Statistics AP Unit 3: Probability: Foundation for Inference

Unit 3: Probability: Foundations for Inference

Chapter 6: Probability: The Study of Randomness

6.1: Randomness
Probability: - the study of the mathematics behind the pattern of chance outcome based on empirical
(observational) data.

Randomness: - a state where the outcome may seems unpredictable (without pattern) at first, but with a
large number of trials, the proportion of outcome will eventually emerge (random
phenomenon).

Experimental Probability: - probability that came from a simulation such as tossing dice, coins etc.
- in order to use experimental probability as an inference, many independent
trials (one trial cannot influence another) will have to be performed in the
simulation.

Number of Favourable Outcomes

Experimental Probability P(A) =
Total Number of Trials

P(A) = Probability (Proportion of Outcome) of Event A

Example 1: Using the TI-83 Plus calculator, simulate spinning the spinner below 100 times and 200 times.
Determine the experimental probability of spinning a 4 for each simulation.

Simulating using 100 trials: Simulating using 200 trials:

8 1
7 2

6 3 To access , press STO

LIST

5 4 To access SortA, press 2nd STAT

Count the number of 4s in L1 Count the number of 4s in L2

From L1(34) to L1(51), there are From L2(73) to L2(96), there are
6.1 Assignment 18 fours. 24 fours.
pg. 316317 #6.2,
6.5 to 6.8 18 9 24 3
P(4) = P(4) = = 0.18 P(4) = P(4) = = 0.12
100 50 200 25

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 73.

Unit 3: Probability: Foundation for Inference Statistics AP

6.2A: Probability Models

Sample Space: - a list of all possible outcomes of an experiment.

Event: - an outcome or a set of occurrences from a subset of a sample space of a random phenomenon.

Complement: - probability of ALL Non-Favourable outcomes.

P (A) = Probability of event A happening. ()

P A = Probability that event A will NOT happen.

In general, P ( A) + P A = 1 () or ()
P A = 1 P ( A)
Example 1: Find the theoretical probability of rolling a 4 using a six-sided dice, and its compliment.

Sample Space of a dice = {1, 2, 3, 4, 5, 6}

1
P (4) = (1 favorable outcome out of a total of 6 outcomes)
6
()
P 4 = 1 P (4) = 1
1
6
()
P4 =
5
6

Tree Diagram: - using branches to list all outcomes.

- following branches one path at a time to list all outcomes.
- its limitation lies in the fact that it can only handle individual events having SMALL
number of outcomes.

Example 2: List the sample space of a family of 3 children. Let event A be at most having 2 girls, find P
(A) and P A . ()
1st child 2nd child 3rd child Outcomes
B (B, B, B) P (A) = Probability of at most 2 girls
B (no girls, 1 girl, and 2 girls)
G (B, B, G)
B 7
B (B, G, B) P (A) =
8
G
G (B, G, G) (7 favorable outcomes out of a total of 8 outcomes)
B (G, B, B)
OR we can say P (A) = 1 P (all girls)
B
1 7
G (G, B, G) P (A) = 1 P (A) =
G 8 8
B (G, G, B)
G
G (G, G, G) ()
P A = 1 P (A) = 1
7
8
()
P A =
1
8
Page 74. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

Sample Space Table: useful when there are TWO items of MANY outcomes for each trial.

Example 3: List the sample space for the sums when 2 fair dice are thrown
a. Find the probability of rolling a sum of 8.
b. What is the probability of rolling at least an 8?

5
Second Dice a. P (8) =
36
Sum 1 2 3 4 5 6 (5 favorable outcomes out of a total of 36 outcomes)
1 2 3 4 5 6 7
First Dice

b. P (at least a sum of 8) = P (sum of 8 and higher)

2 3 4 5 6 7 8
15
3 4 5 6 7 8 9 P (at least a sum of 8) =
36
4 5 6 7 8 9 10
5
P (at least a sum of 8) =
5 6 7 8 9 10 11 12
6 7 8 9 10 11 12
Example 4: Without drawing a tree diagram, determine the number outcomes that will be in the sample
space when 4 coins are tossed. What is the probability of tossing at least one head?
From Example 2, we noticed that when there were 3 events (3 children) of 2 outcomes (boy or girl) each,
there were a total of 2 2 2 = 8 outcomes.

Therefore, it can be assumed that when there are 4 events (tossing four coins) of 2 outcomes (head
or tail) each, there are a total of 2 2 2 2 = 16 outcomes.

P (at least one head) = P (1 or 2 or 3 or 4 heads)

P (at least one head) = 1 P (no head)
P (at least one head) = 1 P (all tails)
1
P (at least one head) = 1
16
(There is only 1 outcome out of a total of 16 outcomes to get all tails.)

15
P (at least one head) =
16

Sampling With Replacement: - sometimes refer to as independent events.

- when the outcome of one event does NOT affect the outcomes of the
events follow.

Example: Drawing a card out of a deck of 52 playing cards. Putting it back in the deck before drawing
another one.

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 75.

Unit 3: Probability: Foundation for Inference Statistics AP
Sampling Without Replacement: - sometimes refer to as dependent events.
- when the outcome of the first event AFFECTS the outcome(s) of
subsequent event(s).

Example: Drawing two cards out of a deck of 52 playing cards without putting the first card back in the
deck.

Multiplication (Counting) Principle: - by multiplying the number of ways in each category of any
particular outcome, we can find the total number if arrangements.
Example 5: How many outfits can you have if you have 3 different shirts, 2 pairs of pants, 4 pairs of socks,
and 1 pair of shoes?
There are 4 categories: shirt, pants, socks and shoes.

3 2 4 1 = 24 outfits
shirts pants socks shoes

When the outcome involves many restrictions, make sure

to deal with the category that is the most restrictive first.
Example 6: How many 5-digits numbers are there if
a. there is no restriction? b. they have to be divisible by 5?
1stdigit 2nddigit 3rddigit 4thdigit 5thdigit 1stdigit 2nddigit 3rddigit 4thdigit 5thdigit
9 10 10 10 10 9 10 10 10 2
(1 to 9) (0 to 9) (0 to 9) (0 to 9) (0 to 9) (1 to 9) (0 to 9) (0 to 9) (0 to 9) (0 or 5)

90000 numbers 18000 numbers

restriction
c. no digits are repeated? d. they have to be less than 40000 with non-repeated
digits?
1stdigit 2nddigit 3rddigit 4thdigit 5thdigit
9 9 8 7 6 1stdigit 2nddigit 3rddigit 4thdigit 5thdigit
(1 to 9) (0 to 9) (0 to 9) (0 to 9) (0 to 9) 3 9 8 7 6
1 digit 2 digits 3 digits 4 digits (1 to 3) (0 to 9) (0 to 9) (0 to 9) (0 to 9)
1 digit 2 digits 3 digits 4 digits
27216 numbers
9072 numbers
e. at least two digits that are the same?
At least two digits the same = same 2 digits + same 3 digits + same 4 digits + same 5 digits
OR
At least two digits the same = Total number of ways with no restriction No digits the same
= 90000 27216

62784 numbers

Page 76. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

f. Find the probability of getting a 5-digits numbers if at least 2 digits have to be the same.
Number of ways at least same 2 digits 62784
P (at least 2 same digits out of a 5-digits number) = =
Total Number of ways 90000
436
or 0.6976
625

Example 7: In an Algebra Midterm Exam, there are 33 multiple-choice questions. Each question contains 4
choices. How many different combinations are there to complete the test, assuming all
questions are answered? What is the probability that a student would guess all the answers
correctly?
For each question, there are 4 ways (choices) to answer.

For 33 questions, there are 4 4 4 4 4 4 There are 33

factors of fours.
Total number of ways to answer the test = 433 or 7.378697629 1019 ways

1
P (guessing all the right answers) = 33
= 1.35525272 1020
4

(There is a way better chance to win the lottery compared to guessing all the answers correctly!)

Example 8: Find the number of ways to select 2 individuals from a group of 6 people.
Lets suppose the names of the people are A, B, C, D, E and F.

The sample space table would be Order is NOT important! These

repeated combinations do not need
AB BA CA DA EA FA to be counted.
AC BC CB DB EB FB
AD BD CD DC EC FC Total Number of ways = 15
AE BE CE DE ED FD
AF BF CF DF EF FE

Summary of the Multiplication (Counting) Principle

Count the outcomes of the
Restrictions? YES most restricted category
YES
NO
ORDER Multiply the number of
Important? ways in each category

1.Count Sample Spaces by using tables or tree diagram.

NO
2.Cross out any duplicated results.

6.2A Assignment
pg. 322323 #6.9, 6.10, 6.12, 6.13, 6.15, 6.16
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 77.
Unit 3: Probability: Foundation for Inference Statistics AP

6.2B: Representing Probability

Basic Rules on Probability

1. Probabilities in a Finite Sample Space: - all values probability must be between 0 and 1.
- when P(A) = 1, it is CERTAIN that Event A will occur.
0 P(A) 1 - when P(A) = 0, it is IMPOSSIBLE that Event A will
happen.
- probability of Event A can consists of many outcomes
(Example: rolling a even number in a 6-sided dice
outcomes are 2, 4, 6 and the event is even numbers)
2. All Probabilities in a Sample Space adds up to 1. P(S) = 1
3. If there are two events in a sample space, A and A (textbook refer to as Ac Compliment of
Event A NOT A), then P( A ) = 1 P(A)

Equally Likely Outcomes: - when each individual outcome in a sample space has the same probability of
happening.

Equally Likely Outcomes where k = number of equal outcomes

1 Number of Outcomes Satisfying Condition of Event A
P( Each Outcome) = P(A) =
k k

Example 1: rolling a even number in a 6-sided dice outcomes are 2, 4, 6 and the event is even numbers
1 3 1
P (each outcome numbers 1 to 6) = P(Even Numbers) = =
6 6 2

Venn Diagram: - a diagram that illustrate all the events and their relationships within the entire sample
space.
The entire rectangle represents all
probabilities in the sample space (S)
P(S) = 1

A A A B This circle
represents P(B)

This circle
This partial This partial represents P(A)
rectangle rectangle The space in the rectangle but NOT part of
represents P(A) the circles represents the probability of
represents P( A )
Neither event A nor B. P( A and B )
Page 78. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

Disjoint (Mutually Exclusive) Events: - when events A & B CANNOT occur at the SAME TIME.
- either event A occurs OR event B occurs.

Disjoint (Mutually Exclusive) Events

A B P (A B) = P (A) + P (B)
means OR
NO Overlap

Example 2: A card is drawn from a standard deck of 52 cards. What is the probability that it is a red card
or a spade?
Since there is no card that is BOTH a spade and red, the events are mutually exclusive.
26 1 13 1
P (red) = = P (spade) = =
52 2 52 4
P (red spade) = P (red) + P (spade)
1 1
P (red spade) = +
2 4

3
P (red spade) =
4

Example 3: Two dice are rolled. What is the probability that either the sum is 3 or the sum is 8?

Since you can either roll a sum of 3 or a sum of 8,

Second Dice but NOT BOTH, the events are mutually exclusive.
Sum 1 2 3 4 5 6 2 5
P (3) = P (8) =
1 2 3 4 5 6 7 36 36
First Dice

2 3 4 5 6 7 8
P (3 8) = P (3) + P (8)
3 4 5 6 7 8 9 2 5
P (3 8) = +
4 5 6 7 8 9 10 36 36

5 6 7 8 9 10 11 7
P (3 8) =
6 7 8 9 10 11 12 36

6.2B Assignment
pg. 330 331 #6.18 to 6.23

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 79.

Unit 3: Probability: Foundation for Inference Statistics AP

6.2C: Independent and Dependent Events

Independent Events: - when the outcome of one event does NOT affect the outcomes of the events follow.

P (A B) = P (A) P (B)
Events A and B means AND

Example 1: What is the probability of rolling at most a 4 from a dice and selecting a diamond out of a
standard deck of cards?
Rolling a Dice and Drawing a Card are Unrelated Events Independent Events
4 13 Reduce before multiply
P (at most 4 diamond) =
6 52
2 1 2 1
P (at most 4 diamond) = = P (at most 4 diamond) =
3 4 12 6

Example 2: Find the probability of getting at least one number correct out a 3-numbers combination lock
with markings from 0 to 59 inclusive.

Each number in the combination is unrelated to each other Independent Events

All numbers are Incorrect (No number is correct) is the Compliment Event of At Least ONE
number is correct.
3
59 59 59 59
P (All numbers are Incorrect) = = (There are 60 numbers from 0 to 59 inclusive.)
60 60 60 60
P (At Least ONE number is correct) = 1 P (All numbers are Incorrect)
3
59
P (At Least ONE number is correct) = 1
60

P (At Least ONE number is correct) 0.0492 4.92%

Example 3: What is the probability of drawing two aces if the first card is replaced (put back into the deck)
before the second card is drawn?

4 4
P (drawing 2 aces) =
52 52
1 1
P (drawing 2 aces) =
13 13
1
P (drawing 2 aces) = 0.005917 0.592%
169

Page 80. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

Dependent Events: - when the outcome of the first event AFFECTS the outcome(s) of subsequent
event(s).

Example 4: What is the probability of drawing two hearts from a standard deck of 52 cards?
If the question doe NOT say, assume the card is NOT Replaced! (WITHOUT
REPLACEMENT is ALWAYS Dependent Events)
13 12 One less heart remains in the deck after the 1st draw.
P (two hearts) =
52 51 One less card total remains in
1 4 the deck after the 1st draw.
P (two hearts) =
4 17
1
P (two hearts) = 0.0588 5.88%
17

Example 5: About 40% of the population likes sci-fi movies, while 80% of the population likes comedies.
If 35% of the population likes both sci-fi and comedies, determine whether preferences for
sci-fi and comedies are independent events. Justify your answer mathematically.

P (sci-fi) = 40% = 0.40 If sci-fi and comedies are independent events, then
P (comedies) = 80% = 0.80 P (sci-fi comedies) = P (sci-fi) P (comedies)
P (sci-fi comedies) = 0.40 0.80
P (sci-fi comedies) = 0.32 (independent events)

Since the question indicates P (sci-fi comedies) is 35% = 0.35, they should be
considered as DEPENDENT EVENTS.

Example 6: There are two identical looking containers. Each contains different number of black and white
balls of the same size. Container A has 8 balls in total and 3 of those are white. Container B
has 10 balls in total and 4 of those are black. If a person has a choice to pick one ball out of
these two containers, what is the probability of selecting a black ball?

P (black) = P (Container A Black) P (Container B Black)

OR means ADD

P (black) = P (A) P (Black) + P (B) P (Black)

Container A

1 5 1 4
P (black) = +
2 8 2 10
5 4
P (black) = +
16 20
41
Container B P (black) = = 0.5125 = 51.25%
80

6.2C Assignment
pg. 335336 #6.24, 6.25, 6.27
pg. 337340 #6.29 to 6.36
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 81.
Unit 3: Probability: Foundation for Inference Statistics AP

6.3A: Union of Two Events

Joint Event: - when two (A and B) or more events can happen simultaneously.
Joint Probability: - the probability of a joint event P(A and B).
- for Independent Events, P (A B) = P (A) P (B)
- for Dependent Events, P (A B) are either given or found
Union of Two (Non-Mutually Exclusive) Events: - the occurrence of at least of one of the events A Or B
when both events A & B can occur at the SAME TIME.
- the probability of the overlapping area, P (A and B), is subtracted.
Union of Two (Non-Mutually Exclusive) Events
A B P (A B) = P (A) + P (B) P (A B)
Overlapping Area Joint Probability

P (A and B) = P (A) P (B) (if A and B are independent events)

P (A and B) must be given or found if A and B are dependent

Example 1: Two dice are rolled. What is the probability that one of the dice is 3 or the sum is 8?

Since you can have sums of 8 that have 3 on one dice, these
Second Dice are Union of Two (Non-Mutually Exclusive) Events.
10
Sum 1 2 3 4 5 6 P (one of the dice is 3) =
36
1 2 3 4 5 6 7 (sum of 6 has two 3s Not included)
First Dice

5 2
2 3 4 5 6 7 8 P (8) = P (3 on one dice sum is 8) =
36 36
3 4 5 6 7 8 9 P (3 on one dice sum of 8) = P (3) + P (8)
4 5 6 7 8 9 10 P (3 on one dice sum of 8)
10 5 2
5 6 7 8 9 10 11 P (3 on one dice sum of 8) = +
36 36 36
6 7 8 9 10 11 12 13
P (3 on one dice sum of 8) =
36
Example 2: Out of 200 people, 80 of them like sci-fi and 160 like comedies. 70 of them like both sci-fi and
comedies. Find the probability that someone will like either sci-fi or comedies.
Since you can have people prefer BOTH sci-fi and comedies, these events are non-mutually exclusive.
80 160 70
P (sci-fi) = P (comedies) = P (sci-fi comedies) =
200 200 200
P (sci-fi comedies) = P (sci-fi) + P (comedies) P (sci-fi comedies)
80 160 70
P (sci-fi comedies) = +
200 200 200
170 17
P (sci-fi comedies) = = = 0.85 = 85%
200 20
Page 82. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

Example 3: In our society, the probability of someone suffering from heart disease is 0.56 and the
probability of developing cancer is 0.24. If these events are independent and non-mutually
exclusive, what is the probability that you will suffer neither illness?

P (Heart Disease) = P (H) = 0.56 P (Cancer) = P (C) = 0.24

P (H C) = P (H) P (C) = 0.56 0.24 (independent events)

P (H C) = 0.1344

P (H C) = P (H) + P (C) P (H C) (non-mutually exclusive events)

P (H C) = 0.56 + 0.24 0.1344

P (H C) = 0.6656

P (No Heart Disease NOR Cancer) = P ( H C )

H C Area OUTSIDE the circles

P ( H C ) = 1 P (H C)
P ( H C ) = 1 0.6656

P ( H C ) = 0.3344 = 33.44%

Example 4: Two cards are drawn from a standard deck of 52 cards. What is the probability that both cards
are face cards or both are hearts?

The question did not mention replacement. Therefore, we can assume there is NO replacement
(Dependent Events).

Since you can have BOTH face cards and hearts, these events are non-mutually exclusive.

12 11 11 13 12 1
P (face cards) = = P (hearts) = =
52 51 221 52 51 17
3 2 1
P (face cards hearts) = = (K, Q, and J of Hearts)
52 51 442

P (face cards hearts) = P (face cards) + P (hearts) P (face cards hearts)

11 1 1
P (face cards hearts) = +
221 17 442

47
P (face cards hearts) = = 0.1063 = 10.63%
442

6.3A Assignment
pg. 345346 #6.37 to 6.40
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 83.
Unit 3: Probability: Foundation for Inference Statistics AP

6.3B: Conditional Probabilities

Conditional Probability: - probability of an event B given that event A occurs first.
Conditional Probability for Dependent Events (Bayess Theorem)
P(A B )
P (B | A) = P (A B) P (A) P (B) (for Dependent Events)
P ( A)
P (B | A) = Conditional Probability for Event B GIVEN that Event A has occurred

Example 1: An independent survey has indicated Mary has a probability of 0.652 of becoming the next
student council president of a university. Joseph, however, has a 0.876 chance of becoming the
vice-president of the same council. Suppose the probability that they are both elected is 0.758
because they campaign together under the guise of some divine approval (or they simply just
look good together). On the day of the election, determine the probability that Mary will
become the president given that Joseph has already been declared a winner.
First, label and define all probability notations involved.
P (J) = Probability that Joseph wins = 0.876
P (M) = Probability that Mary wins = 0.652
P(J M) = Probability that both Joseph and Mary win = 0.758
Note that they are dependent events. P(J M) P (J) P (M) [0.758 0.876 0.652]
P (M | J) = Probability of Mary winning given that Joseph has won = ?
P(J M ) 0.758
P (M | J) = = P (M | J) = 0.865
P( J ) 0.876
Therefore, in contrast to the survey, Mary has a way better chance of winning if Joseph has won.

Conditional Probability for Independent Events

P ( A B ) P ( A) P (B )
P (B | A) = = P (A B) = P (A) P (B) (for Independent Events)
P ( A) P ( A)
P (B | A) = P (B)
P (B | A) = Conditional Probability for Event B GIVEN that Event A has occurred
Intersection of Events:- the occurrence where all favourable events occur.
- basically it is P(A B), for either independent or dependent events.
Example 2: A dice is rolled once and then again. What is the probability that the second roll results a 6
given that the first roll was a 4?
1 1 1 1 1
P (4) = First Roll is a 4 = P (6) = Second Roll is a 6 = P (4 6) = =
6 6 6 6 36
P (6 | 4) = Rolling a 6 given that a 4 is rolled first = ?
P(6 4) (1 36 ) 1
Note that First Roll was already
P (6 | 4) = = P (6 | 4) = 1 1
P(4) (1 6) 6 CERTAIN (4 had happened), 1 =
6 6
Page 84. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

Example 3: The National Cancer Institute has stated that 2.278% of the population will develop colon
cancer with their lives. A new test for colon cancer, without probing, has been developed. It
has an accuracy rate of 95%.
a. Draw a tree diagram to illustrate a randomly selected patient the different outcomes and
their probabilities while underwent this new, non-probing cancer test. Verify the results
with one of the basic rule of probabilities.
b. Find the probability that you will have colon cancer given that your test was positive.
c. To test the accuracy rate, what is the probability that you will be tested negative given that
you have no colon cancer?
a. First, label and define all probability notations involved.
P (C) = Have Colon Cancer = 0.02278 P ( C ) = NO Colon Cancer = 1 0.02278 = 0.97722
P (T) = Test Accurate = 0.95 P ( T ) = Test Inaccurate = 1 0.95 = 0.05
Colon Cancer Test Accuracy Outcomes
P (C T ) = 0.02278 0.95 = 0.021641
0.95 T Have Cancer and Test Accurate (Positive Result)
C
0.02278 P (C T ) = 0.02278 0.05 = 0.001139
0.05 Have Cancer and Test Inaccurate (Negative Result)
Entire T
Population P ( C T ) = 0.97722 0.95 = 0.928359
0.95 T
0.97722 No Cancer and Test Accurate (Negative Result)
C
0.05 P ( C T ) = 0.97722 0.05 = 0.048861
T
No Cancer and Test Inaccurate (Positive Result)
Note that whether the result is positive or negative depends on whether the patient actually has
cancer or not. Test being accurate or inaccurate alone does NOT define positive or negative result.
Verification Using Probability Rule
P (C T ) + P (C T ) + P ( C T ) + P ( C T ) = 1
0.021641 + 0.001139 + 0.928359 + 0.048861 = 1
P (C Positive ) P (C T ) 0.021641
b. P (C | Positive) = = =
P (Postive ) P (C T ) + P (C T ) 0.021641 + 0.048861
P (C | Positive) = 0.306955831 = 30.69%
(There is a bigger chance you DONT have cancer even if the test was positive!)
(Therefore, DONT spend all your money because you think you will die soon!)

P (C Negative) P (C T ) 0.928359
c. P (Negative | C ) = = =
P (C ) P (C T ) + P (C T ) 0.928359 + 0.048861

P (Negative | C ) = 0.95 = 95% (Same as question stated, but rather misleading as seen in part b.)

6.3B Assignment Chapter 6 Review

pg. 350 #6.41 to 6.43 pg. 361 to 364 #6.59 to 6.66
pg. 355358 #6.44 to 6.49, 6.51, 6.53 to 6.55
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 85.
Unit 3: Probability: Foundation for Inference Statistics AP

Chapter 7: Random Variables

7.1: Discrete and Continuous Random Variables

Random Variable (X): - a variable that has numerical values that were from some random incident.
- there are two types of random variables (discrete and continuous)

Discrete Random Variable: - a random variable where the set of values can be counted by their specific
occurrences.
Examples: Rolling a six-sided dice is discrete. X = {1, 2, 3, 4, 5, 6}
The number of girls in a family of 5 children is discrete X = {0, 1, 2, 3, 4, 5}
- the general list of specific occurrence in a set of discrete random variable (commonly known as
sample space) are usually labelled as xi, where i is the individual element of the sample space
- the corresponding probability for each occurrence is pi.

Probability Histogram: - a histogram showing all the probabilities associated with all elements of a
discrete random variable.
- P (any particular outcome) is between 0 and 1
- the sum of all the probabilities is always 1.

a. Uniform Probability Histogram: - a probability histogram where the probability of one

occurrence is the same as all the others.

Example 1: List all the elements in the sample space (xi) and their probabilities (pi) for each outcome of
rolling a fair 6-diced dice. Draw the corresponding probability histogram.

Occurrence (X) x1 = 1 x2 = 2 x3 = 3 x4 = 4 x5 = 5 x6 = 6
1 1 1 1 1 1
Probability (p) p1 = = 0.167 p2 = p3 = p4 = p5 = p6 =
6 6 6 6 6 6
Probability Distribution of Rolling a Fair Dice

0.18
0.16
0.14
0.12
Probability

0.1
0.08
0.06
0.04
0.02
0
1 2 3 4 5 6
Number on a Fair Dice

Page 86. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

b. Non-Uniform Probability Histogram: - a probability histogram where the probability of one

occurrence is the different than others.

Example 2: List all the elements in the sample space (xi) and their probabilities (pi) for each outcome of
number of girls in a family of 5 children. Draw the corresponding probability histogram.
First we need to know that number of possible outcomes. According to the multiplication rule,
there are 2 2 2 2 2 = 32 different outcome (each child has 2 possibilities boy or girl).
1
Therefore, the probability of each outcome is . Listing them gives us the following table.
32
# of Girls (X) x1 = 0 Girl x2 = 1 Girl x3 = 2 Girls x4 = 3 Girls x5 = 4 Girls x6 = 5 Girls
BBBBB GBBBB GGBBB BBGGG BGGGG GGGGG
BGBBB GBGBB BGBGG GBGGG
BBGBB GBBGB BGGBG GGBGG
BBBGB GBBBG BGGGB GGGBG
BBBBG BGGBB GBBGG GGGGB
Combinations
BGBGB GBGBG
BGBBG GBGGB
BBGGB GGBBG
BBGBG GGBGB
BBBGG GGGBB
Probability (p) 1 5 10 5 10 5 5 1
= =
32 32 32 16 32 16 32 32
Note: There is an easier way to determine the theoretical pi, using binomial probability function
(Section 8.1).

Probability of the Number of Girls in a Family of 5 Children

0.35 0.3125 0.3125

0.3
0.25
Probability

0.2 0.15625 0.15625

0.15
0.1
0.05 0.03125 0.03125

0
0 1 2 3 4 5
Number of Girls

Note: Note the shape is normally distributed. If there are infinitely more bars, the histogram will
be like a bell curve.

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 87.

Unit 3: Probability: Foundation for Inference Statistics AP
Continuous Random Variable (X): - a random variable where the set of values are specified by a range
of numbers.
- there are no specific numbers and specific probabilities (no xi and
pi), but rather a specific interval of numbers and the probabilities for
that range, hence the term continuous.
Example: The Proportion of the Population (p) that will agree to a national health care system is 0.6.
The Range of Proportion of a Sample ( p ) in a survey of 2000 people that will reflect this
sentiment within 2% margin is (0.58 p 0.62). This interval has a corresponding
probability, P (0.58 p 0.62), is 94.9%

P = 0.949

p = 0.58 = 0.6 p = 0.62

Probability Distribution: - a density curve that displays the probability (area in the graph) for a
continuous random variable in an experiment.
- P (any range of numbers) is between 0 and 1.
- the sum of all ranges of probabilities is always 1.

a. Uniform Probability Distribution: - a probability distribution where its height is 1 over any
specific intervals between 0 and 1.

P=11=1 1

0 X
1
Example 3: Let X be a random variable between 0 and 1 with a uniform probability distribution. Find the
following probabilities.
a. P (X 0.75) b. P (0.35 X 0.78) c. P (X 0.2 or X 0.83)
P = 0.25 1 P = 0.43 1 P = (0.2 1) + (0.17 1)

1 1 1

X X X
0 0.75 1 0 0.35 0.78 1 0 0.2 0.83 1
0.43 0.2 0.17
0.25
P (X 0.75) = 0.25 P (0.35 X 0.75) = 0.43 P (X 0.2 or X 0.83) = 0.37
Page 88. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

b. Normal Distribution as Probability Distribution: - when a continuous random variable

distributed with the properties a normal distributed curve.
X
- we can say that X has a N (, ) distribution and Z =

Example 4: The national poll on electricity regulation indicates that 0.85 surveyed believes governments
should regulate electricity. In a random survey of 1000 Californians on their views of
electricity regulation, what is the probability that the survey results are different than the
national result by three percentage points if the standard deviation of the state survey is
0.0113?

First, we need to draw the probability distribution with N (0.85, 0.0113). Difference of three
percentage points means X 0.82 or X 0.88. We need to find the probability in this range.

P (0.82 X 0.88)
2nd
= 0.0113
DISTR
P (X 0.82) P (X 0.88) VARS

p = 0.82 = 0.85 p = 0.88

Recall the TI-83 Plus function of normalcdf. We can use it to find the P (0.82 X 0.88).
P (0.82 X 0.88) = normalcdf (0.82, 0.88, 0.85, 0.0113) = 0.9920659522
P (X 0.82 or X 0.88) = 1 P (0.82 X 0.88)
P (X 0.82 or X 0.88) = 1 0.992065922

P (X 0.82 or X 0.88) = 0.00793 = 0.793%

7.1 Assignment
pg. 373374 #7.1 to 7.3; pg. 379 #7.4 and 7.5
pg. 380384 #7.7 to 7.9, 7.13 and 7.15

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 89.

Unit 3: Probability: Foundation for Inference Statistics AP

7.2A: Mean of Random Variable and the Law of Large Numbers

Payoff: - the amount of won or lost for each possible outcome of an event.

Expected Value (Mathematical Expectation) E: - the Average Winning for each time an event occurs.

Mean of Discrete Random Variable (X): - the Expected Value when the random variables are ran
through infinite amount of trials.

When probability of each outcome is different:

E (X) = [P(A) (Payoff for A)] + [P(B) (Payoff for B)] + [P(C) (Payoff for C)] + ...
Mean of Discrete Random Variable (when each outcome is different):

E (X) = X = x1p1 + x2p2 + x3p3 + + xkpk = xipi

where k = total number of outcomes
When the probability of each outcome is the same:
Sum of Payoffs of ALL Outcomes
E(X ) =
Total Number of Outcomes
Mean of Discrete Random Variable (when each outcome is the same):

E (X) = X =
x 1 + x 2 + x 3 + ... + x k
=
x i

k k

Example 1: Find the expected value of the following spinners. The payoff of event is indicated on the
sector of the circle.
a. Since the probability of each sector is the same,
1 3 Sum of Payoffs of ALL Outcomes
E(X ) =
1 3 Total Number of Outcomes
E (X ) =
3
3 + ( 3) + 4 + ( 2) + 5 + ( 4) + 1 + ( 1) 8
E(X ) =
4 4 8
5 2 We can expect that the average payoff is 0.375 point per spin.

b.
Since the probability of each sector is different,
E ( X ) = P( A) (Payoff A) + P (B ) (Payoff B ) + P (C ) (Payoff C )
+ P (D ) (Payoff D )
3 5 1 1 1
The sectors have the fractions , , and
1
that add up to 1.
1
4 3 3 12
1 1 1 1 3 5 4 1
E ( X ) = (3) + (5) + ( 4) + ( 1) = + + +
4 4 3 3 12 4 3 3 12
E (X) = 1
We can expect that the average payoff is 1 point per spin.
Page 90. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

Example 2: Three coins are tossed. He or she wins 5 points if all of the same kind appears; otherwise, the
player losses 3 points. Using a tree diagram to determine the expected value.

1st Toss 2nd Toss 3rd Toss Outcomes 1

P(all H ) =
H (H, H, H) 8
H Payoff of all H = 5
T (H, H, T)
H
H (H, T, H)
T
6 3
T (H, T, T) P(2H or 2T ) = =
8 4
H (T, H, H) Payoff of 2H or 2T = 3
H
T (T, H, T)
T
H (T, T, H)
T 1
P(all T ) =
T (T, T, T) 8
Payoff of all T = 5
1 3 1 5 9 5
E ( X ) = (5) + ( 3) + (5) = + + E (X) = 1
8 4 8 8 4 8
We can expect, on average, losing 1 point per spin.

Example 3: If the expected value for a spin on the spinner below is 2, how many points should be awarded
in the last sector?
Let x = payoff of the unknown sector
1 1 1 1
P(unknown sector ) = 1 = E (X) = 2
3 4 4 6
8 1 1 1 1
x
= 2
35
1 ( 1) + (5) + (8) + ( x ) = 2
3 4 4 6
6 12
x 11
1 5 x =
+ +2+ = 2 6 12
3 4 6
5 12 x = 66
? 35 x
+ =2 66
12 6 x=
12

x = 5.5

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 91.

Unit 3: Probability: Foundation for Inference Statistics AP
Example 4: In a game involving rolling two dice, a player pays $1 per game. If the sum of a roll is 7 or if
the two dice turn up with the same number, the player loses. If the sum is 3 or less, or 10 or
more, the player doubles the money ($2 back, which means $1 gain). Any other sums mean a
push. The player gets the $1 back, which means $0 gain. Using a table, find the probability
of the win, push and lose of this game. Evaluate if the game is fair.

Second Dice 6 6
P(sum of 7 or same number dice) = +
36 36
Sum 1 2 3 4 5 6 1
P(sum = 7 or same # ) =
1 2 3 4 5 6 7 3
2 3 4 5 6 7 8 Payoff of Sum of 7 and two dice with same number = $1
First Dice

3 4 5 6 7 8 9 2 4 1
P(sum 3 or sum 10 ) = + P(sum 3 or sum 10 ) =
4 5 6 7 8 9 10 36 36 6
(NOT including 1&1, 5&5, and 6&6)
5 6 7 8 9 10 11
Payoff for sum 3 or sum 10 = $1
6 7 8 9 10 11 12
1 1 1
P(other sums) = 1 P(other sums) =
1 1 1 1 1 3 6 2
E ( X ) = ( 1) + (1) + (0) = + Payoff for other sums = $0
3 6 2 3 6
1 Since it costs $1 to play and the expected value is about 0.17,
E ( X ) = = $0.166 the player will expect to lose on average $1 ($0.16) = $1.16 to
6
the dealer. Therefore, it is NOT a fair game.
Law of Large Numbers: - with a large enough sample size or trials drawn from a randomized fashion,
the observed mean ( x ) will in time approach the mean of the population ().

Law of Large Numbers

x = X = x1p1 + x2p2 + x3p3 + + xkpk = xipi when k = (infinity)

Law of Small Numbers: - when the sample size or number of trials is small, it is insignificant in
determining the mean of the population because the observed mean is
calculated by an extremely small fraction of an otherwise large sample.
Example: The belief that a tail has a high probability of showing up next when the previous 10 tosses of a
fair coin were all heads. (Each toss is an independent event, and therefore the P (tail) = .)
Example: The belief that if a hockey team was having a good run of the last 5 games, that they will have a
better chance of winning the next game. (The teams performance as a function of probability is
determined by its long-term record with the same team composition. The 5 games winning
streak is a small number compared the number of games the team will play in a season, not to
mentioned winning records of playoffs.)

7.2A Assignment
pg. 389 #7.17 to 7.19
pg. 394395 #7.21 and 7.23
Page 92. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

7.2B: Rules of Means and Variances of Random Variables

Rules for Means: - when there are two independent random variables (X and Y), the total means (X + Y)
are simply the sum of the means of both random variables (X + Y).
- when the values of a particular random variable are increased by a constant factor
(a or b), the mean of this random variable will also be increased by the same factor.
(a + X = a + X and bX = bX).

Rules for Means for Independent Random Variables

1. Addition of Means: X + Y = X + Y
2. Constant Multiple: bX = bX
3. Constant Addition: a + X = a + X

Example 1: A trip to the bank involves a car ride, waiting in line and processing the paperwork at the
bank. Suppose the mean for the time of a car ride, waiting in line at the bank and processing
the paperwork are 12.5, 6.2, and 3.7 minutes respectively. Assuming that all of the above
variables are independent of each other.
a. Determine the combined mean for the entire trip to the bank.
b. If there was a delay due to road construction such that it took twice as long to get to the bank and
the wait was increased by 5 minutes due to increase bank customers during the lunch hour, what is
the new mean for the entire trip to the bank?

a. Let X = Time of Car Ride, Y = Time Waiting in Line, and Z = Time to Process Paperwork
X + Y + Z = X + Y + Z = 12.5 min + 6.2 min +3.7 min X + Y + Z = 22.4 min

b. Let 2X = Delayed Time of Car Ride, Y + 5 = Increased Time Waiting in Line,

and Z = Time to Process Paperwork
2X + (Y + 5) + Z = 2X + (Y + 5) + Z = 2 (12.5 min) + (6.2 min + 5 min) +3.7 min

2X + (Y + 5) + Z = 39.9 min

Variance of Discrete Random Variable (2X): - similar to variance of a sample (s2), variance of discrete
random variable is an average measure of how the
variable is spread out about the mean (X).

Standard Deviation of Discrete Random Variable (X): - the square root of the variance of discrete
random variable (2X).
- it measures the variability of the variable in a distribution and its
most commonly used in normal density curve with the mean (X).

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 93.

Unit 3: Probability: Foundation for Inference Statistics AP

Variance of Discrete Random Variable

Var(X) = (X2) = (x1 X)2 p1 + (x2 X)2 p2 + (x3 X)2 p3 + + (xk X)2 pk
Var(X) = 2X = (xi X)2 pi
where k = total number of outcomes

Standard Deviation of Discrete Random Variable

X = X2 = pi (x i X )
2

Example 2: Using the game in Example 4 of Section 7.2 A, the following table is constructed. Adjusting
with the $1 needed for the player to play this game, calculate the variance and the standard
deviation of this particular game.
xi pi xipi (xi X)2 pi
2
1 1 1 1
$1 $ 1 = 0.2314814815
3 3 6 3
2
1 1 1
$0 $0 0 = 0.0138888889
2 6 2
2
1 1 1 1
$1 $ 1 = 0.2268518519
6 6 6 6
1
X = $ 2X = (xi X)2 pi 2X = 0.4722222222
6

X = X2 = 0.47222222222 X = 0.687

Rules for Variances: - when there are two independent random variables (X and Y), the total variance
(2X + Y) or the difference (2X Y) in variance are simply the sum of the variances of
both random variables (2X + 2Y). This is because variance is calculated by the square
of the difference between the individual outcome and the mean, (xi X)2 pi.
- in general, as one combined variances whether it is a total or difference, the
resulting variance gets larger. (More Random Variables means Larger Variance).
- when the values of a particular random variable are increased by a constant multiple
(b), the variance of this random variable will also be increased by the square of the
same multiple (2bX = b22X).
- the increase on the random variable by addition of a constant (a) does NOT affect
the overall variance (2a + X = 2X).
- the combined standard deviation (X + Y) can be calculate AFTER the variances
have been combined or modified. (Do NOT ADD OR MULTIPLE Standard
Deviations!).

Page 94. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

Rules for Variances for Independent Random Variables

1. Addition or Difference of Variances: 2X + Y = 2X Y = 2X + 2Y
2. Constant Multiple: 2bX = b22X
3. Constant Addition: 2a + X = 2X (NO Change on Variance)
Do NOT Add or Multiply Standard Deviation (X).
Go through Variance first than Square Root!

Example 3: The following table below shows the means and standard deviations from Example 1 in this
section. Again, assuming that all of the above variables are independent of each other.
Driving to the Bank X = 12.5 min X = 2.8 min
Waiting in Line Y = 6.2 min Y = 1.6 min
Processing Paperwork Z = 3.7 min Z = 0.9 min

a. Determine the combined variance and standard deviation for the entire trip to the bank.
b. Suppose again, the drive to the bank is doubled and the wait in line was 5 minutes longer, what are
the new variance and standard deviation of the trip to the bank?
a. First, we have to calculate the variance of each random variable by squaring the standard
deviations given.
2X = (2.8)2 = 7.84 2Y = (1.6)2 = 2.56 2Z = (0.9)2 = 0.81

2X + Y + Z = 2X + 2Y + 2Z = 7.84 + 2.56 + 0.81 2X + Y + Z = 11.21

X + Y + Z = 11.21 X + Y + Z = 3.348

b. We have to examine the changes of the variances in question before combing them.
22X = 22 (2.8)2 = 31.36 2Y +5 = (1.6)2 = 2.56 (No Change) 2Z = (0.9)2 = 0.81

22X + (Y + 5) + Z = 222X + 2Y + 5 + 2Z = 31.36 + 2.56 + 0.81 22X + (Y + 5) + Z = 34.73

2X + (Y + 5) + Z = 34.73 2X + (Y + 5) + Z = 5.893

7.2B Assignment Chapter 7 Review

pg. 397 #7.24; pg. 402403 #7.25 to 7.28; pg. 406410 #7.34 to 7.37,
pg. 404405 #7.29 to 7.32 7.39, 7.41, 7.42, 7.44

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 95.

Statistics AP Unit 3: Probability: Foundation for Inference

Chapter 8: The Binomial and Geometric Distributions

8.1A: The Binomial Distributions

Binomial Setting: - a sample space with a finite number of trials (n) where there is only two outcomes
for each trial (favorable and non-favorable or success and failure).
- the probability of success (p) of each trial is the same and independent of each other.

Binomial Random Variable (X): - is defined as the number of trials that was successful or favourable.
- the range of X is all whole numbers between 0 to n, where n represents
the total number of trials.

Examples: tossing a coin, boy or girl, pass or fail, rolling a dice if specified a particle number versus not
that number.

Binomial Distribution: - the probability distribution of a binomial random variable from 0 to n number of
successes.
- there is a total of (n + 1) number of outcomes for a binomial setting with n
number of trials.
- commonly denotes as B(n, p).

Example 1: Determine if the following are binomial distribution.

a. Having children until a boy is born where X = number of children.
b. The number of correct responses in a 30 multiple choice questions test with 5 choices for each
question.
c. A survey of a proposition for an upcoming election where the only responses are yes or no.

a. Having children until a boy is born where X = number of children.

This is NOT a binomial distribution since in this case, X cannot be a range of numbers
but rather the number of trials (n).
b. The number of correct responses in a 30 multiple choice questions test with 5 choices for
each question.
This is a Binomial Distribution since there are two possible results from each question
(right or wrong), X = whole numbers between 0 to 30, all questions have the same
success probability, which is p = 0.2

c. A survey of a proposition for an upcoming election where the only responses are yes or
no.
This is a Binomial Distribution since there are two results from each respondent (yes or
no), X = whole numbers between 0 to n, all questions have the same success probability,
which is p = 0.5

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 96.

Statistics AP Unit 3: Probability: Foundation for Inference

Binomial Probability (P(X)): - the individual probability of each possible outcome of a binomial
distribution.

For Binomial Distribution where {X | 0 X n, n W},

P(at least 1) = P(1 X n) = 1 P(0)

Probability Distribution Function (pdf): - a mathematical function (formula) to calculate all probabilities
for each value of X.

binompdf (Binomial Probability Distribution Function):

- displays a binomial probability distribution when the number of trials and the theoretical
probability of the favourable outcome are specified.

B(n, p) = binompdf (Number of Trials, Theoretical Probability of Favorable Outcome)

To access binompdf:

1. Press 2nd DISTR

VARS

2. Select Option 0

To Calculate the Binomial Probability of a Particular Outcome, X:

P(X) = B(n, p, X) = binompdf (Number of Trials, Theoretical Probability of Favorable Outcome,

Particular Outcome Desire)

Cumulative Distribution Function (cdf): - a mathematical function (formula) to sum up all probabilities
for a specified range of X.
binomcdf (Binomial Cumulative Distribution Function):
- displays the sum of a specified range of binomial probabilities when the number of trials
and the theoretical probability of the favourable outcome are defined.
To Calculate the Binomial Cumulative Probabilities of a range of Outcome from 0 to X:
P(0 to X) = binomcdf(n, p, X) = binomcdf (Number of Trials, Theoretical Probability of
Favorable Outcome, 0 to Particular Outcome Desire)
To access binomcdf:

1. Press 2nd DISTR

VARS

2. Select Option A

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 97.

Statistics AP Unit 3: Probability: Foundation for Inference

Example 2: Using your graphing calculator, determine the probabilities of having any number of girls in a
family of 5 children.

1. binompdf (5, ) 2. Store answer in L2 of the STAT Editor. 3. Enter 0 to 5 in L1 of the STAT
Editor.
STO
STAT

2nd L2
ENTER
2
4. WINDOW Settings 5. Select Histogram in STAT PLOT and graph.
x: [xmin, xmax, xscl] = x: [0, 6, 1]
y: [ymin, ymax, yscl] = y: [0, 0.35, 0.05] 2nd STAT PLOT ENTER
Y=
GRAPH
Select a number
WINDOW slightly higher than the Plot1 is
maximum in L2. ON.

Must be 1 more
than the number
of trials.
Type in L2 as Frequency
L2 Select
by pressing:
2nd
2 Histogram
6. Transfer the graph on the calculator to paper.

Probability of the Number of Girls in a Family of 5 Children

Note: Distribution is symmetrical because of probability being 0.5.
0.35 0.3125 0.3125
0.3
0.25
Probability

0.2
0.15625 0.15625
0.15
0.1
0.05 0.03125 0.03125

0
0 1 2 3 4 5
Number of Girls

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 98.

Statistics AP Unit 3: Probability: Foundation for Inference

Example 3: The first week of February marks the tradition of Groundhog Day. If the groundhog sees its
own shadow, it means 6 more weeks of winter. Otherwise, spring is just around the corner.
Recent statistics has shown that the groundhog sees its shadow 90% of the time on Groundhog
Day.
a. Graph a binomial distribution to illustrate the probability that the groundhog will see its
shadow for the next ten years.
b. Find the probability that the groundhog will see its shadow 9 time out of the ten years.
c. Calculate the probability that the groundhog will see its shadow at least 6 times out of the next
10 years.
d. Determine the probability that spring is just around the corner at least 8 years out of the next
ten years.

a. Graph Binomial Distribution

1. binompdf (10, 0.90) 2. Store answer in L2 of the STAT Editor. 3. Enter 0 to 10 in L1

STO STAT

2nd L2 ENTER
2
4. WINDOW Settings 5. Select Histogram in STAT PLOT and graph.
x: [xmin, xmax, xscl] = x: [0, 11, 1]
y: [ymin, ymax, yscl] = y: [0, 0.4, 0.05]
2nd STAT PLOT ENTER
Y= GRAPH
Select a number
WINDOW
slightly higher than the Plot1 is
maximum in L2. ON.

Must be 1 more
than the number
of trials.
Type in L2 as Frequency L2
by pressing: Select
2nd 2 Histogram

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 99.

Statistics AP Unit 3: Probability: Foundation for Inference
6. Transfer the graph on the calculator to paper.
Groundhog Day Predictions for the next Ten Years

0.45 Note: Distribution skewed to the left

because of high probability of 0.90 0.387420
0.40 0.348678
0.35
Probability

0.30
0.25
0.193710
0.20
0.15
0.10 0.057396
0.05 1.0E-10 9.0E-09 3.65E-07 8.75E-06 1.38E-04 1.49E-03 0.011160

0.00
0 1 2 3 4 5 6 7 8 9 10
Number of time groundhog will see its shawdow

b. P (Sees Shadow 9 times out of 10) c. P (Shadow at least 6 times)

P(X = 9) = binompdf (10, 0.9, 9) P (6 X 10) = 1 P (0 X 5)
P (6 X 10) = 1 binomcdf (10, 0.9, 5)

P(X = 9) = 0.387420 38.7%

(or read from Table or TRACE on Graph) P (6 X 10) = 0.998364 99.8%

d. P (No Shadow at least 8 times) = P (Shadow at most 2 times)

P (Shadow at most 2 times) = P (0 X 2) = binomcdf (10, 0.9, 2)

P (No Shadow at least 8 times) = 3.736 107 = 0.000 03736%

8.1A Assignment
pg. 418 #8.1 to 8.4; pg. 423424 #8.5 to 8.7

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 100.

Statistics AP Unit 3: Probability: Foundation for Inference

8.1B: The Binomial Formulas and Simulations

Factorial (n!): - the number of ways to arrange n elements when they are in order.

Factorial
n! = n (n 1) (n 2) (n 3) 3 2 1
where n W 0! = 1
To access n!:
1. Press MATH

2. Press three times

3. Select Option 4

Example 1: Simplify or evaluate the following factorial expressions.

8! n!
a. 5! b. c.
5! (n 2)!
5! = 5 4 3 2 1 8! 8 7 6 5! n! n (n 1) (n 2 )!
= = 8 7 6 =
5! 5! (n 2)! (n 2)!
= n (n 1)

5! = 120 n!
= n2 n
8!
(n 2)!
= 336
5!

Combinations (nCr): - the number of ways r restrictions can be chosen from n elements where the order is
unimportant.
- sometimes phrased as n choose r.

Binomial Coefficient (nCk): - the number of ways to k successes can be chosen form n trials disregarding
order in a binomial experiment.
- it is an application of combinations in binomial probabilities.

Combinations To access nCr: MATH Binomial Coefficient

n n! n n!
= =
nCr = =
nCk
r r ! (n r ) ! k k! (n k ) !
n = total number of elements n = total number of trials
r = number of restrictions k = number of successes
r n where n and r W k n where n and k W

Select Option 3

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 101.

Statistics AP Unit 3: Probability: Foundation for Inference

Example 2: Evaluate the following binomial coefficients.

a. 8C3 b. 7C7 c. 8C5
8! 8! 7! 7! 1 8! 8!
8C3 = = 7C7 = = = 8C5 = =
3!(8 3)! 3! 5! 7!(7 7 )! 7! 0! 1 5!(8 5)! 5! 3!

8C3 = 56 7C7 =1 8C5 = 56

Notice 8C3 = 8C5

Binomial Probability Formula

n
P(X = k) = p (1 p )
k n k

k
n = total number of trials X = k = number of successes k n where n and k W
p = probability of success for each trial (1 p ) = q = probability of failure for each trial
P(X = k) = binomial probability of X success for n trails
Example 3: It is estimated that 15% of the population suffers some form of mental illness at one point of
their lives. In a small company that employs 20 workers,
a. determine the probability that exactly 2 employees suffer or had suffered some form of mental
illness using the binomial probability formula and verify your answer using binompdf.
b. calculate the probability that at least 1 employee suffers or had suffered some form of mental
illness using the binomial probability formula and verify your answer using binompdf.
a. P (X = 2) = ? b. P (X 1) = 1 P (X = 0)
n = 20 p = 0.15 (1 p) = 0.85 k=2 n = 20 p = 0.15 (1 p) = 0.85 k=0
n
n k 1 P(X = k) = 1 p (1 p )
k nk
P(X = k) = p (1 p )
nk
k
k
20
20 1 P(X = 0) = 1 (0.15) (0.85)
0 20 0
P(X = 2) = (0.15) (0.85)
2 20 2

2 0
P(X = 2) = 20C2 (0.15)2 (0.85)18 1 P(X = 0) = 1 [20C0 (0.15)0 (0.85)20]
1 P(X = 0) = 1 [20C0 (0.85)20]
Calculate and Verify using binompdf
Calculate and Verify using binompdf

P (X = 2) = 0.2293
P (X 1) = 0.9612

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 102.

Statistics AP Unit 3: Probability: Foundation for Inference

Simulating Binomial Experiments:

Experimental Probability: - probability that came from a simulation such as tossing dice, coins etc.

randBin (Random Binomial): generates and displays a random integer from a specified binomial
distribution.
randBin (1, p, n) = randBin (number assigned to represent success, Theoretical Probability of
favorable outcome, number of trials)
Always set to 1

To access randBin:

1. Press MATH

2. Use to access PRB

3. Select Option 7

Theoretical Probability: - probability from calculations by using sample space or other formulas.

Example 4: Find the probability of obtaining tails when a coin is tossed 500 times theoretically and
experimentally.
Theoretical Probability
1
Sample Space of Tossing a Coin = (Head), (Tail) P (Tail) =
2

Experimental Probability using Graphing Calculator

1. randBin (1, , 500) 2. Summing all trails of favorable outcome. sum (Ans)
2nd 2nd

LIST ANS
STAT ()

258 129
Experimental P (Tail) = = 0.516 =
500 250

(Your answer might be different depending on when you last RESET your calculator.)

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 103.

Statistics AP Unit 3: Probability: Foundation for Inference

Example 5: Find the theoretical and experimental probability of NOT landing on a 5 using the spinner
below if it was spun 100 times.
9 1
8 2 Theoretical Probability

7 3 P (5) =
1
9
()
P 5 = 1
1
9
()
P5 =
8
9
6 4
5 P (NOT 5)

Experimental Probability using Graphing Calculator

1. randBin (1, 8/9, 100) 2. Summing all simulations of favorable outcome. sum (Ans)
2nd 2nd

LIST ANS
STAT ()

()
Experimental P 5 =
87
100
= 0.87

Example 6: A world-class tennis player has an average of 60% win whenever she serves. In the most
recent tennis match she served 80 times and has won 45 serves. Design an experiment that will
simulate 10 matches of an upcoming tournament. What is the experimental probability that she
will reproduce the same result or better compared to her last match? Contrast this result with
the theoretical probability using the same parameters.

Experimental Probability n = 80 serves per match; 10 matches = 10 simulations

1. randBin (1, 0.60, 80), store results of each set simulation in L1, sum L1 and store the sum in the
first cell of L2.
Press ENTER ten times (10 matches) and write
randBin (1, 0.60, 80) L1: sum (L1) L2 (1)
down results as they appear.
To access press STO

1st match 1st five

To access : press 2nd CATALOG
matches
0
There are 7 matches
that she has will win
45 serves or more.
a few times Last five
matches
Experimental P (X 45)
Theoretical P(X 45) = 1 P (X 44) 7
Theoretical P(X 45) = 1 binomcdf (80, 0.60, 44) = 0.70
10
Theoretical P(X 45) = 0.7885

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 104.

Statistics AP Unit 3: Probability: Foundation for Inference

Mean and Standard Deviation of a Binomial Probability Distribution

When the frequency distribution involves binomial probabilities, we can use the following formulas
to estimate the mean and standard deviation.
= np = np (1 p )
where n = number of trials and p = probability of favourable outcome.

Example 7: Find the mean and standard deviation of the number of male students in a class of 35. Graph
the binomial distribution.
n = 35 p = 0.5 (probability of a male student from any student)
= np (1 p )
= np
= 35(0.5)
= (35)(0.5)(1 0.5)
= 35 0.5 0.5
= 17.5 = 2.96

To Graph the Binomial Distribution:

1. binompdf (35, 0.5) 2. Store answer in L2 of the STAT Editor. 3. Enter 0 to 35 in L1.
STO STAT

2nd L2
ENTER
2
4. WINDOW Settings 5. Select Histogram in STAT PLOT and graph.
x: [xmin, xmax, xscl] = x: [0, 36, 1] 2nd STAT PLOT ENTER
y: [ymin, ymax, yscl] = y: [0, 0.14, 0.01]
Y= GRAPH
Select a number
WINDOW
slightly higher than the
Plot1 is
maximum in L2.
ON.

Must be 1 more
than the number
of trials.
Type in L2 as Frequency
by pressing: L2 Select
2nd Histogram
2
8.1B Assignment
pg. 427 #8.8 to 8.10; pg. 429 #8.11, 8.13;
pg. 431432 #8.15 to 8.18; pg. 433434 #8.19 to 8.23
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 105.
Statistics AP Unit 3: Probability: Foundation for Inference

8.2: The Geometric Distributions

Geometric Setting: - where the variable, X, is defined as the number of trials to realize the first
success.
- a sample space where there is only two outcomes for each trial (favorable and non-
favorable or success and failure).
- the probability of success (p) of each trial is the same and independent of each
other.

Examples: tossing a coin until a head, keep having children until a boy is born, rolling a dice continuously
until a specified number appears.

Geometric Distribution: - the probability distribution of a geometric random variable from 1 to X number
of trials where X N.
- there is an infinite number of trials and the probability of each trial
decreases as X increases exponentially.

Example 1: Determine if the following are geometric distribution.

a. Guessing an answer to a single multiple-choice question with 5 choices until you get the right
answer.
b. Testing each toaster coming of the same assembly line until one fails to meet specification.

a. Guessing an answer to a single multiple-choice question with 5 choices until you get the
right answer.
This is NOT a geometric distribution since in this case, each time you guess on the same
question, you have one less choice to choose from. Thus, the probability of each trial is
not the same nor independent.
b. Testing each toaster coming of the same assembly line until one fails to meet
specification.
This is a Geometric Distribution since there are two possible results from each toaster
(pass or fail), X = natural numbers between 1 to infinity, all toasters have the same
probability of failing the test.

Geometric Probability (P(X = n)): - the individual probability of each possible outcome of a geometric
distribution.

Geometric Probability Formula

P(X = n) = (1 p)n 1p
X = n = number of trials until success where n 1 and n N
p = probability of success for each trial (1 p ) = q = probability of failure for each trial
P(X = n) = geometric probability of X or n trials until success is realized

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 106.

Statistics AP Unit 3: Probability: Foundation for Inference

geometpdf (Geometric Probability Distribution Function):

- displays a geometric probability distribution when the theoretical probability of the
favourable outcome and the number of trials until success are specified

geometpdf (p, X) = geometpdf (Theoretical Probability of Success, Number of Trials Until Success)

To access geometpdf:

1. Press 2nd DISTR

VARS

2. Select Option D

To Calculate the Geometric Probabilities of a Range of Trials, (1 to X): geometpdf (p, {List of Trials})

Geometric Cumulative Probabilities: - the sum of all geometric probabilities for a range of X from 1 to n.
- the Sum of ALL Geometric Probabilities = 1

X=n Geometric Probability P(X) Geometric Series (Sum of Geometric Sequence)

1 p
2 (1 p) p Sn =
(
a r n 1 )
= P (X n)
3 (1 p)2 p r 1
4 (1 p)3 p a = value of 1st term = p r = common ratio = (1 p)
5 (1 p)4 p n = the number of terms in series

P (X n) =
(
p (1 p ) 1
n
=
) (
p (1 p ) 1
n
)
= [(1 p)n 1]
(1 p ) 1 p
n
P (X n) = 1 (1 p)

Geometric Cumulative Probabilities Formula

P(X n) = 1 (1 p) n
Sum of ALL Geometric Probability = P (x ) = 1
i =1
i

X n = a range of number of trials until success where n 1 and n N

p = probability of success for each trial (1 p ) = q = probability of failure for each trial
P(X n) = geometric cumulative probability of a range of 1 to n trails until success is realized

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 107.

Statistics AP Unit 3: Probability: Foundation for Inference

geometcdf (Geometric Cumulative Distribution Function):

- displays the sum of a specified range of geometric probabilities when the theoretical probability of
the favourable outcome and the list of number of trials within a range from 1 to n are defined.
To Calculate the Geometric Cumulative Probabilities of a range of Outcome from 1 to n:
P(X n) = geometcdf(p, X)
P(X n) = geometcdf (Theoretical Probability of Success, Number of Trials from 1 to X)
To access geometcdf:

1. Press 2nd DISTR

VARS

2. Select Option E

Example 2: The probability that anyone over the age of 50 had experienced heart attack at least once is
0.32. A room full of people who are over the age of 50 were asked if they had ever
experienced a heart attack. You are to ask each person if they have experienced a heart attack
until one admits it.
a. Using the formula and pdf of your calculator, determine the probability that the first person
that experienced heart attack was the fourth person asked.
b. Construct a probability distribution graph of this experiment for the first 20 people asked.
c. Calculate the probability that at most four people are healthy before a respondent had a heart
attack was asked. How is this question different than the one in part a?
a. P (X = 4) = ? b. Graphing Geometric Probability Distribution
n = X = 4 p = 0.32 (1 p) = 0.68 1. Type 1 to 20 in L1 using seq (X, X, 1, 20, 1) L1
n1
P(X = n) = (1 p) p To access seq:
P(X = 4) = (1 0.32)4 1(0.32)
P(X = 4) = (0.68)3(0.32) a. Press 2nd LIST
Calculate and Verify using geometpdf STAT

b. Use to access OPS

c. Select Option 5

To access X: press X,T,,n

P (X = 4) = 0.1006

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 108.

Statistics AP Unit 3: Probability: Foundation for Inference

2. Generate Geometric Probabilities and store in L2 3. WINDOW Settings

x: [xmin, xmax, xscl] = x: [1, 21, 1]
a. Use geometpdf (0.32, {1, 2, 3,, 19, 20}) L2 OR
y: [ymin, ymax, yscl] = y: [0, 0.40, 0.05]
b. Use geometpdf (0.32, L1) L2
WINDOW

Select a number
slightly higher
than the
maximum in L2.
4. Select Histogram in STAT PLOT and graph. Must be 1 more than
the number of trials.
2nd Plot1 is
ON.
GRAPH
STAT PLOT
Y=

ENTER Type in L2 as Frequency

Select
by pressing: L2
Histogram
2nd 2 Note: The geometric distribution graph
forms an exponential curve.
c. P (X 5) because at most 4 healthy people means maximum of 5 people asked (4 healthy and 1
had heart attack) = 5 trials.
P (X 5) = P(X = 1) + P(X = 2) + P(X = 3) + P(X = 4) + P(X = 5)
P (X 5) = 0.32 + (0.68)(0.32) + (0.68)2(0.32) + (0.68)3(0.32) + (0.68)4(0.32)
Or Using the Formula: p = 0.32 and n = 5 Calculate and Verify using geometcdf
P (X n) = 1 (1 p)n
P (X 5) = 1 (1 0.32)5
P (X 5) = 0.8546

Mean, Variance, Standard Deviation of a Geometric Probability Distribution

When the frequency distribution involves geometric probabilities,

1 1 p 1 p
= 2 = =
p p2 p

= mean or expected value of a geometric probability distribution

2 = variance of a geometric probability distribution
= standard deviation of a geometric probability distribution
p = probability of favourable outcome

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 109.

Statistics AP Unit 3: Probability: Foundation for Inference

Cumulative Geometric Probabilities of More than n Trials P(X > n)

P(X > n) = 1 P (X n) = 1 [1 (1 p)n]

P(X > n) = 1 1 + (1 p)n

Cumulative Geometric Probabilities of More than n Trials

P(X > n) = (1 p)n
X > n = a range of number of trials until success where n 1 and n N
p = probability of success for each trial (1 p ) = q = probability of failure for each trial
P(X > n) = geometric cumulative probability of n to infinity trials until success is realized
Example 3: A company that manufactures high precision communication products finds that 98% of its
newest product meets the pre-set standard. The manager decides to test each unit as it comes
off the assembly line. As soon as a unit is found to be defective, the line will be shut down and
engineers will analyze the process before production resumes.
a. How many units are expected to come off the assembly line before a defective one is found?
Calculate its variance and standard deviation of the assembly line.
b. What is the probability that the first 50 units are produced before the line is shut down due to
an appearance of a defective unit?
c. Calculate the probability that more than 50 units are produced before the line is shut down.
d. Determine the probability that at most 50 units are produced before a defective one is found.
a. Expected Number = = ? b. P (X = 50) = ? Calculate and Verify using
p = 0.02 defective (98% passes) n = X = 50 p = 0.02 geometpdf
(1 p) = 0.98
1 1
= = = 50
p 0.02 P(X = n) = (1 p)n 1p
1 p 1 0.02 2 P(X = 50) = (1 0.02)50 1(0.02)
2 = = = 2450
p 2
(0.02)2 P(X = 4) = (0.98)49(0.02)
P (X = 50) = 0.007432
= 2 = 2450 = 49.5

c. P (X > 50) = ? d. P (X 51) because at most 50 good units means maximum of

p = 0.02 defective n = 50 51 units produced (50 good ones and 1 defective) = 51 trials.
p = 0.02 and n = 51 Calculate and Verify using geometcdf
P(X > n) = (1 p)n
P(X > 50) = (1 0.02)50
P (X n) = 1 (1 p)n
P (X 51) = 1 (1 0.02)51
P(X > 50) = 0.3642
P (X 51) = 0.6431

Note that P (X 51) 1 P (X > 50)

8.2 Assignment
pg. 439440 #8.24 to 8.26;
pg. 443 #8.27 to 8.29; Chapter 8 Review
pg. 444446 #8.30 to 8.35 pg. 447451 #8.37, 8.39,
8.41, 8.43 and 8.44
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 110.
Statistics AP Unit 3: Probability: Foundation for Inference

Chapter 9: Sampling Distributions

9.1: Sampling Distributions

Parameters: - any values or measures that describe the population.

Statistic: - a value or measure was obtained from data that were samples of the population.

Example 1: It was found that the mean age for all teachers in Suburbia High School was 42 years old. This
is low compared to the national average taken from a recent survey to be 48 years old. Identify
the statistic and parameter from the above statements.
Parameter: Mean Age of All Teachers in Suburbia High School being 42 years old.
(Population defined as all teachers in Suburbia High.)
Statistic: National Average of Teachers Age from Recent Survey to be 48 years old.
(The data was from a Sample of the Population of all Teachers in the Nation.)

Population Proportion (p): - a parameter that describes the proportion of the population that fits a certain
category on any topic.

Example: From all income tax filed last year, the government has published that 14% of all two parents
families has one parent working.

Sample Proportion ( p ): - a statistic that describes the proportion of a sample that fits a certain category
on any topic.

Example: A sample of 2500 college students was randomly chosen, and it is found that 64% have college
loans to support their tuitions.

Sampling Variability: - the natural difference between the similar repeated random samples on the same
survey topic.

Sampling Variability is NOT a Concern as long as:

1. The random sample is taken from a large population, and the sample size is big.
2. If repeated random samples are taken, they are of the same size and from the same population.

When Sampling Variability is Minimized, Sample Proportion ( p ) = Population Proportion (p)

Sampling Distribution: - the distribution of the all sample proportions from repeated random sampling
surveys from the same population.
- have the same features as other distributions (such as shape, outliers, centre and
spread).

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 111.

Unit 3: Probability: Foundation for Inference Statistics AP
Describing Sampling Distributions:

If a fair sized sample is taken randomly from a large population, then

1. The Overall SHAPE of the distribution is Symmetrical and very similar to a Normalized curve.
2. There will be No Outliers.
3. The CENTRE of the Sampling Distribution is the Sample Proportion ( p ), which is the same as the
Population Proportion (p).
4. The SPREAD of the Sampling Distribution is the same as the Standard Deviation of the curve. It
depends on the Size of the Sample. The bigger the sample size, the smaller is the spread.

Count of Samples (size 2500)

Count of Samples (size 100)

= 0.05 = 0.01

2 1 p = p +1 +2 2 1 p = p +1 +2
0.40 0.45 0.50 0.55 0.60 0.48 0.49 0.50 0.51 0.52
Sample Proportion Sample Proportion
The Standard Deviation of Sample Size of 100 for The Standard Deviation of Sample Size of 2500
Sample Proportion of 0.50 is 0.05. This means for Sample Proportion of 0.50 is 0.01. This means
95% of the time (19 times out of 20), the accuracy 95% of the time (19 times out of 20), the accuracy
is only within 10% of the Population Proportion. is only within 2% of the Population Proportion.

Example 2: Simulate an opinion poll of 40 people with the population proportion of 0.3 says yes using
line 106 of the Random Digits Table.
a. Decide on a method to select the digits that will represent the yes response.
b. Repeat this simulation 20 times by using successive rows. Record the results in the table.
c. Input your data to your calculator and graph the resulting sampling distributions.
d. Determine the mean from the 20 sampling proportions. How does it compare with the actual
population proportion of 0.3?

a. Since p = 0.3 or 30%, we can define digits 1, 2, and 3 as yes, where as digits 4 to 9 and 0 as no.

b. Line 106: 68417 35013 15529 72765 85089 57067 50211 47487 p = 10/40 = 0.25
Line 107: 82739 57890 20807 47511 81676 55300 94383 14893 p = 11/40 = 0.275
(Repeat to Line 125 and record in Table)

Page 112. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

Line # p Line # p Line # p Line # p

106 10/40 = 0.25 111 10/40 = 0.25 116 15/40 = 0.375 121 12/40 = 0.3
107 11/40 = 0.275 112 8/40 = 0.2 117 16/40 = 0.4 122 11/40 = 0.275
108 8/40 = 0.2 113 13/40 = 0.325 118 8/40 = 0.2 123 12/40 = 0.3
109 12/40 = 0.3 114 13/40 = 0.325 119 5/40 = 0.125 124 11/40 = 0.275
110 8/40 = 0.2 115 15/40 = 0.375 120 13/40 = 0.325 125 17/40 = 0.425
c. Input Data into Calculator and Graph Sampling Distribution.
1. Enter data in L1 2. Turn STAT PLOT On and Set up Window
Since p = 0.3,
we double it for
a symmetric
view
The smallest
interval is
1/40 = 0.025
There is no L2. Select
3. Graph (Leave it as 1) Histogram
Sampling
Population
Distribution Notice that the Sampling Distribution is Skewed
Distribution
to the right and NOT Normally Distributed. This
is due to the Sample Size (40) and the Number of
Repetitions (20) are relatively SMALL. As they
increase, the sampling distribution will become
the normal distribution of the population.

p = 0.30
d. Mean of 20 repetitions of Sample Size of 40.

2nd LIST
STAT

1. Press twice

2. Select Option 3

p = 0.285 where as p = 0.30. Again, the reason for the difference is because the Sampling
Distribution is Skewed to the right due to Small Sample Size and Small Number of Sampling
Repetitions.

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 113.

Unit 3: Probability: Foundation for Inference Statistics AP
Bias Statistic: - occurs when the sample proportion is somewhat above or below the population
proportion due to overestimation, undercoverage or underestimation.

Unbiased Statistic: - when the sample proportion matches very close or exactly with the population
proportion.

Variability of a Statistic: - sometimes refer to as the spread of the sampling distribution.

- the higher the variability, the more spread out are the sampling
proportions if the samples are taken repeatedly.

Example 3: Classify the following statistics as bias versus unbiased along with their variability as high or
low.
a. A state wide survey of a 100 college students has found that 30% still smokes regularly in
Arizona. The survey is accurate within 9%, 19 times out of 20. A state university
administration has found through their mandatory health plan that 16% of its students actually
smokes.
b. Another statewide survey of 2000 college students has found the same result with 30% of the
respondent smokes regularly in California. This survey is accurate within 1.8%, 19 times out
of 20. The same university administration analyzed its student population still finds 16% of
them smokes.
c. A poll to predict the upcoming proposition on regulating electricity in California reported 80%
would vote yes. The poll uses 200 people across the state and is accurate with 5.7%, 19 times
out of 20. The actual proposition result was 81% in favour of regulating electricity.
d. The same poll on the regulating electricity in California is now done with 1000 people. The
same result was found to be 80% in favour and the reported accuracy is within 2.5%, 19 times
out of 20. Again, the actual proposition was 81% in favour of regulating electricity.

a. High Bias ( p = 30% versus p = 16%): Sampling Proportion is too far from Population Proportion.
High Variability: Standard Deviation is 4.5% (9% at 2 = 95% of the time; 4.5% at 1).

b. High Bias ( p = 30% versus p = 16%): Sampling Proportion is too far from Population Proportion.
Low Variability: Standard Deviation is 0.9% (1.8% at 2 = 95% of the time; 0.9% at 1).

c. Low Bias ( p = 80% versus p = 81%): Sampling Proportion is very close to Population Proportion.
High Variability: Standard Deviation is 2.85% (5.7% at 2 = 95% of the time; 2.85% at 1).

d. Low Bias ( p = 80% versus p = 81%): Sampling Proportion is very close to Population Proportion.
Low Variability: Standard Deviation is 1.25% (2.5% at 2 = 95% of the time; 1.25% at 1).

Note: For a random sample, the Sample Size does NOT necessary affect the Sample Proportion or its
Bias, but it has a Strong Effect on the Standard Deviation and thus the Variability of the
distribution.

9.1 Assignment
pg. 457 #9.1 to 9.4;
pg. 468 #9.9 and 9.10;
pg. 471 #9.13 and 9.14
Page 114. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

9.2: Sampling Proportions

Normal Approximation of Sampling Proportion in a Binomial Setting

When an unbiased survey involves a binomial setting using a categorical variable, and the sample
size along with the sampling proportion met the following conditions, we can calculate the mean and
standard deviation, and use the normal approximation to determine the relative probability.
Conditions for Normal Approximation of a Sampling Proportion in a Binomial Setting.
1. The Population must be At LEAST 10 times the Sample Size (npop 10nsample).
2. To Ensure a good fit to the Normal Curve, np 10 and n(1 p) 10
Mean and Standard Deviation of Normal Approximation of a Sampling Proportion in a
Binomial Setting (Categorical Variable n N)
p(1 p )
p = p p =
n
n = Sample Size and p = Population Proportion or Theoretical Probability

Using the bell curve to approximate a binomial distribution really depends on the number of trials, n.
When n is small, there are very few bars on the binomial distribution and the bell curve does not fit
the graph well. However, when n is large, the bell curve fits the binomial distribution much better.
Therefore, we can use the area under normal bell curve to approximate the cumulative sum of the
binomial probabilities.
n=4 p = 0.5
np = 4 0.5 = 2 (less than 10)
n(1 p) = 4 (1 0.5) = 2 (less than 10)
CANNOT use Normal Approximation
Bell-Curve does NOT fit the binomial distribution well.

n=9 p = 0.5
np = 9 0.5 = 4.5 (less than 10)
n(1 p) = 9 (1 0.5) = 4.5 (less than 10)
CANNOT use Normal Approximation. Bell-Curve does NOT
fit the binomial distribution well. Its still not good enough.

p (1 p ) p = p n = 20 p = 0.5
p =
n
p = 0.5 np = 20 0.5 = 10 (at least 10)
p = 0.1118
n(1 p) = 20 (1 0.5) = 10 (at least 10)
CAN use Normal Approximation. Bell-Curve fits the
binomial distribution well. (MORE bars means better fit!)

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 115.

Unit 3: Probability: Foundation for Inference Statistics AP
Example 1: A multiple-choice test has 10 questions. Each question has 4 possible choices.
a. Determine whether the conditions for normal approximation are met.
b. Graph the resulting binomial distribution.
c. Find the probability that a student will score exactly 6 out of 10 on the test.
d. Calculate the probability that a student will at least pass the test.
a. Determining Conditions for Normal Approximation
1
n = 10 questions p= = 0.25 (probability of guessing a question correct)
4
np = 10 0.25 n(1 p) = 10 (1 0.25)
np = 2.5 (less than 10) n(1 p) = 10 0.75
n(1 p) = 7.5 (less than 10)
Since the np and n(1 p) conditions are NOT met, we CANNOT use the normal approximation.

b. To Graph the Binomial Distribution:

1. binompdf (10, 0.25) 2. Store answer in L2. 3. Enter 0 to 10 in L1.

STO STAT

2nd L2 ENTER
2
4. WINDOW Settings 5. Select Histogram in STAT PLOT and Graph.
x: [xmin, xmax, xscl] = x: [0, 11, 1]
y: [ymin, ymax, yscl] = y: [0, 0.30, 0.05] 2nd STAT PLOT ENTER
Y= GRAPH
WINDOW Select a number
slightly higher than the Plot1 is
maximum in L2. ON.

Must be 1 more
than the number Type in L2 as Select Histogram
of trials. Frequency

c. P (6) = binompdf (10, 0.25, 6) d. P (passing) = P (at least 5 out of 10) = P (X 5)

P (X 5) = 1 P (X 4)
P (X 5) = 1 binomcdf (10, 0.25, 4)
exactly 6
successes
P (passing) = 0.07813
P (6) = 0.01622

Page 116. Copyrighted by Gabriel Tang B.Ed., B.Sc.

Statistics AP Unit 3: Probability: Foundation for Inference

Example 2: A multiple-choice test has 50 questions. Each question has 4 possible choices.
a. Determine whether the conditions for normal approximation are met.
b. Find the mean and standard deviation
c. Graph the resulting binomial distribution.
d. Find the probability that a student will score exactly 30 out of 50 on the test.
e. Calculate the probability that a student will fail the test.

a. Determining Conditions for Normal Approximation

1
n = 50 questions p = = 0.25 (probability of guessing a question correct)
4
np = 50 0.25 n(1 p) = 50 (1 0.25)
np = 12.5 (greater than 10) n(1 p) = 50 0.75 = 37.5 (greater than 10)
Since both np and n(1 p) condition are met, we CAN use the normal approximation.

b. Mean and Standard Deviation of Binomial Distribution (NOT Sampling Distribution)

= np = 50(0.25) = 12.5
= 12.5
= 3.062
= np(1 p ) = (50)(0.25)(1 0.25) = 3.061862178
c. To Graph the Binomial Distribution:
1. Enter 0 to 50 in L1 using seq (X,X, 0, 50, 1). 2. binompdf (50, 0.25, L1) and store answer in L2.

3. WINDOW Settings 4. Select Histogram in STAT PLOT and Graph.

x: [xmin, xmax, xscl] = x: [0, 47, 1]
y: [ymin, ymax, yscl] = y: [0, 0.14, 0.02]

Calculator can
only handle 48
bars (0 to 47)

5. To Overlay the Binomial Distribution with a Normal Bell Curve

[Enter Equation: normalpdf (X, , )] GRAPH
Y=
= 12.5
= 50 0.25 0.75
use exact value for
= 12.5
Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 117.
Unit 3: Probability: Foundation for Inference Statistics AP
d. P (30) = binompdf (50, 0.25, 30) e. P (fail) = P (0 X 24)
Since Normal Approximation is allowed, we can use
normalcdf to calculate area under the bell curve.
P (0 X 24) = normalcdf (0, 24, 12.5, 3.062)
7
P (30) =1.296 10
P (0 X 24) = 0.99989
We can also use binomcdf (50, 0.25, 24) = 0.99987
Example 3: Suppose that 85% of underage college students drink alcohol regularly.
a. What is the probability that in a simple random survey of 100 students the sample proportion
with be within 5 percentage points?
b. Determine the sample size needed to achieve the probability of 0.95 of the time the sample
proportion will be within 1 percentage point.

a. p = 0.85 n = 100 Within 5% = 0.85 0.05 X 0.85 + 0.05 P (0.80 X 0.90) = ?

First, we have to decide on the whether the conditions are met.
1. It is likely that there are more than 1000 college students (10 100) in the population.
2. np = 100 0.85 = 85 (greater than 10) and n(1 p) = 100 (1 0.85) = 15 (greater than 10).
Since the conditions are met,

p = p p (1 p )
p = p = 0.85
n
p = 0.85
0.85(1 0.25 ) p = 0.03571
p = = 0.0357071421
100
P (0.80 X 0.90) = normalcdf (0.80, 0.90, 0.85, 0.0357071421) P (0.80 X 0.90) = 0.8386
With the population proportion being 85%, using the sample size of 100, the sample proportion will
fall within 5 percentage points of the population proportion 83.86% of the time.
b. First we have to determine the z-scores corresponding with the mid 95% of the bell curve.
P = 0.95 + 0.025 = 0.975 P (0.86 X 0.87) = 0.95 n=?
X p
zupper = invNorm (0.975) Using z =
zupper = 1.959963986 p
P = 0.025
P = 0.95 X p 0.86 0.85
p = =
z 1.959963986
p = 0.0051021346
Xlower = 0.84 p = 0.85 Xupper = 0.86
zlower zupper

p (1 p ) p(1 p ) (0.85)(1 0.85)

Since p = n= = = 4897.860004 n = 4898 people
n ( )
p
2
(0.0051021346)2

9.2 Assignment
pg. 477480 #9.15 to 9.25
Page 118. Copyrighted by Gabriel Tang B.Ed., B.Sc.
Statistics AP Unit 3: Probability: Foundation for Inference

9.3: Sampling Means

Parameter and Statistic Means and Standard Deviations of Quantitative Variables

When an unbiased survey on a large population involves quantitative variable, we can calculate
the mean and standard deviation and use the normal approximation to determine the relative
probability.
Symbols need to be distinguished between parameter and statistic measures because they are
both quantitative variables.
Parameter Mean of Population () Parameter Standard Deviation of Population ()
Statistic Mean of Sample ( x or x ) Statistics Standard Deviation of Sample (s or x )

Mean and Standard Deviation of Normal Approximation of a Sample that uses Quantitative
Variable (X R)

x = x = s = x = n = Sample Size
n

Example 1: The average life span for all Americans in 2000 is 76.9 years of age with a standard deviation
of 8.4 years. In a random sample 20 Americans who has died in 2000, calculate the mean and
standard deviation of the sample.

= 76.9 years = 8.4 years n = 20 x = ? x = ?

8 .4
x = x = 76.9 years x = = x = 1.878 years
n 20

Central Limit Theorem: - states that the sampling of any form of probability distributions will always

result in a sampling distribution resembling the Normal Distribution N ,
n
- the larger the sample size n, the smaller the standard deviation or the
spread of the sample distribution, s. Hence, the sampling distribution gets
closer to the sample mean (central limit).

(Check out Central Limit Theorem Animations at:)

1. Central Limit Theorem in Action: http://www.rand.org/methodology/stat/applets/clt.html
2. Uniform Probability Distribution: http://www.statisticalengineering.com/central_limit_theorem.htm
3. Various Form of Sampling Distributions: http://www.ruf.rice.edu/~lane/stat_sim/sampling_dist/
4. Sampling Exponential Distribution: http://bcs.whfreeman.com/ips4e/cat_010/applets/CLT-
SampleMean.html)

Examples of Exponential Distributions: the probability of a life time of a light bulb or any electronic
component once it is turned on, and the amount of time to serve customer.

Copyrighted by Gabriel Tang B.Ed., B.Sc. Page 119.

Unit 3: Probability: Foundation for Inference Statistics AP
Example 2: In 2002, the Dow Jones Industrial Average was normally distributed and had an average lost of
16.1% with a standard deviation of 10.4%. If 15 trading days were randomly selected from all
trading days in 2002,
a. what are the mean and standard deviation of the sample?
b. calculate the probability in the sample that the Dow will at least break even? How does this
sample probability compare to the similar probability of the entire population?

a. = 16.1% = 10.4% n = 15 x = ? x = ?

10.4
x = x = 16.1% x = = x = 2.685%
n 15
b. For the Sample:
P (X 0%) = normalcdf (0, 1 1099, x , x )
10.4
P (X 0%) = normalcdf (0, 1 1099, 16.1, )
15
P (X 0%) = 1.02 109

For the Population:

P (X 0%) = normalcdf (0, 1 1099, , )
P (X 0%) = normalcdf (0, 1 1099, 16.1, 10.4)

P (X 0%) = 0.0608

Example 3: A call centre manager finds that the time it takes to wait for a call representative to a phone call
is exponentially distributed with a mean of 2.8 minutes and a standard deviation of 0.8
minutes. If a sample of 5 calls were taken in a given day, what is the sample probability that a
call will be answered within the first two minute?
b. = 2.8 minutes = 0.8 minutes n=5
P (X 2 min) = normalcdf (0, 2, x , x )
0 .8
P (X 2 min) = normalcdf (0, 2, 2.8, )
5
P (X 2 min) = 0.01267

Law of Large Number: - states that as the number of sample size increases, the sample mean, x or
x will get increasingly closer to the population mean, . This is the direct

result of the central limit theorem due to s = x = .
n
9.3 Assignment
pg. 485486 #9.26 to 9.29; Chapter 9 Review
pg. 491493 #9.30 to 9.35; pg. 497498 #9.41 to 9.47
pg. 494495 #9.36 to 9.39
Page 120. Copyrighted by Gabriel Tang B.Ed., B.Sc.