise 5
• An economics consulting firm has created a model to predict recessions.
The model predicts a recession with probability 80% when a recession is
indeed coming and with probability 10% when no recession is coming.
The unconditional probability of falling into a recession is 20%. If the
model predicts a recession, what is the probability that a recession will
indeed come?
 Exercise 6
In a company, 51% of the employees are full-time. It was found that
9.5% of the full-time employees in the company work more than 9
hours a day on average, whereas only 1.7% of the part-time
employees in the company work more than 9 hour a day. What is
the probability that if an employee works more than 9 hours a day
on average, then the employee is full-time?
Thank You!

Any Questions？

53