Chapter 4-Probability

Judgement
A Course in Behavioral Economics by Erik Angner
Instructor: Tolga Yuret
Probability

• People’s intiutive judgements about probability of events fail

miserably.

• Here are two examples that are used in the experiments.

Probability Judgement-Example
• You are visiting your new neighbor, Mrs Jones. Mrs Jones tells you
that she has two children, who are playing in their room. Now, Mrs
Jones tells you that at least one of the children is a girl. What is the
probability that the other child is a girl too?
Probability Judgement-Example
Linda is 31 years old, single, outspoken and very bright. She majored in
philosophy. As a student, she was deeply concerned with issues of
discrimination and social justice and also participated in anti-nuclear
demonstrations.
(a) What is the probability that Linda is a bank teller?
(b) What is the probability that Linda is a bank teller and a feminist?
Definitions
• The outcome space is the set of all possible individual outcomes.

• An outcome is a subset of the outcome space.

• e.g. {1} for one, {6} for six, {1, 2, 3} for a number less than or equal to
three, and {2, 4, 6} for even numbers
Definitions
• The probability function is a function Pr (·) that assigns a real number
to each outcome. The probability of an outcome A is the number Pr
(A) assigned to A by the probability function Pr (·).

• e.g. Pr ({2, 4, 6}) (probability of an even toss)

• The probability of any outcome A is a number between 0 and 1

inclusive; that is, 0 ≤Pr (A) ≤ 1.
The equiprobability rule
• If there are n equally probable individual outcomes {A1, A2, … , An},
then the probability of any one individual outcome Ai is 1/n; that is, Pr
(Ai) = 1/n.

• e.g. “If you have something that can happen, and something that
won’t necessarily happen, it’s going to either happen or it’s going to
not happen, and so the best guess is one in two.” Why is this not a
correct application of the equiprobability rule?
The equiprobability rule

• Now let’s go back to the very first example.

• the outcome space is {GG, GB, BG, BB}

• Once you learn that at least one of the children is a girl, the outcome
space has been reduced to {GG, GB, BG}. Because these three
outcomes are equally likely, you can apply the equiprobability rule to
find that the probability that the other child is a girl equals Pr (GG) =
1/3
The equiprobability rule
• Instead of telling you that at least one of the children is a girl, Mrs
Jones tells you that her oldest child is a girl. Now, what is the
probability that the other child is also a girl?

• Homework: Solve Exercise 2.11

Example: Three-card swindle
• The deck consists of only three cards. The first card is white on both
sides. The second card is red on both sides. The third card is white on
one side and red on the other. The deck is put on the table in such a
way that you can see the visible face of the top card only.
(a) What is the outcome space? Write “W/R” to denote the outcome
where the visible side of the top card is white and the other side is red,
and so on.
(b) After shuffling, the visible side of the top card is white. What is the
new outcome space?
(c) Given that the visible side of the top card is white, what is the
probability that the other side of the top card is red?
Example: Three-card swindle
(a) {W/W, W/W, R/R, R/R, W/R, R/W}.
(b) {W/W, W/W, W/R}.
(c) 1/3.

• You can earn money without cheating! Many people believe that the
probability that other side is Red is ½!

• Homework. Do exercise 4.13

Example: The Monty Hall Problem
• You are on a game show. The host gives you the choice of three doors, all
of which are closed. Behind one door there is a car; behind the others are
goats. Here is what will happen. First, you will point to a door. Next, the
host, who knows what is behind each door and who is doing his best to
make sure you do not get the car, will open one of the other two doors
(which will have a goat).

• Finally, you can choose to open either one of the remaining two closed
doors; that is, you can keep pointing to the same door, or you can switch. If
you do not switch, what is the probability of finding the car?
Definition
• If two outcomes A and B are mutually exclusive (see below), then the
probability of A or B equals the probability of A plus the probability of
B; that is, Pr(A ∨ B)=Pr(A)+Pr(B).

• e.g. Suppose that you want to know the probability of rolling a one or
a two when you roll a fair six-sided die. The or rule tells you that the
answer is
Pr(1 ∨ 2)=Pr(1)+Pr(2)=1/6+1/6=1/3.
Axioms.
• The probability of the entire outcome space is equal to one.

• The probability that some outcome A will not ocur is equal to one
minus the probability that it does. That is, Pr(¬A)=1-Pr(A).
Axioms
• Two outcomes A and B are independent just in case the fact that one
occurs does not affect the probability that the other one does.

• If two outcomes A and B are independent, then the probability of A

and B equals the probability of A multiplied by the probability of B;
that is, Pr (A & B) = Pr (A) * Pr (B).
Exercise 4.22
• What pairs of outcomes are independent? More than one answer
may be correct.
(a) You sleep late; you are late for class.
(b) You are a remarkable student; you get a good job after graduation.
(c) You write proper thank-you notes; you get invited back.
(d) The first time you flip a silver dollar you get heads; the second time
you flip a silver dollar you get tails.
(e) General Electric stock goes up; General Motors stock goes up.
Lucky in cards, unlucky in love
• According to a well-known saying: “Lucky in cards, unlucky in love.” Is
this to say that luck in cards and luck in love are independent or not
independent?

• Homework: Do exercise 4.24 and 4.25

Solution to bank teller example.
• Let F mean that Linda is a feminist and B that Linda is a bank teller.
Then the probability that she is both a feminist and a bank teller is
• Pr (B&F) = Pr (B) * Pr (F). (In order to apply the and rule here, I am
assuming that the outcomes are independent; the general result,
however, holds even if they are not.)
• Because Pr (F) ≤ 1, we know that
• Pr (B) * Pr (F) ≤ Pr (B).
• This mistake is referred to as the conjunction fallacy
• Homework: Do exercises 4.26 and 4.27
Conditional Probability

• About one in five Americans believe that “the most practical way for
them to accumulate several hundred thousand dollars is to win the
lottery.” The poor, least educated, and oldest are particularly likely to
think of the lottery as a smart way to get rich. So it might be useful to
ask just how likely or unlikely it is to win common lotteries.

• So let’s compute the probability of lotteries...

Conditional Probability
• In Lotto 6/49, players circle 6 numbers out of 49.

• You win the grand prize if all 6 are correct. What is the probability
that you win the Lotto 6/49 any one time you play?

• Note: The idea of anchoring and adjustment explains why people

believe that they have a good chance of winning these lotteries.
The answer is:
Conditional Probability
• Suppose that H means “The patient has a headache” and T means
“The patient has a brain tumor.”
(a) How do you interpret the two conditional probabilities Pr (H | T)
and Pr (T | H)?
(b) Are the two numbers more or less the same?
Conditional Probability
• If A and B are two outcomes, Pr (A | B) = Pr (A & B)/ Pr (B).

• Proposition 4.36 Pr(A | B) * Pr(B) = Pr(B | A) * Pr(A).

Proof. Just apply the above definition to LHS and RHS

Total Probability Rule
• Suppose that you are running a frisbee factory and that you want to
know the probability that one of your frisbees is defective. You have
two machines producing frisbees: a new one (B) producing 800
frisbees per day and an old one (¬ B) producing 200 frisbees per day.

• Thus, the probability that a randomly selected frisbee from your

factory was produced by machine B is Pr (B) = 800/(800 + 200) = 0.8;
the probability that it was produced by machine ¬B is
• Pr (¬ B) = 1 - Pr (B) = 0.2
Total Probability Rule
• Among the frisbees produced by the new machine, one percent are
defective (D); among those produced by the old one, two percent are
defective. That is:

• Pr (D | B) = 0.01

• Pr (D | ¬ B)=0.02
Total Probability Rule
• What is the probability that the frisbee defective? We can use the
concept of total probability. Total probability of D can be written by
adding up conditional probabilities as follows. (note: easy to prove
that this is so, just use the definitions.)

Homework: Do exercises 4.40 and 4.41

Bayes Rule
• Suppose you pick up one of the frisbees produced in your factory and
find that it is defective.
• What is the probability that the defective frisbee was produced by the
new machine?
• That is what is Pr (B | D)?
• Pr (D&B) = Pr (D | B) * Pr (B) = 0.01 * 0.8 =0.008
• Pr (D &¬ B) = Pr (D | ¬ B) * Pr (¬ B) = 0.02 * 0.2 = 0.004
• Pr (B | D) = 0.008/0.012 = 2/3.
• Why is this the case that defective comes more probably from the
new factory?
Bayes Rule

Homework: Do exercises 4.43 and 4.44

Bayes Rule
You are considering asking L out for a date, but you are a little worried that L may
already have started dating somebody else. The probability that L is dating
somebody else, you would say, is 1/4. If L is dating somebody else, he/she is
unlikely to accept your offer to go on a date: in fact, you think the probability is only
1/6. If L is not dating somebody else, though, you think the probability is much
better: about 2/3.
(a) What is the probability that L is dating somebody else and will accept your
offer to go on a date anyway?
(b) What is the probability that L is not dating somebody else and will accept your
offer to go on a date?
(c) What is the probability that L will accept your offer to go on a date?
(d) Suppose L accepts your offer to go on a date. What is the probability that L is
dating somebody else, given that L agreed to go on a date?
Answers (a) 1/4 *1/6 = 1/24. (b) 3/4 * 2/3 = 6/12 = 12/24. (c) 1/24+12/24 = 13/24.
(d) (1/24)/(13/24) = 1/13. Good news!
Bayesian Updating
• We update beliefs in light of new evidence all the time.

• In science, we update our assessment about the plausibility of a

hypothesis or theory in light of evidence, which may come from
experiments, field studies, or other sources.
Bayesian Updating
• Let H stand for the hypothesis and E for the evidence.
• The probability of H, Pr (H), is called the prior probability: it is the
probability that H is true before you learn whether E is true.
• The probability of H given E, Pr (H | E), is called the posterior
probability: it is the probability that H obtains given that the evidence
E is true.
• The question is what the posterior probability should be.
Bayesian Updating
• Suppose that John and Wes are arguing about whether a coin brought to class by a student has
two heads or whether it is fair. Let H be the hypothesis that the coin has two heads, so that ¬ H
means that the coin is fair. Let us consider John first. He thinks the coin is unlikely to have two
heads: his prior probability, Pr (H), is only 0.01. Now suppose the student flips the coin, and that it
comes up heads. Let E mean “The coin comes up heads.” The problem is this: What probability
should John assign to H given that E is true?
Bayesian Updating
• Now, suppose you toss the coin once more and this one is also a
head.

What if the third toss is tails?

Homework: Do problems 4.46 and 4.47, 4.51