Austin Menger 1

STAT 1100Q-001

LECTURE 6
Probability
Motivation

Remember when we went over the statistics lifecycle:

We’ve discussed Steps 1 and 2 (Producing Data & EDA), and we’re now moving on to Step 3,
Probability. Step 4 is the last 3rd of the course.

Why do we use probability?

Example – Who Prefers Fall to Summer?

Group 1:

Group 2:

From this, we can see that there is variation due to the random sample.
Austin Menger 2
STAT 1100Q-001

Probability is not always intuitive

Example – Let’s Make a Deal

In the 1970’s show “Let’s Make a Deal,” a contestant was first given a choice of 3 doors to select
the one they believed the prize lay behind ( a goat was behind the 2 other). After the contestant
chose which door they thought the prize was behind, the host would then reveal one of the
other two doors that contained a goat instead of the prize. Finally, the host would ask if the
contestant would like to switch doors to the other remaining door.

The question is: Should the contestant switch? Are the odds of winning higher if he/she
switches?

Let’s take this step by step…

Step 1: One door has the prize, and the other two have goats. Whatever is behind your door is
what you take home. Suppose that you, the contestant, select door 1.

You

Right now at this step, the chance of the contestant selecting the door with the prize behind it
is
Austin Menger 3
STAT 1100Q-001

Step 2: Now, the Monty Hall (the host) opens one of the other two doors that he knows has a
goat behind it , say Door 3, and then asks if you’d like to switch your door selection from Door 1
to Door 2:

However, the odds of Door 2 containing the prize are

Comparing this to your odds at the start, you should

Terminology:

Probability → quantification of randomness and uncertainty

Random Experiment → An experiment that produces an outcome that cannot be determined

in advance (i.e. involves uncertainty)

Example 1: Toss a coin once

Example 2: Roll a die once
Example 3: A couple decides to have children until they have one boy and one
girl, but no more than 3 children.

Sample space, S, → the list of all possible outcomes of a random experiment

Austin Menger 4
STAT 1100Q-001

Event → a statement about the nature of the outcome that we’re actually going to get once the
experiment is conducted

- Sidenote – Subset →

Set A = {1,2,3,4}
Set B = {2,4}
Set C = {2,3,5}
Set D = {1,2,3,4}

B is a subset of A, but C isn’t a subset of A because 5 is not in Set A

Outcome → a possible result of the experiment

So, now let’s re-examine Example 1 – One Coin Toss

Looking at just one coin being flipped once, we have two possible events:

Event A:

Event B:

Now, let’s consider an example with more event possibilities, Example 2 – One Die Roll

We could have the following events that we are interested in:

Event A:

Event B:

Event C:

Event D:
Austin Menger 5
STAT 1100Q-001

Example 3 takes a bit more work. Here, we are looking at a couple that decides to have children
until they have one boy and one girl, but no more than 3 children.

Let’s first look at the sample space for this example:

So, some possible events we could look at are:

Event A:

Event B:

Note how an event is always a subset of the sample space.

How Can We Visualize Probability (and make our lives easier…)

Venn Diagrams are your best friend! ALWAYS DRAW ONE.

Let’s look at Example 2 again:

2 events we were interested in were:

Event B: roll an even number

Event C: roll a number less than or equal to 4

We can visualize this using a Venn Diagram:

B C

What does the overlap imply?

Austin Menger 6
STAT 1100Q-001

What if there was no overlap?

Now, how does this blend into probability? We can talk about the probability of an event
occurring:

P(B) →

P(C) →

The probability associated with an event describes the likelihood of that event occurring, i.e.
the outcome is one of the possible values dictated by the event.

If P(B) = 0, then

If P(B) = 1/2, then

If P(B) = 1, then

The Probability Lifecycle:

So, how do we calculate the probability that an event will occur?

Austin Menger 7
STAT 1100Q-001

I. Relative Frequency Approach

Let’s start simple. The probability that you’ll get a Head on a coin flip is 1/2. We all know that.
How is this calculated, though? It ties into relative frequency.

Remember that we defined relative frequency in Lecture 1 as the percent of all subjects that fall
into the category.

Formulaically, this was written as

In the case of probability, we can tweak the definition to write it as

Example – Coin Flips

Trial 1: ___ , ___ , ___ , ___ , ___ , ___ r.f. of Heads =

Trial 2: ___ , ___ , ___ , ___ , ___ , ___ r.f. of Heads =

In theory, we could flip the coin an infinite number of times. The relative frequency of heads
will be half the number of flips.

In practice, however, we can’t flip a coin infinitely many times. So, we must instead choose
what is considered a very large number, 40,000 times for example like a couple of Berkeley
undergrads did (https://www.stat.berkeley.edu/~aldous/Real-World/coin_tosses.html) . Out of
40,000, theoretically 20,000 should be heads.

However, the relative frequency approach only provides an estimate for P(A). As you can
imagine, if we tossed a coin 40,000 times (1 hour a day for an entire semester according to the
Berkeley undergrads…) we wouldn’t get exactly 20,000 heads. It would vary due to variability,
but the average would be 20,000 if you repeated the experiment a number of times.
Austin Menger 8
STAT 1100Q-001

Note: We control how good this estimate is by the number of times we repeat the random
experiment (increase from 40,000 to 100,000). The more repetitions performed, the closer
this relative frequency gets to the true probability P(Head).

II. Equally Likely Outcomes

If all the outcomes in the sample space S are equally likely, then we can find P(A) using
the following formula:

Example – Roll a Die

What is the probability that when you roll a die, the number will be less than 3?

Note: It’s NOT always the case that all outcomes are equally likely.

Example - Gender of babies

Re-examining Example 3 from above, we said the sample space for this random
experiment was

Not each outcome has the same probability!

Note: if looking at pairs, order may or may not matter depending on the context.
Austin Menger 9
STAT 1100Q-001

Example – Job Selection

4 people have applied to work as a Data Scientist at Facebook, 2 males (M1, M2) and 2
females (F1, F2). There are only 2 positions, but their HR department has told those
involved with the hiring process that they can’t hire 2 people of the same sex. Assuming
all candidates are equally qualified and thus equally likely to secure the opening, what is
the probability that the HR department’s mandate would be violated?

III. Probability Principles (Rules)

1. The Complement Principle

Let’s look at the Complement Principle visually first

Formulaically:

Note: this rule is especially helpful when finding P(A) is quite complicated or laborious.

2. The Intersection of 2 Events: (which outcomes are in both events)

Austin Menger 10
STAT 1100Q-001

Visually:

Notation:

Note: 2 events are called disjoint if they don’t have any outcomes in common

Visually:

Mathematically:

3. The Union of 2 Events (all the outcomes that are in A or B (or both) → all the
outcomes that are in at least one of the 2 events)

Visually:
Austin Menger 11
STAT 1100Q-001

Notation:

So, let’s summarize by labelling each part of the Venn Diagram:

Example – Throwing a Die

Suppose we are only throwing 1 die once. Event A is the event that the roll is less than or equal
to 2. Event B is that the roll is odd.

a. Are all the possible outcomes equally likely? What is the sample space?

b. What is P(A)? P(B)?

c. What is the 𝑃(𝐴 ∩ 𝐵)?

d. What is the 𝑃(𝐴 ∪ 𝐵)?

Austin Menger 12
STAT 1100Q-001

e. What is the 𝑃(𝐴𝑐 )? What does the event 𝐴𝑐 actually mean in context?

Example – Symptoms
Suppose that we are looking at a drug to treat high blood pressure. We know that there is a
14% chance of headache, an 11% chance of indigestion, and a 6% chance of getting both.

a. Write down the 3 pieces of information given in terms of probabilities of H

and I

b. Summarize the info using a Venn Diagram

c. Are the two events H and I disjoint?

d. What is the probability of experiencing only Indigestion?

e. What is the probability of experiencing only headache?

Austin Menger 13
STAT 1100Q-001

f. What is the probability of experiencing neither?

Another way to dissect and summarize this information is to use something called a
probability table:

A fundamental principle used to fill in the probability table is that when there are only 2 events
A,B:

To put this in context from our previous example, 14% of all people who use the drug get
headaches. Of that 14%, 6% also get indigestion and the other 8% experience only headaches.

The information needed to be able to completely fill in a probability table is as follows:

Austin Menger 14
STAT 1100Q-001

Example – Package Delivery

As he does with most things, Austin waited until the day before his taxes were due to mail them
in. It’s vital that the documents reach the destination within one day. To maximize the chances
of on-time delivery, two copies of the document are sent using 2 services, Fed-Ex (F) and
Amazon’s new delivery service Prime Delivery. We know that there is a 92% chance of on-time
delivery by Fed-Ex and a 94% chance of on-time delivery by Prime Delivery. We also know there
is a 2% chance of on-time delivery by neither Fed-Ex or Prime Delivery.

a) Write down the 3 pieces of given information in terms of probabilities involving event F
and event P

b) Summarize the info using a probability table

c) What is the probability that only Prime Delivery is on time? (put this in symbols and
then answer using the table)
Austin Menger 15
STAT 1100Q-001

d) What is the probability that both services render on-time delivery? (put this in symbols
and then answer using the table)

e) What is the probability that only Fed-Ex renders on-time delivery? (put this in symbols
and then answer using the table)

4. The Addition Principle (for finding 𝑷(𝑨 ∪ 𝑩))

First, what is 𝑃(𝐴 ∪ 𝐵) in words?

Now, visually what is 𝑃(𝐴 ∪ 𝐵) using a Venn Diagram?

So, the formula is

What happens when the 2 events are disjoint?

First off, what does it mean to be disjoint?

Austin Menger 16
STAT 1100Q-001

So, to find 𝑃(𝐴 ∪ 𝐵) in the case where the 2 events are disjoint, we have

Example – Car Ownership

For this example, we are looking at the town of Greenwich, CT and studying the types of
cars families own. We know that 15% own a Jeep Wrangler and 21% own an Audi. 3% of
families own both a Jeep Wrangler and an Audi (must be nice…).

a. Write down the 3 pieces of info given in terms of the probabilities of events J
and A

b. Construct a probability table

c. What is the probability that a family living in Greenwich, CT owns a Jeep

Wrangler or an Audi?

Note: There is another way to word the same question: What is the probability that a family
living in Greenwich, CT owns at least one of the two types of cars?
Austin Menger 17
STAT 1100Q-001

d. Now, what is the probability that a family living in Greenwich, CT owns exactly
one of the two car types mentioned?

Visually:

Formulaically: