Date: 07/18/98 at 22:44:54 From: Carole Black Subject: Conditional Probability I will be a beginning math teacher in the fall and will be teaching Statistics. I am "boning up" on conditional probabilities and I have a question about an example in the Basic Probability information from the Ask Dr. Math faq. The example is discussing the independent events of drawing red or blue marbles. There are 6 blue marbles and 4 red marbles. The discussion goes on to talk about two events, the second outcome dependent upon the first. The actual example is: But suppose we want to know the probability of your drawing a blue marble and my drawing a red one? Here are the possibilities that make up the sample space: a. (you draw a blue marble and then I draw a blue marble) b. (you draw a blue marble and then I draw a red marble) c. (you draw a red marble and then I draw a blue marble) d. (you draw a red marble and then I draw a red marble) The calculation for b is given as: your probability of drawing a blue marble (3/5) multiplied by my probability of drawing a red marble (4/9): 3/5 x 4/9 = 12/45 or, reduced, 4/15. My question is: is this the same thing as P(Red|Blue)? I believe these are two different things, but I am confused as to how to explain the difference. For P(Red | Blue) I calculate this probability as: (4/15)/(6/10) = 4/9. Can you help clear up my confusion so I can explain this clearly to my students in the fall? Thank you, Carole Black
Date: 07/19/98 at 08:07:58 From: Doctor Anthony Subject: Re: Conditional Probability Your second answer P(Red|Blue) = 4/9 is correct This means that the probability of drawing a Red given that the first draw was a blue is 4/9. Note the word 'given'. We know before making the second draw that the first draw was a blue. This must be contrasted with the probability of red-blue before we start making any draw. The probability of red-blue is 6/10 x 4/9 = 4/15 and this probability is calculated before the result of the first draw is known. The word 'conditional' alerts us to the fact that we are calculating probabilities 'conditional' on knowing further information partway through the experiment. These probabilities are also referred to as 'Bayesian' probability, named after the probability theorist Thomas Bayes (1702-61) who gave this theorem: P(A and E) P(E|A) = ---------- P(A) In other words, if we know that A has occurred, then the sample space is reduced to the probability of event A, and the denominator for P(A and E) is not 1 but P(A). - Doctor Anthony, The Math Forum Check out our web site! http://mathforum.org/dr.math/
Date: 07/19/98 at 18:50:40 From: Carole Black Subject: Conditional Probability Dr. Anthony, thank you for your very quick and wonderfully clear explanation to my question. Carole Black
Search the Dr. Math Library:
Ask Dr. MathTM
© 1994- The Math Forum at NCTM. All rights reserved.