251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) Graded Assignment 3 Name: Key x 16 for the following distributions (Use tables in d, e, f, h and j.): a. Continuous Uniform with c 1, d 14 (Make a diagram!). b. Continuous Uniform with c 10 , d 25 (Make a diagram!). c. Continuous Uniform with c 1, d 10 (Make a diagram!). d. Binomial Distribution with p .35, n 25 . e. Binomial Distribution with p .65, n 25 . f. Binomial distribution with p .03, n 20 (Approximate Solution) g. Geometric Distribution with p .35. 1. (Source 03) Find P 10 h. Poisson Distribution with parameter of 9. i. Show how you would do this for a Hypergeometric Distribution with p .35, n 25 , N 100 . Remember M Np . j. Hypergeometric Distribution with p .35, n 25 , N 1000 . (Approximate Solution) p .35, n 25 , N 100 , find Px 1 . (Extra credit if you actually k. For a Hypergeometric Distribution with evaluate this.) l. (Extra credit) Find P 10 x 16 for an exponential distribution with c .01 . 2. Find the Mean and Standard deviation for the following distributions. a. Continuous Uniform Distribution with c 10 , d 25 . p .65, n 25 . c. Geometric Distribution with p .35 b. Binomial Distribution with d. Poisson Distribution with parameter of 9. e. Hypergeometric Distribution with p .35, n 25 , N 100 . f. Compare means and standard deviations for a Binomial distribution with Hypergeometric Distribution with p .35, n 25 , and a p .35, n 25 , N 1000 . c .01 g. (Extra credit) Exponential distribution with 3. Identify the distribution and do the following problems. Use tables where possible. a. This problem uses all four discrete distributions. (i) An IRS auditor is about to audit 6 returns from a group of returns for improper deductions. Assume that these returns come from a group of 80 returns in which 25% have improper deductions, what is the probability that exactly one of the 6 has improper deductions? (ii) What is the probability that at least one of the six has improper deductions? (iii) If two or more of the 6 returns have improper deductions, the entire group of 80 will be audited. What is the chance that that happens? (iv) Assume that the 6 returns come from a large group of returns (more than 120) that is 25% improper, and if two or more are improper, the whole group will be audited, what is the probability that the whole group will be audited? (Answers that use a specific population size will not be counted.) (v) Assume that a large number of returns are audited from this group that is 25% improper, what is the chance that the second return is the first one found with improper deductions? (vi) What is the chance that the first improper return will be among the first ten audited? (vii) Assume that only 2% of the returns are improper and a sample of 100 taxpayers is taken, what is the chance that at least 3 have improper returns? b. A fast food hamburger must be cooked a minimum of 90 seconds to be safe. At my lunch spot the time that burgers are on can be represented by a continuous uniform distribution with limits of 80 and 120 seconds. (i) What is the chance that a hamburger will be safe? (ii) What is the chance that, if I order 2 hamburgers both will be safe? (iii) What is the chance that if I order 3 hamburgers, at least one will be unsafe? c. The number of students who come to my office has a Poisson distribution with a parameter (mean) of 2.5 a day. (i) What is the chance that no students come in in a day? (ii) In a 5-day week? (iii) What is the chance that more than 5 come in in a day (iv) in a 5-day week? d. Complaints come in to my website at an average of seven an hour and take an hour to adjust. How many complaint handlers do I need to keep the probability of a complainer having to wait below 1%? (Hint: Look at a Poisson table with a mean of 7. What is the probability that x is above 10?) 1 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) e. If 80% of shoppers at a website abandon their shopping carts before quitting and there are currently 20 shoppers on my site, what is the chance that at least one will buy something:? What is the average number that will buy something? What is the most likely number that will buy something? f. According to Bowerman and O’Connell, in the movie Coma a young nurse finds that 10 of the last 30000 patients of a Boston Hospital went into a coma during anesthesia. Upon investigation she finds that nationally 6 out of 100000 patients go into comas after anesthesia. Assume that the national probability is correct and that the Binomial distribution is appropriate, what is the mean number of patients out of 30000 that will go into a coma? In a hypothesis test, a researcher finds the probability of getting a number as large as or larger than what actually happened. So, find the probability of 10 or more people out of 30000 going into a coma. Obviously, we don’t have a binomial table that will do this. Show that this is an appropriate place to use a Poisson distribution and use the Poisson tables to find the probability that x 10 . On the basis of your result; do you think that something strange is going on? (The answer is shorter than the problem.) g. (Extra credit) A telemarketer finds that the length of a call has an exponential distribution with a mean of 1 2 c minutes. Find the probability that the length of a call will be (i) No more than 3 minutes, (ii) Between 1 and 2 minutes, (iii) More than 5 minutes. h. A coin is tossed 6 times. Define the following events. A : Exactly one head B : Exactly two heads C : At least one head (i) Find P A - Use a table in all three parts of this problem. (ii) Find P B . (iii) Find P C Solution to Question 1 1. Find P10 x 16 for the following distributions (Use tables in d, e, f, h j.): a. Continuous Uniform with c 1, d 14 (Make a diagram!). 1 1 1 .07692 . In the diagram below, shade the area between 10 and 14. (There is d c 14 1 13 no area between 14 and 16.) The horizontal line has a height of 113 . 0 1 10 14 16 14 10 .3077 . 14 1 Another way to do a problem of this type is to remember that for any continuous distribution, we can use differences between cumulative distributions, Pa x b F b F a where the xc cumulative distribution is F x0 Px x0 and F x for c x d , d c F x 0 for x c and F x 1 for x d . If we use the cumulative distribution method, So P10 x 16 P10 x 14 remember F x 1 for x d . So P10 x 16 F 16 F 10 1 10 1 9 1 .3077 . 14 1 13 b. Continuous Uniform with c 10 , d 25 (Make a diagram!). 1 1 1 .066667 . In the diagram below, shade the area between 10 and 16. d c 25 10 15 The horizontal line has a height of 115 . 2 251solngr3-08b 6/25/08 0 (Open this document in 'Page Layout' view!) 10 16 25 16 10 .4000 . If we use the cumulative distribution method, So P10 x 16 25 10 remember F x 1 for x d . So P10 x 16 F 16 F 10 16 10 10 10 6 .4000 . 25 10 25 10 15 c. Continuous Uniform with c 1, d 10 (Make a diagram!). Find P10 x 16 . 1 1 1 . In the diagram below, try to shade the area between 10 and 16. (Surprise! d c 10 1 9 There is no area between 10 and 16.) The probability is thus zero. 0 1 10 16 So P10 x 16 F 16 F 10 1 1 0 . d. Binomial Distribution P10 x 16 with p .35, n 25 . P10 x 16 Px 16 Px 9 .99921 .63031 .3689 e. Binomial Distribution P10 x 16 with p .65, n 25. P10 x 16 can't be done directly with tables that stop at p .5 , so try to do it with failures. (The probability of failure is 1 - .65 = .35.) 10 successes correspond to 25 – 10 = 15 failures out of 25 tries. 16 successes correspond to 9 failures. So try 9 to 14 successes when p .35 and n 25 . P9 x 15 Px 15 Px 8 .99706 .46682 .53078 f. Binomial distribution P10 x 16 with p .03, n 20 (Approximate Solution) You do not have a Binomial table for p .03 . However since n 20 666 .67 500 , you can p .03 use the Poisson distribution for this problem. The mean is m np 20.03 0.6 . If you use the 0.6 part of the Poisson table, you get P10 x 16 Px 16 Px 9 1 1 0 g. Geometric Distribution P10 x 16 with p .35. Remember that F c Px c 1 q c , because success at try c or earlier implies that there cannot have been failures on the first c tries. q 1 p 1 .1 .9 P10 x 16 Px 16 Px 9 F 16 F 9 1 .6516 1 .659 .65 .65 9 16 .02071 .00102 .0197 3 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) h. P10 x 16 for Poisson Distribution with parameter of 9. (Parameter of 9 means 9 m 2) P10 x 16 Px 16 Px 9 .98889 .58741 .40148 i. Show how you would do this for a Hypergeometric Distribution with p .35, n 25 , N 100 . Remember M Np . We have no cumulative tables or cumulative distribution formula for this distribution, so the only method is to add together probabilities over the range 10 to 16. 65 C 35C 25 C M C N M x Px x Nn x and M Np 100 .35 35 so Px x 100 Cn C 25 and P10 x 16 65 C x35C 25 1 x 100 100 C 25 x 10 C 25 16 C 16 45 65 x C 25 x x 10 j. P10 x 16 for a Hypergeometric Distribution with p .35, n 25 , N 1000 . (Approximate Solution). Since 20 n 2025 500 1000 , we can use the Binomial distribution with p .35, n 25 . P10 x 16 Px 16 Px 9 .99921 .63031 .3689 k. For a Hypergeometric Distribution with p .35, n 25 , N 100 , find Px 1 . (Extra credit if you actually evaluate this.) Px 1 1 Px 0 , M Np 100 .35 35 and Px C xM C nNxM C nN . So 1 P0 1 65 C 035 C 25 100 C 25 35! 65! 35!0! 40! 25! 1 100! 75! 25! 65 64 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 100 99 98 97 96 95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 I get 1 .018512 .981488 . This is pretty likely to be right, since it’s close to the Binomial probability. 1 l. (Extra credit) Find P10 x 16 for an exponential distribution with c .01 . In ‘Great Distributions I Have Known, we have the following information. x is usually the Exponential f x ce cx and amount of time 1 1 cx you have to wait F x 1 e c c when x 0 and until a success. the mean time to a 1 success is . Both c are zero if x 0 . Since this is a continuous distribution, P10 x 16 F 16 F 10 1 e 0.16 1 e .10 e .10 e .16 .90484 .85214 .0527 Note: F 10 .09516 and F 16 .14786 . 4 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) Solution to Question 2 2. Find the Mean and Standard deviation for the following distributions. Some of these averages could be expressed in words – try it! (For example, 'The average number of successes when …… is ……') a. Continuous Uniform Distribution with c 10 , d 25 . d c 25 10 225 18.75 c d 10 25 17 .5 , 2 2 2 12 12 12 2 2 So 18.75 4.3301 b. Binomial Distribution with p .65, n 25 . (Remember that q 1 p and that both p and q must be between 0 and 1.) q 1 p 1 .65 .35, np 25.65 16 .25 , 2 npq 16.25 .35 5.6875 so npq 5.6875 2.3848 . The average number of successes in 25 tries, when the probability of a success in an individual try is 65% is 16.25. c. Geometric Distribution with p .35 If q 1 p 1 .35 .65, 1 1 q .65 .65 2.8571 , 2 2 5.3061 2 . 1225 p .35 p .35 so 5.3061 2.3035 . If we play a game repeatedly and the probability of a success in an individual try is 35%, on the average our first success will occur on the 2nd or 3rd try. d. Poisson Distribution with parameter of 9. m 9 , 2 m 9 , so m 9 3 . e. Hypergeometric Distribution with p .35, n 25 , N 100 . M and q 1 p then np 25.35 8.75 and N N n 100 25 75 8.75 .65 .757576 5.6875 4.3087 . npq 25.35 .65 N 1 100 1 99 If p 2 So 4.3087 2.0757 . If we take a sample of 25 from a population of 100 of which 35% are successes, on the average we will get 8.75 successes. f. Compare means and standard deviations for a Binomial distribution with p .35, n 25 and a Hypergeometric Distribution with p .35, n 25 and N 1000 . If p M and q 1 p 1 .35 .65 , then np 25.35 8.75 for both N distributions. For the binomial distribution 2 npq 25.35.65 5.6875 so that 5.6875 2.3848 . For the Hypergeometric distribution, N n 1000 25 975 5.6875 2 npq 25.35 .65 N 1 1000 1 999 0.97598 5.6875 5.5509 . So 0.97598 5.6875 .987915 2.3848 2.3560 . There is a difference of less than 2% between these two. Note that the finite population correction factor in the variance formula has relatively little effect this time, which is why we can justify using the binomial distribution in place of the hypergeometric distribution here. 5 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) g. (Extra credit) Exponential distribution with c .01 , 1 1 1 1 We have 100 and 100 . c .01 c .01 Solution to Question 3 3. Identify the distribution and do the following problems. Remember: (i) If you are looking for numbers of successes when the number of tries is given and the probability of success is constant, you want the Binomial distribution. (ii) If you are looking for the try on which the first success occurs out of many possible tries when the probability of success is constant, you want the Geometric distribution. (iii) If you are looking for numbers of successes when the number of tries is given and the average number of successes per unit time or space is given, you want the Poisson distribution. (iv) If you are looking for numbers of successes when the number of tries is given and the probability of success is not constant because the total number of successes in the population is limited, you want the Hypergeometric distribution. a. This problem uses all four discrete distributions. (i) An IRS auditor is about to audit 6 returns from a group of returns for improper deductions. Assume that these returns come from a group of 80 returns in which 25% have improper deductions, what is the probability that exactly one of the 6 has improper deductions? 20 out of 80 returns are improper. This is Hypergeometric because we want to know the probability of a number of ‘successes, but the total number in the population is limited. C M C N M N 80, n 6, M Np 80.25 20 . Px x nn x and CN 60! 60 59 58 57 56 20 20 60 59 58 57 56 6 55!5! 5 4 3 2 1 P1 .3635 . If we 80 80! 80 79 78 77 76 75 80 79 78 77 76 75 C6 74!6! 6 5 4 3 2 1 wrongly used the Binomial distribution here, we would have found Px 1 when n 6 and C120 C 560 20 p .25. The binomial formula says Px C xn p x q n x , so if p .25, Px 1 C16 p1q 5 6.25 .75 5 .3560 , but we would be better off using a binomial table with n 6 and p .25. P1 Px 1 Px 0 .53394 .17798 .3560 . (ii) What is the probability that at least one of the six has improper deductions? Hypergeometric: 60! 60 59 58 57 56 55 1 C 020 C 660 54 ! 6 ! 6 5 4 3 2 1 Px 1 1 Px 0 1 1 1 80 80 ! 80 79 78 77 76 75 C6 74!6! 6 5 4 3 2 1 60 59 58 57 56 55 1 1 .16660 .8333 . This can be argued more directly if we realize 80 79 78 77 76 75 that on the first try, the probability of not getting an improper return is 60 80 , on the next try the probability becomes 59 79 and so on, so that on the last try the probability is 55 75 . Since these are 60 59 58 57 56 55 conditional probabilities, we can multiply them together to get P0 80 79 78 77 76 75 .16660 . If we had wrongly used the Binomial distribution, we would have found Px 1 when 6 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) n 6 and p .25. If p .25, Px 1 1 P0 1 C 06 p 0 q 6 1 .75 6 .8220 , but we would have been better off using a binomial table with n 6 and p .25. Px 1 1 Px 0 1 .17798 .8220 . (iii) If two or more of the 6 returns have improper deductions, the entire group of 80 will be audited. What is the chance that that happens? Hypergeometric: Px 2 1 Px 0 Px 1 1 .16660 .3635 .8333 .3635 .4698 (iv) Assume that the 6 returns come from a large group of returns (more than 120) that is 25% improper, and if two or more are improper, the whole group will be audited, what is the probability that the whole group will be audited? (Answers that use a specific population size will not be counted.) Binomial: Px 2 when n 6 and p .25. If we use the binomial table with n 6 and p .25, Px 2 1 Px 1 1 .53394 .4661 . (v) Assume that a large number of returns are audited from this group that is 25% improper, what is the chance that the second return is the first one found with improper deductions? Geometric: Px q x1 p . p .25, q 1 p .75 , Px 2 .75 .25 .1875 (vi) What is the chance that the first improper return will be among the first ten audited? Geometric: F c Px c 1 q c , so F 10 Px 10 1 .7510 1 .0531 .9437 . (vii) Assume that only 2% of the returns are improper and a sample of 100 taxpayers is taken, what is the chance that at least 3 have improper returns? Binomial: Px 3 when n 100 and p .02. But we do not have the appropriate table. Unless we want to do the problem by hand, call on the Poisson distribution, since n 100 p .02 5000, which is safely over 500. Our parameter is m np 100 .02 2. Px 3 1 Px 2 1 .67668 .3233 . b. A fast food hamburger must be cooked a minimum of 90 seconds to be safe. At my lunch spot the time that burgers are on can be represented by a continuous uniform distribution with limits of 80 and 120 seconds. (i) What is the chance that a hamburger will be safe? x c Continuous Uniform: The cumulative distribution is F x0 Px x0 0 d c 90 80 10 1 .75 for c x d , c 80 and d 120 . Px 90 1 Px 90 1 120 80 40 (ii) What is the chance that, if I order 2 hamburgers both will be safe? Binomial: Px 2 when n 2 and p .75. The binomial formula says Px C xn p x q n x , so if p .75, P2 C 22 p 2 .75 2 .5625 . . It’s probably not worthwhile to use the binomial table, since we don’t have one for p .75. But, if the probability of success is .75, the probability of failure is .25 and 2 successes is no failures. Sure enough, for n 2 and p .25, P0 .5625 . (iii) What is the chance that if I order 3 hamburgers, at least one will be unsafe? 7 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) Binomial: Px 1 when n 3 and p .25. Note that if the probability of a safe hamburger is .75, the probability of an unsafe one is .25. The binomial formula says Px C xn p x q n x , so if p .25, Px 1 1 P0 1 C 03 p o q 3 1 .75 3 1 .421875 .5781 . It’s probably not worthwhile to use the Binomial table, since we don’t have one for p .75. But, if the probability of success (a safe burger) is .75, the probability of failure is .25 and we want 1 or more failures. So we have the same thing from the table Px 1 1 P0 1 .42187 .5781 . c. The number of students who come to my office has a Poisson distribution with a parameter (mean) of 2.5 a day. (i) What is the chance that no students come in during a day? From the Poisson table with m 2.5 , P0 .08208 . (ii) What is the chance that no students come in during a 5-day week? From the Poisson table with m 52.5 12 .5 , P0 .000004 . (iii) What is the chance that more than 5 come in during a day From the Poisson table with m 2.5 , Px 5 1 Px 5 1 .95798 .0420 . (iv) What is the chance that more than 5 come in during a 5-day week? From the Poisson table with m 12 .5 , Px 5 1 Px 5 1 .01482 .09852 . d. Complaints come in to my website at an average of seven an hour and take an hour to adjust. How many complaint handlers do I need to keep the probability of a complainer having to wait below 1%? Poisson. If we look at the table with parameter of 7, the first cumulative probability above 99% is Px 14 .99428 . We need to hire 14 people. e. If 80% of shoppers at a website abandon their shopping carts before buying and there are currently 20 shoppers on my site, (i) What is the chance that at least one will buy something? Binomial: p .20 , n 20 . Px 1 1 P0 1 .01153 .9885 (ii) What is the average number that will buy something? Binomial: p .20 , n 20 . np 20.2 4. (iii) What is the most likely number that will buy something? P0 Px 0 .01153 P1 Px 1 Px 0 .06918 .01153 .05765 P2 Px 2 Px 1 .20608 .06918 .13690 P3 Px 3 Px 2 .41145 .20608 .20537 P4 Px 4 Px 3 .62965 .41145 .21815 P5 Px 5 Px 4 .80421 .62965 .17456 If we check the rest of the table we find that no difference between subsequent values of the cumulative distribution is more than .15, so 4 seems to be the mode. Since the Binomial distribution is not usually symmetrical for low values of n, we do have to check individual probabilities. We should be able to do this by calculus since Px C xn p x q n x C x20 .2 x .8n x should be differentiable, and the derivative could be set equal to zero, but I can’t remember how to differentiate C xn . One of my manuals says that the mode is any integer value or values between q and p . This seems to work when the mean is not an integer. 8 251solngr3-08b 6/25/08 (Open this document in 'Page Layout' view!) f. According to Bowerman and O’Connell, in the movie Coma a young nurse finds that 10 of the last 30000 patients of a Boston hospital went into a coma during anesthesia. Upon investigation she finds that nationally 6 out of 100000 patients go into comas after anesthesia. Assume that the national probability is correct and that the Binomial distribution is appropriate, what is the mean number of patients out of 30000 that will go into a coma? In a hypothesis test, a researcher finds the probability of getting a number as large as or larger than what actually happened. So, find the probability of 10 or more people out of 30000 going into a coma. Obviously, we don’t have a binomial table that will do this. Show that this is an appropriate place to use a Poisson distribution and use the Poisson tables to find the probability that x 10 . On the basis of your result, do you think that something strange is going on? (The answer is shorter than the problem.) 6 This is a binomial problem. We want Px 10 when p and n 30000 . Note 100000 6 n 30000 1.8 . According that 5 10 8 is above 500. The mean is m np 30000 6 100000 p 100000 to the Poisson (1.8) table Px 10 1 Px 9 1 .99998 .00002 . This is an extremely small probability and should make us suspect that something is wrong. g. (Extra credit) A telemarketer finds that the length of a call has an exponential distribution with a mean of 1 2 minutes. Find the probability that the length of a call will be c (i) No more than 3 minutes, F x 1 ecx . Since 1 c 2 , c 1 2 .5 calls per minute. If x 3, cx .53 1.5 and Px 3 F 3 1 e 1.5 1 .22313 .7769 (ii) Between 1 and 2 minutes, F x 1 ecx If x 1, cx .51 0.5 , if x 2, cx .52 1 and P1 x 2 F 2 F 1 1 e 1 (1 e .0.5 ) e 0.5 e 1 .60653 .36788 .2387 . (iii) More than 5 minutes. If x 5, cx .55 2.5 and Px 5 1 F 5 1 1 e 2.5 e 2.5 .0820. h. A coin is tossed 6 times. Define the following events. A : Exactly one head B : Exactly two heads C : At least one head (i) Find P A - Use a table in all three parts of this problem. (ii) Find PB . (iii) Find PC . This is a binomial problem with p .5 and n 6. (i) P1 Px 1 Px 0 .10938 .01563 .09375 or C16 p 1 q 5 6.56 6 .09375 . 64 6! .56 6 5 1 (ii) P2 Px 2 Px 1 .34375 .10938 .23437 or C 26 p 2 q 4 4! 2! 2 1 64 .234375 1 6 .984375 . (iii) Px 1 1 Px 0 1 .01563 .98437 or 1 C 06 p 0 q 6 1 1.5 1 64 9