ECO 231: Intermediate Economic Statistics I

(2 units)​

Course Outline

Probability theory and probability distributions [including discrete and

continuous probability distributions]​

Probability sampling from tables of random numbers​

Applications of probability theory to economics​

Basic probability theorems​

Types and distributions of random variables: Discrete and continuous

random variables​

Exponential density and gamma density​

Bernoulli, Binomial, Poisson, Geometric, Hyper-geometric, Normal

distributions​

Generation of statistical events from set-theory​

Combinatorial Analysis​

Expectations and moments of random variables

NB: ​

I. This table of content above is based on the department’s standard

course module and not really directly gotten from the lecturer.

Nonetheless, it quite resonates with the note and might be helpful for

further personal study.​

II. This note was made was made using Ebuka Julien’s handwritten

lecture note and transcribed to textual format by Ugwu Martins who

also made some adjustments for enhanced readability and

comprehensibility

III. More adjustments will be made to this note as there could be some

corrections to make. For any observation, question, correction or

suggestion, DM: Ugwu Martins

ECO 231 - 21/03/2025
Intermediate Economic Statistics

Partial Linear Probability

Tossing a coin is called a random trial. The different possible outcomes of random trials are
called Basic Outcomes. The state of possible outcomes in basic outcomes is called sample
space.

Example of Sample Space:​

𝑆 = {1, 2, 3, 4, 5}

Therefore, the sample space for tossing a coin is:​

𝑆 = {𝐻, 𝑇}

The basic outcomes are called sample points or sample events, and they are mutually
exclusive & mutually exhaustive → at least one event must happen.

Example:​
𝑆 = {𝐻𝐻, 𝐻𝑇, 𝑇𝐻, 𝑇𝑇}

Recall

A subset of a sample space is called an event.​

𝐸 = {𝐻𝐻, 𝑇𝑇}

Examples

1.​ Consider an event for at least one Head:​

Answer:​
𝐸1​ = {𝐻𝐻, 𝐻𝑇, 𝑇𝐻}​

2.​ Consider an event that both faces are tails:​

𝐸2​ = {𝑇𝑇}​

Intersection & Complementary Sets

●​ Intersection: A set of outcomes or sample points not contained in an event is called a

complementary set and is denoted as C.​
 ●​ The complementary event of 𝐸1 is:​
𝐶
𝐸1​ = {𝑇𝑇}
●​ The complementary event of 𝐸2is:​
𝐶
𝐸2​ = {𝑇𝑇}

The set that belongs to both 𝐸1 and 𝐸2 is called intersection.

●​ Intersection of 𝐸1 and 𝐸2 is a null set:​

∅
●​ The intersection of an event and its complement is always a null set.​
Here is the neatly typed content from your notebook:

Univariate, Bivariate, Multivariate Sample Space

H T

H HH HT

T TH TT

Probability

We will now go beyond random trials where actions take place. Probability is a likelihood
measure that an event will occur. It is attached to sample points, providing the probability of all
basic outcomes, ensuring they sum up to one.

Probability assignments must follow certain rules:

1.​ The probability of any basic outcome or event consisting of basic outcomes must be
between 0 and 1:​
0≤𝑃(𝐸)≤1
2.​ 𝑃(𝐸) = 1 or 0​

○​ If 𝑃(𝐸) = 1 or 𝑃(𝑂) = 1, it means that a certain outcome is bound to occur.​

3.​ 𝑃(𝐸𝑖)≤𝑃(𝐸𝑗​)​

○​ The probability of 𝑃(𝐸𝑖) is a subset of 𝑃(𝐸𝑗​).​

 4.​ If 𝐸1 and 𝐸𝑗 are mutually exclusive events of a sample space, then their intersection is
null (∅) and their union is 𝐸𝑖 ∪ 𝐸𝑗.​

𝑛
5.​ ∑ 𝑃(𝐸𝑖​) = 1
𝑖=1
○​ This ensures that the total probability across all possible outcomes is 1.

Odd Ratios
Probability can be expressed as odd ratios. If the probability of an event 𝐸𝑖 is 𝑥, then the odd
ratio of that event occurring is:

𝑥
1−𝑥

For example, if the probability that you get into an accident on your way home from work is 0.2,
then the odd ratio for you getting into an accident is:

0.2 0.2
1−0.2
= 0.8
= 0. 25

If the odd ratio is b over a given interval, then the probability of an event occurring is:

𝑎
𝑎+𝑏

Example

If the odd ratio that a car would break down is 1 to 10, then the probability that the car will
break down is:

1 1
1+10
= 11

Conditional Probability
The probability of an event occurring given another event is called conditional probability.
The probability of event A occurring given event B is:

𝑃(𝐴∩𝐵)
𝑃(𝐴∣𝐵) = 𝑃(𝐵)

Example

Suppose we are looking at the behavior of four stocks listed on the Nigerian Stock Exchange.
 Increase 𝐴1 No Change 𝐴2​ Decrease 𝐴3 Sum

𝐵1 0.20 0.05 0.05 0.30

𝐵2 0.15 0.10 0.15 0.40

𝐵3 0.05 0.05 0.20 0.30

Sum 0.40 0.20 0.40 1.00

Question

What is the conditional probability that the price of Stock A will increase given that Stock B
has increased?​
Here is the transcription of the left side of the note:

Solution of 𝑃(𝐴∣𝐵):
0.2−0.66
𝑃(𝐴∣𝐵) = 0.3
= 0. 3​
𝑃(𝐴∩𝐵)⇒ particular 𝑃(𝐵)

Interpretation:

0.66, which is greater than the unconditional probability of an increase in the price of stock A.

This suggests that the probability that the price of stock A is greater if the price of stock B also
increases.

Assignment:

Consider the probability that the price of stock A will fall conditional on a fall of stock B.

Solution:

𝑃(𝐴∩𝐵) = 0. 05​
𝑃(𝐵) = 0. 3​
0.05
𝑃(𝐴∣𝐵) = 0.3
= 0. 16​
0.2−0.66
𝑃(𝐴) = 0.3
Probability Theory

FO 231 – 28th February 2025

Probability theory is a branch of mathematics that gives analysis of random events. It provides a
foundation for statistical inference & decision-making under uncertainty.

Axioms of Probability:

1.​ Non-negativity: Probability values must be 𝑃(𝐴)≥0.​

2.​ Normalization: The sum of all probabilities must equal 1.​

3.​ Additivity: If two events are mutually exclusive, the probability of their union is:​
𝑃(𝐸1 ∪ 𝐸2​) = 𝑃(𝐸1​) + 𝑃(𝐸2​)

Probability Distribution

A probability distribution describes how probabilities are assigned to different outcomes of a

random variable.

Random Variable Types:

●​ Discrete Variable​

●​ Continuous Variable

Discrete Probability Distribution

A discrete random variable takes a countable number of values.

For a discrete random variable XX, a probability mass function (PMF) assigns probability to
different outcomes.

Features of PMF

1.​ Non-Negativity:​
𝑃(𝑋 = 𝑥)≥0 for all 𝑥.
2.​ Summation Property:​

∑ 𝑃(𝑋 = 𝑥) = 1.
𝑥∈𝑋

Some Common Discrete Distributions

1. Bernoulli Distribution
This models a single trial with two possible outcomes: success (1) or failure (0).

The PMF for this distribution is:

𝑥 1−𝑥
𝑃(𝑋 = 𝑥) = 𝑝 (1 − 𝑝)

where xx is an element of {0,1}, i.e.,

𝑋 ∈ {0, 1}.

2. Binomial Distribution

This models the number of successes in nn independent Bernoulli trials.

The PMF is given by:

3. Poisson Distribution

This distribution models the number of events occurring in a fixed interval of time or space.

The PMF is:

𝑘 −λ
λ𝑒
𝑃(𝑋 = 𝑘) = 𝑘!

Continuous Probability Distribution

For a continuous random variable 𝑋, a probability density function (PDF) assigns

probabilities over an interval of values.

Properties of PDF

1.​ Non-Negativity:​
𝑓(𝑥) ≥ 0 for all 𝑥.
2.​ Total Probability Rule:​

3.​ Probability of an Interval [a,b][a, b]:​

Expectation of a Discrete Random Variable

If 𝑋 is a discrete random variable with PMF 𝑃(𝑋 = 𝑥), its expected value (expectation) is
given by:

𝐸(𝑋) = ∑𝑥𝑃(𝑋 = 𝑥)

Example: Suppose that a random variable has only four values

𝑋 𝑃(𝑋) 𝑋𝑃(𝑋)

0 0.4 0

1 0.15 0.15

2 0.3 0.6

4 0.15 0.6

Totall 1.0 1.35

𝐸(𝑋) = ∑𝑋𝑃(𝑋) = 1. 35

Expectation of a Continuous Random Variable

If a random variable 𝑋 has a continuous distribution with PDF 𝑓(𝑥), then the expectation of 𝑋
is defined as:

Let the PDF of a continuous variable be:

2
𝑓(𝑥) = 3𝑥 , 0 ≤ 𝑥 ≤ 1

Solution:

​
4 4
3(1) 3(0) 3
= 4
− 4
= 4
Bernoulli Distribution Explained

For a Bernoulli distribution, the probability function is:

𝑥 1−𝑥
𝑃(𝑋 = 𝑥) = 𝑝 (1 − 𝑝)

If 𝑋 follows a Bernoulli distribution with parameter 𝑝, the expectation of 𝑋 is:

𝐸(𝑋) = ∑𝑥𝑃(𝑋 = 𝑥)

Substituting values:

1×𝑝+0×𝑞=𝑝+0=𝑝

Thus,

𝐸(𝑋) = 𝑝

Variance of a Bernoulli Distribution

2 2
𝑉𝑎𝑟(𝑋) = 𝐸(𝑋 ) − (𝐸(𝑋))

Expanding:

2
𝑉𝑎𝑟(𝑋) = 𝑝 − 𝑝 = 𝑝(1 − 𝑝) = 𝑝𝑞
ECO 231 - 9th March 2025

Binomial & Bernoulli Distribution

The proportion of defective items in a manufacturing plant is 25%.​

In a random sample of 8 items, what is the probability that there will be no defective item?

Solution:

The binomial formula is given as:

where:

●​ 𝑛 = 8​

●​ 𝑥 = 0​

●​ 𝑝 = 0. 25​

●​ 𝑞 = 0. 75​

​
= 1 × 0. 1001 = 0. 1001

(i) Probability of finding no more than one defective item

(ii) Probability of finding not less than two defective items

Using the binomial probability formula:

​
8−1
= 0. 25×0. 75 ​
= 0. 25×0. 1334 = 0. 03335​
+ 0. 267 + 0. 1001 = 0. 3691

For 𝑃(𝑋 ≥ 2):

1 − 𝑃(𝑋 ≤ 1) = 1 − 0. 3608 = 0. 6392

Poisson Distribution

λ represents the expected number of occurrences per unit of time/space.

Question:

Let calls going through the MTN exchange be distributed according to a Poisson law at the rate
of 5 per minute.

I.​ What is the probability that:​

A.​ There will be no call in a minute?​

B.​ Exactly two calls per minute?​

C.​ More than two calls in a minute?​

II.​ What is the expected number of calls in 7 minutes?​

Solution

Let 𝑋 represent the number of calls per minute. Using the Poisson formula:
𝑥 −λ
λ𝑒
𝑃(𝑋 = 𝑥) = 𝑥!

where λ = 5.

(i) Probability of no call in a minute

For 𝑥 = 0:
0 −5 −5
5𝑒 𝑒
𝑃(𝑋 = 0) = 0!
= 1

(ii) Probability of exactly two calls in a minute

For 𝑥 = 2:
2 −5 −5
5𝑒 25𝑒
𝑃(𝑋 = 2) = 2!
= 2
= 0. 08422

(iii) Probability of more than two calls in a minute

𝑃(𝑋 > 2) = 1 − [𝑃(𝑋 = 0) + 𝑃(𝑋 = 1) + 𝑃(𝑋 = 2)]​
= 1 − 0. 111172

Expected Value in Poisson Distribution

●​ The expected value in a Poisson distribution is given by 𝐸(𝑋) = λ.​

●​ PMF (Probability Mass Function) → Discrete Random Variable​

●​ PDF (Probability Density Function) → Continuous Random Variable

Normal Distribution

●​ One of the most commonly observed continuous random variables.​

●​ It has a bell-shaped curve, as shown below.​

●​ It is known as the normal random variable, and its distribution is called the normal
distribution.​

●​ The two sides of a normal distribution are equal, which means that it is symmetric.

Probability Density Function (PDF) of Normal Distribution

2
(𝑥−µ)
−
1 2σ
2

𝑃𝐷𝐹 = 𝑓(𝑥) = 𝑒
σ 2π

Where:

●​ μ is the mean of the random variable 𝑋.​

●​ σ is the standard deviation.​

●​ π = 3. 146​

●​ 𝑒 = 2. 718​

Finding Probability using Normal Distribution

●​ 𝑃(𝑋 ≤ 𝑥) is obtained from the table of normal distribution.

Properties of Normal Distribution

 ●​ If 𝑋 is a normal random variable with mean μ\mu and standard deviation σ\sigma,
then the normal random variable 𝑍 is given by:​

𝑋−μ
𝑍= σ

●​ The value 𝑍 describes the number of standard deviations between 𝑋 and µ.

Application of Normal Distribution

Problem Statement

Suppose that 𝑋 has a normal distribution with mean 5 and variance 16.​
Find the probability that:

1.​ 𝑋 lies between 2 and 12.​

2.​ 𝑋 is greater than 15.​

3.​ 𝑋 is less than 12.​

4.​ 𝑋 lies between 8 and 15.

Solution

(i) Probability that 𝑋 lies between 2 and 12

Using standardization:

2−5 12−5
𝑃(2 ≤ 𝑋 ≤ 12) = 𝑃( 4
≤𝑍≤ 4
)​
= 𝑃(− 0. 75 ≤ 𝑍 ≤ 1. 75)

Using normal tables:

𝐴(1. 75) − 𝐴(− 0. 75) = 0. 4599 + 0. 2237 = 0. 9353

ECO 231 – 14/03/2025
Finding the Probability that a Standard Normal Random Variable (SNR)
Exceeds 1.96 in Absolute Value

𝑃(∣𝑍∣ > 1. 96)

The probability is the area of the figure. Note: The total shaded area is the sum of the two
areas.

0. 96 = 0. 4750​
0. 5 − 0. 4750 = 0. 025 ⇒ 𝐴1​
𝐴2​ = 0. 025​
𝐴1​ + 𝐴2​ = 0. 025 + 0. 025 = 0. 05

The same value in one side of the normal table is also the same for the other side.

Thus,

𝑃(∣𝑍∣ > 1. 96) = 𝐴1​ + 𝐴2​ = 0. 05

Application of Normal Topic in Process Control

Problem Statement

Suppose a paint manufacturing firm has a daily production that is normally distributed with:

●​ Mean (μ) = 100k gallons​

●​ Standard deviation (σ) = 10k gallons​

The management would like to create an incentive bonus program where production
exceeds the 90th percentile.​
This is expected to make workers more productive over time.

Question:

What production level should management pay incentive bonuses for?

Hint: The 90th percentile of a standard normal distribution (SND) is 1.28.

Solution

Let 𝑋0 be the required production level:

Since 90% of daily values fall below 𝑋0, we write:

𝑃(𝑋 ≤ 𝑋0​) = 0. 90

Using the Z-score transformation:

𝑋0​−μ
𝑃(𝑍 ≤ σ
​) = 0. 90

𝑋0​−100,000
𝑃(𝑍 ≤ 10,000
​) = 0. 90

Since the 90th percentile of a standard normal distribution (SND) is 1.28, we set:

𝑃(𝑍 ≤ 1. 28) = 0. 90

Consequently, we know that the production level X0X_0 at which the incentive bonus is paid
corresponds to the Z-score = 1.28.

Setting up the equation:

𝑋0−100,000
10,000
​ = 1. 28

Cross multiply:

𝑋0​ − 100, 000 = 1. 28 × 10, 000​

𝑋0​ − 100, 000 = 12, 800​
𝑋0​ = 100, 000 + 12, 800​
𝑋0​ = 112, 800

Thus, the 90th percentile of the production level is 112,800.​

So, management should pay an incentive bonus when a day's production exceeds this
level.

Central Limit Theorem (CLT)

The Central Limit Theorem (CLT) is a fundamental concept in applied and theoretical
probability.
Theory Statement

Let 𝑋1​, 𝑋2​, 𝑋3​,..., 𝑋𝑛​ be a sample of nn independent random variables with:

●​ Expectation: 𝐸(𝑋𝑖​) = μ​

2
●​ Variance: 𝑉(𝑋𝑖​) = σ ​

Then, the sample mean 𝑋 follows a normal distribution for large 𝑛.

Example:

A sample of 100 tickets from an observation has:

●​ Mean = 80​

●​ Variance = 25​

Find the probability that the sum is greater than 8120.

Here's a neatly typed version of your notes:

ECO 231

Date: 26/08/2025

EXPONENTIAL PROBABILITY DISTRIBUTION

The amount of time the occurrence of an event takes must be monitored. The failure of
RV (random variable) can be described as an Exponential Probability Distribution (EPD).
It is sometimes called a lifetime distribution.

The Probability Density Function (PDF) is given as:

−λ𝑥
𝑓(𝑥) = λ𝑒 , 𝑥 > 0​
−λ𝑥
𝐹(𝑥) = 1 − 𝑒 ,𝑥 > 0

●​ Mean:​
1
𝐸(𝑋) = λ
●​ Standard Deviation:​
1
σ= λ
●​ Variance:​
1
𝑉𝑎𝑟(𝑋) = 2
λ

Problem Statement & Solution

A microwave oven manufacturer is trying to determine the length of the warranty period
that should be attached to its magnetron tube. Marketing testing has shown that the
length of life (years) of a magnetron tube follows an EPD with:

λ = 0. 16

Find the Mean, Standard Deviation, and Variance of 𝑋:

1
1.​ Since 𝐸(𝑋) = λ
, this implies that:​
1
𝐸(𝑋) = 0.16
= 6. 25
1
2.​ Since 𝑆𝐷 = λ
, then:​
1
σ= 0.16
​ = 6. 25
3.​ Variance:​
1 1
𝑉𝑎𝑟(𝑋) = 2 ​= 2 ​ = 39. 02
λ (0.16)
(a) Find the fraction of tubes to be replaced before 5 years usually expires.

We need to find the probability:

−λ𝑥
𝑃(𝑋 > 5) = 𝑒

Substituting the given values:

−0.16(5)
𝑃(𝑋 > 5) = 𝑒 ​
−0.8
=𝑒 ​
= 0. 449

Thus, the fraction of tubes lasting more than 5 years is 0.449 (or 44.9%).

(b) Suppose the warranty period of 5 years is attached to the magnetron

tube. What fraction of the tubes fail within the warranty period?

𝑃(𝑋≤5) = 1 − 𝑃(𝑋 > 5)​

= 1 − 0. 449​
= 0. 55

Thus, 55% of the magnetron tubes fail and have to be replaced within the warranty
period(5 years).