Section 3.5 – Conditional Probability Experiment: Toss a single (balanced) die. Observe the up face. S={1, 2, 3, 4, 5, 6} P(1) = P(2) =…P(6) = 1/6 Let A = {Toss an even #} = {2, 4, 6} B = {Toss a # 3} = {1, 2, 3} 4 A 1 B 2 6 3 5 S What is the probability of observing an even number? P(Toss an even #) = P(A) = P(1) + P(2) + P(3) = 3/6 = 1/2 Does the probability of observing an even number change if you know that the # was 3, i.e. that event B occurred? Out of the 3 (equally likely) sample points in the event B, that was known to have occurred, only one sample point, i.e. 2, was even, so… P(even # given that the # was 3) = # evensample points 3 =1/3 # sample points 3 The condition “you know that the # was 3” causes the event B to become the reduced sample space of a new conditional experiment. def: The conditional probability of an event A is the probability that event A occurs knowing that the event B has occurred on a single replication of the experiment. Notation: P(A|B) – “the probability of A given B” Above, we calculated P(A|B) = # evensample points 3 = 1/3 # sample points 3 This only works if the sample points are equally likely. 4 A General formula for a conditional probability P( A / B) P( A B) P(B) 3.5 Conditional Probability - 1 B 2 6 5 1 3 S Example: 1000 people classified by smoker & presence of cancer. Yes No Smoker Let Yes 50 30 80 Cancer No 200 720 920 250 750 1000 C S = { smoker} Sc = {non-smoker} C = { cancer} Cc = { no cancer} P(S) = P(Sc) = P(C ) = P(Cc ) = P(C|S) = P(C S)/P(S) = P(C|Sc) = P(C Sc)/P(Sc) = S You have a 5 times greater risk of getting cancer if you smoke. What are the chances you are a smoker (non-smoker) if you have cancer? P(S|C) = P(Sc|C) = P(S) = .25; P(S’) = .75; P(C) = .08; P(C’) = .92; P(C|S) = .2; P(C|S’) = .04; P(S|C) = 5/8; P(S’|C) = 3/8 3.5 Conditional Probability - 2