Inferential Statistics and Probability

a Holistic Approach

Chapter 4
Probability

This Course Material by Maurice Geraghty is licensed under a Creative Commons Attribution-
ShareAlike 4.0 International License.
Conditions for use are shown here: https://creativecommons.org/licenses/by-sa/4.0/

1
Probability
 Classical probability
 Based on mathematical formulas
 Empirical probability
 Based on the relative frequencies of
historical data.
 Subjective probability
 “one-shot” educated guess.

2
Examples of Probability
 What is the probability of rolling a four
on a 6-sided die?
 What percentage of De Anza students
live in Cupertino?
 What is the chance that your favorite
team will win the championship?

3
Classical Probability
 Event
 A result of an experiment, usually expressed as a letter (A, B,…)
 Outcome
 A result of the experiment that cannot be broken down into smaller
events
 Sample Space
 The set of all possible outcomes
 Probability Event A Occurs - written as P(A)
 Number of Outcomes in Event A / Number of Outcomes in Sample Space

 Example – flip two coins, find the probability of exactly 1 head.

 Sample Space = {HH, HT, TH, TT} A= {HT, TH}
 P(A) = 2/4 = 0.50

4
 Example – Field Bet in Craps
Field Bet
• 2 dice are rolled and
totaled
• Player wins even money
on
3, 4, 9, 10, 11,
• Player wins double on 2
• Player wins triple on 12
• Player loses bet on
Who5, has
6, 7,the
8 advantage in this game?

5
Sample Space
36 possible pairs of rolls

6
 More ways to make 5, 6, 7, 8

P(Win) = 16/36 = 0.444 P(Lose) = 20/36 = 0.556

7
Empirical Probability
 Historical Data National: Rate Your
community
 Relative Frequencies

Percentage of Sample
 Example: What is 60 51
50
the chance someone 40 32
30
rates their 20 13
10 3 1
community as good 0
or better?

ir

er
l

or
ce
oo

Fa

th
Po
0.51 + 0.32 = 0.83
Ex


O
Rating

8
Rule of Complement
 Complement of an
event
 The event does not
A
occur
 A’ is the A’
complement of A
 P(A) + P(A’) =1
 P(A) = 1 – P(A’)

9
 Joint Probability
 The UNION of two events A and B is
that either A or B occur (or both).
(All colored parts)

 The INTERSECTION of two events A

and B is that both A and B will occur.
(Purple Part only)

 Additive Rule:
P(A or B) = P(A) + P(B) – P(A and B)

10
Example
 In a group of students, 40% are taking
Math, 20% are taking History.
 10% of students are taking both Math
and History.
 Find the Probability of a Student taking
either Math or History or both.
 P(M or H) = 40% + 20% - 10% = 50%

11
 Mutually Exclusive
 Mutually Exclusive
 Both cannot occur
 If A and B are mutually exclusive, then
 P(A or B) = P(A) + P(B)
 Example roll a die
 A: Roll 2 or less B: Roll 5 or more
 P(A)=2/6 P(B)=2/6
 P(A or B) = P(A) + P(B) = 4/6

12
Conditional Probability
 The probability of an event occuring GIVEN another
event has already occurred.
 P(A|B) = P(A and B) / P(B)

 Example: Of all smart phone users in the US, 21% have an

Apple iPhone and AT&T. 35% of all smart phone users have
AT&T. Given a selected smart phone user has AT&T, find the
probability the user also has an Apple iPhone.

 A=AT&T subscriber B=Apple iPhone

 P(A and B) = 0.21 P(A)=0.35
 P(B|A) = 0.21/0.35 = 0.60

13
Marginal, Joint and
Conditional Probability
 Marginal Probability means the probability
of a single event occurring.
 Joint Probability means the probability of
the union or intersection of multiple events
occurring (and/or statements).
 Conditional Probability means the
probability of an event occurring given that
another event has already occurred.

14
Contingency Tables
 Two data items can be displayed in a
contingency table.
 Example: auto accident during year and DUI
of driver.
Accident No Accident Total
DUI 70 130 200
Non- DUI 30 770 800
Total 100 900 1000

15
 Marginal Probability
Accident No Accident Total
A = Accident
DUI 70 130 200
Non- DUI 30 770 800 D = DUI
Total 100 900 1000
 Find the Probability a Driver had an Accident

 P(A) = 100/1000 = 0.10

 Find the Probability was not DUI

 P(D’) = 1 – 200/1000 = 0.80

16
 Joint Probability
Accident No Accident Total
A = Accident
DUI 70 130 200
Non- DUI 30 770 800 D = DUI
Total 100 900 1000
 Find the Probability a Driver had an Accident and was DUI

 P(A and D) = 70/1000 = 0.07

 Find the Probability a Driver had an Accident or was DUI

 P(A or D) = P(A) + P(D) – P(A and D) = (100+200-70)/1000 = 0.23

17
 Conditional Probability
Accident No Accident Total
A = Accident
DUI 70 130 200
Non- DUI 30 770 800 D = DUI
Total 100 900 1000
 Find the Probability a DUI Driver had an Accident

 P(A|D) = 70/200 = 0.35

 Find the Probability a Driver who had an Accident is also DUI

 P(D|A) = 70/100 = 0.70

18
Creating Contingency Tables
 You can create a hypothetical contingency table from
reported cross tabulated data.
 First choose a convenient sample size (called a radix)
like 10000.
 Then apply the reported marginal probabilities to the
radix of one of the variables.
 Then apply the reported conditional probabilities to
the total values of one of the other variable.
 Complete the table with arithmetic.

19
 Example
Create a two-way table from the cross tabulation of
gender from the 2016 election results (from CNN)

20
 Example
First select a radix (sample size) of 10000

21
 Example
Then apply the marginal probabilities to the radix
(53% female, 47% male)

22
 Example
Then apply the cross tabulated percentages for each
gender. Make sure the numbers add up.

23
 Example
Finally, complete the table using arithmetic.

24
Multiplicative Rule
 P(A and B) = P(A) x P(B|A)
 P(A and B) = P(B) x P(A|B)
 Example: A box contains 4 green balls and 3 red
balls. Two balls are drawn. Find the probability of
choosing two red balls.
 A=Red Ball on 1st draw B=Red Ball on 2nd Draw
 P(A)=3/7 P(B|A)=2/6
 P(A and B) = (3/7)(2/6) = 1/7

25
Multiplicative Rule – Tree Diagram

26
Multiplicative Rule – Tree Diagram

27
Multiplicative Rule – Tree Diagram

28
Multiplicative Rule – Tree Diagram

29
Independence
 If A is not dependent on B, then they
are INDEPENDENT events, and the
following statements are true:

 P(A|B)=P(A)
 P(B|A)=P(B)
 P(A and B) = P(A) x P(B)

30
 Example
Accident No Accident Total
DUI 70 130 200
Non- DUI 30 770 800
Total 100 900 1000

A: Accident D:DUI Driver

P(A) = .10 P(A|D) = .35 (70/200)

Therefore A and D are DEPENDENT events as P(A) < P(A|D)

31
 Example
Accident No Accident Total
Domestic Car 60 540 600
Import Car 40 360 400
Total 100 900 1000

A: Accident D:Domestic Car

P(A) = .10 P(A|D) = .10 (60/600)
Therefore A and D are INDEPENDENT events as P(A) = P(A|D)

Also P(A and D) = P(A)xP(D) = (.1)(.6) = .06

32
Random Sample
 A random sample is where each
member of the population has an
equally likely chance of being chosen,
and each member of the sample is
INDEPENDENT of all other sampled
data.

33
Tree Diagram method
 Alternative Method of showing
probability
 Example: Flip Three Coins
 Example: A Circuit has three switches. If at least two
of the switches function, the Circuit will succeed.
Each switch has a 10% failure rate if all are
operating, and a 20% failure rate if one switch has
already failed. Find the probability the circuit will
succeed.

34
 Circuit Problem
.9 .1
Pr(Good)=
.81+.072+.064=.946

A A’
.9 .1 .8 .2

B’ B B’
B .2
.8 .8
.2 .02
.81
C C’ C C’
.072 .018 .064 .016
35
Switching the Conditionality
 Often there are questions where you desire to
change the conditionality from one variable to
the other variable
 First construct a tree diagram.
 Second, create a Contingency Table using a
convenient radix (sample size)
 From the Contingency table it is easy to
calculate all conditional probabilities.

36
Example
 10% of prisoners in a Canadian prison are
HIV positive.
 A test will correctly detect HIV 95% of the
time, but will incorrectly “detect” HIV in non-
infected prisoners 15% of the time (false
positive).
 If a randomly selected prisoner tests positive,
find the probability the prisoner is HIV+

37
 Example
.1 .9

A A’
.95 .05 .15 .85

B’ B B’
B
.095 .005 .135 .765

A=Prisoner is HIV+
B=Test is Positive for HIV
38
Example
HIV+ HIV-
A A’ Total
Test+ 950 1350 2300
B

Test- 50 7650 7700

B’

Total 1000 9000 10000

950
PA | B    .413
2300
39