STAT 153 : Statistical Methods I Lecture 7 and 8: Introduction to Probability Carole F. Amoako-Yirenkyi Department of Statistics and Actuarial Science KNUST March 8, 2024 1 / 55 Introduction to Probability Basic Concepts 2 / 55 Probability Probability, a general concept, may be defined as the chance of an event occurring. More formally, it is the likelihood or chance that a particular outcome or event from a random experiment will occur. The basic concepts of probability include: - probability or random experiments - sample spaces - addition and multiplication rules - probabilities of complementary events. 3 / 55 Basic Concepts A random experiment is a chance process that leads to well defined results. - The word ”chance” means that even when the experiment is repeated in the same manner every time it may result in different outcomes. A trial is a single performance of an experiment (that is, a repetition of an experiment.) - A trial means rolling one die once, flipping a coin once, choosing one participant from a group of people etc. An outcome is the result of a single trial of an experiment. 4 / 55 Basic Concepts: Sample space A sample space is the set of all possible outcomes of a random experiment. Examples: - The single toss of a coin yield the possible outcomes in the set, {H,T} - A single roll of a die{1, 2, 3, 4, 5, 6} - A play of a football match yields {win(W ), loss(L), draw(D)}. 5 / 55 Tree diagrams A sample space may be defined by using a tree diagram Figure 1: Tree diagram to find sample space for gender of three children in a family 6 / 55 Basic Concepts: Events An event (E) is a subset of the sample space of a random experiment. - Events are therefore sets of outcomes. - An event with one outcome is called a single event. Eg. Getting a result of 3 when a die is rolled. - The event of getting an odd number when a die is rolled is called a compound event since it consists of three outcomes or simple events. - The event of observing a total score of 8 on two tosses of a die. B = {(2, 6), (3, 5), (4, 4), (5, 3), (6, 2)} 7 / 55 Types of Probability: Classical Probability Probability Based on Equally-Likely Outcomes Uses sample spaces to determine the numerical probability that an event will happen. Classical probability assumes that all outcomes in the sample space are equally likely to occur. Equally likely events are events that have the same probability of occurring. Example: When a die is rolled, each outcome has the same probability of occurring. The probability of any event E is Number of outcomes in E P(E) = Total number of outcomes in the sample space n(E) = n(S) 8 / 55 Interpreting Probability Figure 2: Interpreting probability values Example: If a family has three children, find the probability that exactly two of the three children are boys. 9 / 55 Types of Probability: Relative frequency probability (Empirical probability) While classical probability assumes that certain outcomes are equally likely, relative frequency probability relies on actual experience to calculate probabilities. It based on how often an event A occurs if an experiment is repeated a large number of times n. In empirical probability, one might actually roll a die 6000 times, observe the various frequencies, and use these frequencies to determine th probability of the outcome. This employs the Law of Large Numbers: If an experiment is repeated again and again, in the long-run the proportion of times the event E actually occurs is the theoretical probability. 10 / 55 Frequency interpretation in more detail What do we mean by P(E) is the long-run proportion of times E occurs in repeated experiments? As more and more information (data!) are collected, we can estimate the probability almost perfectly. In some books the empirical probability is defined as number of times A occurs out of n experiments n→∞ n n(A) = lim n→∞ n P(A) = lim 11 / 55 Types of Probability: Subjective probability Subjective probability: prob assigned to an event based on subjective judgment, experience, information and belief. Example: “There is a 90% chance that I’ll pass my driving test.” 12 / 55 Complement of an Event Another important concept in probability is that of complementary events. - The complement of an event is the set of outcomes in the sample space that are not contained in the event. ′ - Denoted as E or E c or Ē. 13 / 55 Rules of Probability Probability is a number that is assigned to each member of a collection of events from a random experiment that satisfies the following properties: If S is the sample space and E is any event in the random experiment, Rule 1: P(S) = 1 Rule 2: 0 ≤ P(E) ≤ 1 Rule 3: If an event E cannot occur, its probability is 0. Rule 4: If an event E is certain to occur, its probability is 1. 14 / 55 Implication of rules The rules of probability imply that - P(ϕ) = 0 - P(E ′ ) = 1 – P(E) or P(E) = 1 – P(E ′ ) or P(E ′ ) + P(E) = 1 - If E1 is contained in E2 , then P(E1 ) ≤ P(E2 ). 15 / 55 Probability of any Event Let all experimental outcomes be listed as simple events E1 , E2 , E3 , · · · , Ek . We can make new events from these, e.g A = {E2 , E3 }. The probability of any event is the sum of the probabilities of the experimental outcomes in the event. X P(A) = P(Ei ) Ei in A Computing probabilities involves a lot of counting and summing. 16 / 55 Example: Blood Type A random experiment for blood types of individuals has the following as all possible outcomes: S = {A, B, AB, O}. These outcomes are not equally-likely and suppose the probabilities in Ghana are P(O) = 0.44, P(A) = 0.42, P(B) = 0.10 and P(AB) = 0.04. All of these are between 0 and 1. P(O) + P(A) + P(B) + P(AB) = 1 The probability that a randomly selected individual does not have type AB is P(AB C ) = 1 − P(AB) = 1 − 0.04 = 0.96. The probability of either A or AB is P{A, AB} = P(A) + P(AB) = 0.42 + 0.04 = 0.46 17 / 55 Operations on Events In general, a compound event is a combination of two or more outcomes or simple events. We can make a new event by taking a union (∪) or intersection (∩) of two events E1 and E2 . - The Union of two events consists of all outcomes that are contained in one event or the other, denoted as E 1 ∪ E 2 . - The Intersection of two events consists of all outcomes that are contained in one event and the other, denoted as E 1 ∩ E 2. 18 / 55 Venn Diagrams Events A & B contain their respective outcomes. The shaded regions indicate the event relation of each diagram. Figure 3: (a) = Union of two sets; (b)= Intersection of two sets 19 / 55 Disjoint Events (Mutually Exclusive Events) Two events A and B are said to be disjoint mutually exclusive if they share no common outcomes. In other words, if they cannot happen at the same time. For a 6-sided die, E1 = {1, 3, 5} and E2 = {2, 4, 6} are disjoint. If E1 and E2 are disjoint, what is their intersection? 20 / 55 Disjoint Events (Mutually Exclusive Events) Two events A and B are said to be disjoint mutually exclusive if they share no common outcomes. In other words, if they cannot happen at the same time. For a 6-sided die, E1 = {1, 3, 5} and E2 = {2, 4, 6} are disjoint. If E1 and E2 are disjoint, what is their intersection? - Symbolically, A∩B =ϕ 20 / 55 Disjoint Events (Mutually Exclusive Events) Two events A and B are said to be disjoint mutually exclusive if they share no common outcomes. In other words, if they cannot happen at the same time. For a 6-sided die, E1 = {1, 3, 5} and E2 = {2, 4, 6} are disjoint. If E1 and E2 are disjoint, what is their intersection? - Symbolically, A∩B =ϕ A Venn diagram for disjoint events.. Figure 4: Venn diagram showing two disjoint events 20 / 55 Mutually Exclusive Events - Laws 1 Commutative law (event order is unimportant): - A∩B =B∩A 2 and A∪B =B∪A Distributive law (like in algebra): - (A ∪ B) ∩ C = (A ∩ C) ∪ (B ∩ C) - (A ∩ B) ∪ C = (A ∪ C) ∩ (B ∪ C) 3 Associative law (like in algebra): - (A ∪ B) ∪ C = A ∪ (B ∪ C) - (A ∩ B) ∩ C = A ∩ (B ∩ C) 4 DeMorgan’s law: ′ ′ ′ ′ - (A ∪ B) = A ∩ B ′ The complement of the union is the intersection of the complements. ′ - (A ∩ B) = A ∪ B The complement of the intersection is the union of the complements. 21 / 55 Addition Rules 22 / 55 Probability of a Union: Addition rules Addition rules are used to compute probabilities for a union of events (both disjoint and non-disjoint cases). Addition Rule 1: For two mutually exclusive events, the probability that A or B will occur is P(A ∪ B) = P(A) + P(B) since P(A ∩ B) = ϕ Addition Rule 2: If A and B are not mutually exclusive P(A ∪ B) = P(A) + P(B) − P(A ∩ B) 23 / 55 Probability Rules We therefore extend the rules or axioms of probability to include the following: Rule 5: For any two events E1 and E2 with E1 ∩ E2 = ϕ P(E1 ∪ E2 ) = P(E1 ) + P(E2 ) Rule 6: For any two events E1 and E2 with E1 ∩ E2 non-empty, P(E1 ∪ E2 ) = P(E1 ) + P(E2 ) − P(E1 ∩ E2 ) 24 / 55 Addition Rules Extended For a collection of events Ei that are pairwise mutually exclusive, that is, Ei ∩ Ej = ϕ, for all i , j, P(E1 ∪ E2 ∪ ... ∪ Ek ) = k X P(Ei ) i=1 For three or more events, P(A ∪ B ∪ C) = P(A) + P(B) + P(C) − P(A ∩ B) P(A ∩ C) − P(B ∩ C) + P(A ∩ B ∩ C) Note the alternating signs. 25 / 55 Example: Hair dye and Eye colour The table below shows the relationship between hair dye and eye colour for 1770 German men Find the probability that a randomly chosen person from this group has 1 black hair 2 black hair or red hair 3 blue eyes 4 black hair and blue eyes 5 black hair or blue eyes 26 / 55 Probability of an Intersection Multiplication Rules 27 / 55 Independent events and Exhaustive events Two events are said to be independent if the occurrence of one has no effect on the occurrence of the other. A collection of events E1 , E2 , · · · , Ek are said to be exhaustive if E1 ∪ E2 ∪ · · · ∪ Ek = S 28 / 55 Probability of an Intersection: Multiplication Rules The multiplication rules can be used to find the probability of two or more events that occur in sequence. Recall the definition of independent events. Two events A and B are independent if the fact that A occurs does not affect the probability of B occurring. Examples: 1 Rolling a die and getting a 6, and rolling a second die and getting a 3. 2 Drawing a card from a deck and getting a queen, replacing it, and drawing a second card and getting a queen. 29 / 55 Multiplication Rules Multiplication Rule 1: When two events are independent, the probability of both occurring is P(A ∩ B) = P(A) · P(B) Example: A coin is flipped and a die is rolled. Find the probability of getting a head on the coin and a 4 on the die. Ans: The sample space for a coin is S1 = {H, T }. The sample space for a dis is S2 = {1, 2, 3, 4, 5, 6} 1 1 1 Then P(H and 4) = P(H ∩ 4) = · = ≈ 0.083. 2 6 12 30 / 55 Example: Blood Type From the previous blood type example we stated that 44% of the Ghanaian population has type O blood. Suppose it is also true that 15% of the population is Rh negative and that this is independent of blood group. If this is the case, if someone is chosen at random, the probability that the person has type O, Rh negative blood is P(O ∩ Rh) = P(O) × P(Rh) = 0.44 × 0.15 = 0.066 31 / 55 Multiplication Rules Sometimes we have information before we compute a probability. Events that are not independent are said to be dependent events. In this case the outcome or occurrence of the first event affects the outcome or occurrence of the second event in such a way that the probability is changed. When events are independent, we need to consider the conditional probability of one event, given that the other event has happened. We use the notation P(E2 |E1 ) to represent the probability of E2 happening when E1 has already occurred. 32 / 55 Conditional Probability For example, we might know an individual from a group has blue eyes. Knowing this fact reduces the population to only those that have blue eyes. This can change the probability of having black hair. 33 / 55 The conditional probability of an event E2 given an event E1 , denoted as P(E2 |E1 ), is: P(E2 |E1 ) = P(E1 ∩ E2 ) P(E1 ) for P(E1 ) > 0. Example: Consider choosing a man at random from the group. The probability of the man having blue eyes given that he has black hair is P(Blue eyes|Black hair) = = P(Black hair and blue eyes) Black hair 200/1770 200 = = 0.40 500/1770 500 34 / 55 Multiplication Rule 2 Multiplication Rule 2: When two events are dependent, the probability of both occurring, P(A ∩ B) = P(A) · P(B|A) Example: Consider choosing a man at random from the group, what is the probability that the man will have red hair and brown eyes? Hair color and eye color are dependent, so finding this probability involves using a conditional probability. P(R hair and Br eyes) = P(red hair) × P(brown eyes|red hair) = 70/1770 × 20/70 = 20 1770 35 / 55 Sampling without replacement Another situation that gives rise to conditional probabilites is when we sample from a group withour replacing. Since the object is not replaced, the probability of the first event changes the probability with which the second event occurs. Example: Three cards are drawn from an ordinary deck and not replaced. Find the probability of these events. a. Getting 3 jacks b. Getting an ace, a king, and a queen in order c. Getting a club, a spade, and a heart in order 36 / 55 Independence with Multiple Events The events E1 , E2 , · · · , Ek are independent if and only if, for any subset of these events: P(Ei1 ∩ Ei2 ∩ · · · , ∩Eik ) = P(Ei1 ) · P(Ei2 ) · · · · · P(Eik ) Example: 37 / 55 Figure 5 38 / 55 Total Probability Rule Sometimes the probability of an event is given under each of several conditions. With enough of these conditional probabilities, the probability of the event can be recovered. Recall the definition of mutually exclusive events, in this case subsets: disjoint; nothing in common; cannot occur at the same time. Figure 6 ′ A and A are mutually exclusive. ′ A ∩ B and A ∩ B are mutually exclusive ′ B = (A ∩ B) ∪ (A ∩ B) 39 / 55 Total Probability Rule For any two events A and B P(B) ′ = P(B ∩ A) + P(B ∩ A ) ′ ′ = P(B|A) · P(A) + P(B|A ) · P(A ) This concept can be extended to as many mutually exclusive and exhaustive events as there are: Then P(B) = P(B ∩ E1 ) + P(B ∩ E2 + ... + P(B ∩ Ek = P(B|E1 ) · P(E1 ) + P(B|E2 ) · P(E2 ) + ... + P(B|Ek ) · P(Ek ) 40 / 55 Example: Picking a red ball Box 1 contains 2 red balls and 1 yellow ball. Box 2 contains 3 yellow balls and 1 red ball. A coin is tossed. If it falls heads up, box 1 is selected and a ball is drawn. If it falls tails up, box 2 is selected and a ball is drawn. Find the probability of selecting a red ball. 41 / 55 Example: Hand size Consider choosing someone at random from a population that is 60% female and 40% male. 1 Suppose a woman is chosen. What is the probability of having a hand size smaller than 100 cm2 ? 2 For a man, what is the probability of having a hand size smaller than 100 cm2 ? 3 What is the probability that the randomly chosen person will have a hand size smaller than 100 cm2 ? 42 / 55 Solution Let SH be the event of choosing someone with a small hand size, W the event of selecting a woman and M the event of selecting a man. 186 186/1000 = 600/1000 600 = 0.31 P(SH|W ) = 32/1000 32 = 400/1000 400 = 0.08 P(SH|M) = Therefore P(SH) = P(W ) · P(SH|W ) × P(M) · P(SH|M) = 0.6(0.31) + 0.4(0.08) = 0.186 + 0.032 = 0.218 43 / 55 Bayes’ Theorem Named after Rev. Thomas Bayes (1702-1761), an English mathematician and Presbyterian minister. He was one of the first to work on revising or estimating probabilities when new information on a random sample is obtained. The conditional probability of E2 given E1 is defined to be, P(E2 |E1 ) = P(E2 ) · P(E1 |E2 ) for P(E1 ) > 0 P(E1 ) E1 and E2 are mutually exclusive and exhaustive. 44 / 55 Bayes’ Theorem Their corresponding probabilities P(E1 ), P(E2 ), · · · P(Ek ) are called prior probabilities because they were obtained before any new information was taken into account. A probability that is revised based on new information is called a posterior probabilities It is the cornerstone of the statistical methods known as ‘Bayesian Statistics’ 45 / 55 Example: Hand size We found P(SH) = 0.218 using the total probability rule. Then if we are interested in finding P(W |SH), P(W |SH) = = P(W ) · P(SH|W )P(H) P(SH) 0.60 · 0.31 = 0.8532 0.128 Interpretation: Of the people who have hands less than 100cm2 , approximately 83.3% are women. Note: Note that this value can also be obtained straight from the table. Find P(M|SH). 46 / 55 Bayes Theorem with Total Probability If E1 , E2 , ...Ek are k mutually exclusive and exhaustive events and B is any event, P(E1 |B) = P(B|E1 )P(E1 ) P(B|E1 )P(E1 ) + P(B|E2 )P(E2 ) + ... + P(B|Ek )P(Ek ) where P(B)>0 Note : Numerator expression is always one of the terms in the sum of the denominator. 47 / 55 In biological sciences More relevant examples might be - P(heart attack this year | cholestrol > 190) - P(cavities | do not brush) - P(low stress | exercise) - P(pertussis | vaccinated) 48 / 55 Independent events Two events are independent if knowing one has occurred does not change the probability of the other. This is true for any one of the following equivalent statements: 1 P(A|B) = P(A) 2 P(B|A) = P(B) 3 P(A ∩ B) = P(A) · P(B) Recall that P(Black hair) = 0.28 and P(Black hair|blue eyes) = 0.19. Are “black hair” and “blue eyes” independent events? 49 / 55 Summary From a relative frequency perspective of n equally likely outcomes: Marginal Prob: P(A) = (number of outcomes in A) / n Union Prob: P(A ∪ B) = (number of outcomes in A ∪ B) / n Joint / Intersection Prob: P(A ∩ B) = (number of outcomes in A ∩ B) / n Conditional Prob: P(B|A) = number of outcomes in A ∩ B / number of outcomes in A Bonus advice: Always identify which of these you have been given in the question! 50 / 55 Using Probability Trees If our event E is the result of two (or more) smaller experiments, probability trees provide a convenient way to find the probability of all possible outcomes. The first experiment results in either A or B. After the first experiment, a second experiment is carried out resulting in either C or D. The four possible events are AC , AD, BC , or BD and their probabilities are obtained from a probability tree by multiplying numbers going down the appropriate branch. 51 / 55 Example: Medical testing Suppose that a particular test has a 95% chance of detecting the disease if the person has it (this is called the sensitivity of the test) and a 90% chance of correctly indicating that the disease is absent if the person really does not have the disease (this is called the specificity of the test). Suppose 8% of the population has the disease. What is the probability of testing positive? 52 / 55 53 / 55 A more relevant question... Given that my test come out positive, what is the probability that I have the disease? This is the conditional probability P(disease|test positive) = = P(disease and test positive) P(test positive) 0.076 0.168 = 0.452 54 / 55 Thank you 55 / 55