Chapter 17 Probability Models 269 b) SD(Y)=.Jnpq = (50.13)(0.87) ©) Use Geom(0.13). Let X = the number of people checked until the first lefty is discovered. i para” People 16. More arrows. .75 lefties E(X)= a) In Exercise 10, we determined that the shots could be considered Bernoulli trials. Since the archer is shooting 6 arrows, use Binom(6,0.80). Let ¥ = the number of bull’s-eyes in n = 6 shots. E(Y) = np = 6(0.80) = 48 bull’s-eyes. 1b) SD(Y)= Jnpq = Y6(0.80)(0.20) = 0.98 bull’s-eyes. ©) Use Geom(0.80). Let X = the number of arrows shot until the first bull’s-eye. 11.25 shots. p 080 17. Still more lefties. a) In Exercise 9, we determined that the selection of lefties (and also righties) could be considered Bernoulli trials. Since our group consists of 12 people, and now we are considering the righties, use Binom(12,00.87) E(X)= Let ¥ = the number of righties among n = 12. E(¥) = np = 12(0.87) = 10.44 righties SD(Y) = «npg = J 12(0.87)(0.13) = 1.16 righties b) P(not all righties) = 1— Pall righties) =1-P(Y=12) 1-(2)0.87)"013" 812 P(no more than 10 righties ) = P(Y < 10) = PUY =0)+ P(Y =1)+ P(Y =2)+...+ P(Y =10) -(?Je. 87/013)" +(?)087)'(013)" + -+()Jos7"0:13) 0475 d) Plexactly six of each ) = P(Y = 6) =(2)o.87013" = 0.00193270 Part IV Randomness and Probability °) P(majority righties ) = PY >7) = P(Y =7)+ P(Y =8)+ P(Y =9)+...+ P(Y =12) =(? 087/013" +(2}osn"0.13) +...+(72}087)" 0.13" =0.998 18. Still more arrows. a) In Exercise 10, we determined that the archer’s shots could be considered Bernoulli trials. Since our archer is now shooting 10 arrows, use Binom(10,0.80). the number of bull’s-eyes hit from n = 10 shots, 10(0.80) =8 bull’s-eyes hit. SD(¥) = s{npq = {10(0.80)(0.20) ~ 1.26 bull’s-eyes hit. b) P(no misses out of 10 shots) = P(all hits out of 10 shots) =P(Y=10) 2\0.80)"0.20)° =(32}0.80"@20% = 0.107 ¢) (no more than 8 hits) = P(Y <8) = P(Y =0)+ PU =1) + PY =2)+... + PUY =8) -(Joso'c 20)" +(?Jo 80)'(0.20)" + --+('2)0.80)"0.20 = 0.624 d) P(exactly 8 out of 10 shots) = P(Y = 8) -(° (0.80)°(0.20)? =0:302 9 P(more hits than misses) = P(Y 26) = PLY =6)+ PLY =7) +...+ PO = 10) -(2\0.80'0.20 + (0.807 0.20" Pet (i)o £80)°(0.20)° =0.967Chapter 17 Probability Models 271 19. Tennis, anyone? The first serves can be considered Bernoulli trials. ‘There are only two possible outcomes, successful and unsuccessful. The probability of any first serve being good is given as p=0.70. Finally, we are assuming that each serve is independent of the others. Since she is serving 6 times, use Binom(6,0.70). Let X= the number of successful serves in » = 6 first serves. a) b) Pall six serves in ) = P(X =6) P(exactly four serves in) = P(X = 4) =(3)0.70%0:30) =(s}o70%0307 =0.118 0324 9 Pat least four serves in) = P(X = 4) + P(X =5)+ P(X =6) = (so 70)*(0.30)? + (Jo -70)*(0.30)" +(9}070» (030 = 0744 d) (no more than four serves in) = P(X = 0)-+ P(X =1) + P(X =2)+ P(X =3)+ P(X=4) = (Sho.70,"00.30° +(%}0.70)'0.30" +($}070700.30)" + (se 70)°(0.30) + (; Jo. 70)*(0.30)? ~ 0.580 20. Frogs. a) The frog examinations can be considered Bernoulli trials. There are only two possible outcomes, having the trait and not having the trait. If the frequency of the trait has not, changed, and the biologist collects a representative sample of frogs, then the probability of a frog having the trait is constant, at p = 0.125. The trials are not independent since the population of frogs is finite, but 12 frogs is fewer than 10% of all frogs. Since the biologist is collecting 12 frogs, use Binom(12,0.125), Let X= the number of frogs with the trait, from n = 12 frogs. P(no frogs have the trait) = P(X =0) -(?)0.25"'0875) = 0.201272 Part IV Randomness and Probability > peat east two fogs) = P22) = P(X =2)+P(X=3)+...+ P(X =12) (Ze 12570875)" +(2}0.125)°0.875) +...+(3025)"0.875) 2 = 0.453 Co} P(three o four frogs have trait) = P(X =3)+ PX = 4) -(2)o 125)*(0.875)" + (2)o225) (0875 ~0171 a) P(no more than four) = P(X $4) = P(X =0)+ P(X =1)+...+ P(X =4) -(2jo 125)°(0.875)"? + (7Joa2sre. 875)" +. _+(?}0.125)*00.875)" = 0.989 21. And more tennis. ‘The first serves can be considered Bernoulli trials. There are only two possible outcomes, successful and unsuccessful. The probability of any first serve being good is given as p=0.70. Finally, we are assuming that each serve is independent of the others. Since she is serving 80 times, use Binom(80,0.70).. Let X = the number of successful serves in n = 80 first serves. a) EX) 80(0.70) = 56 first serves in. SD(X) = npg = (80(0.70)(0.30) = 4.10 first serves in. b) Since np = 56 and ng = 24 are both greater than 10, Binom(80,0.70) may be approximated by the Normal model, N(56, 4.10). ©) According to the Normal model, in matches with 80 serves, she is expected to make between 51.9 and 60.1 first serves approximately 68% of the time, between 47.8 and 64.2 first serves approximately 95% of the time, and between 43.7 First serves made out of 80 56 60.1 642 68.3 and 68.3 first serves approximately 99.7% of the BT 478,519Chapter 17 Probability Models 273 4) Using Binom(80, 0.70): Pat least 65 first serves) = P(X > 65) = P(X =65)+ P(X = 66) +...4 P(X =80) -(2)0.70%0.30" +(2 7090.30)" +...+($ 0.70030)" = 0.0161 According to the Binomial model, the probability that she makes at least 65 first serves out of 80 is approximately 0.0161. Using NG6, 4.10): According to the Normal model, the probability that she makes at least 65 first serves out of 80 is approximately 0.0141. First serves ‘made out of 80 252.195 P(X 265) ~ Pe> 2.195) = 0.0141 22, More arrows. ‘These may be considered Bernoulli trials, There are only two possible outcomes, hitting the bull’s-eye and not hitting the bull’s-eye. The probability of hitting the bull’s-eye is given, p = 0.80. The shots are assumed to be independent. Since she will be shooting 200 arrows, use Binom(200, 0.80). Let ¥ = the number of bull’s-eyes in = 200 shots. = 200(0.80) = 160 bull’s-eyes. SD(Y) = Jnpq = {200(0.80)(0.20) =5.66 bull's-eyes. b) Since np = 160 and ng = 40 are both greater than 10, Binom(200,0.80) may be approximated by the Normal model, N(160, 5.66). ©) According to the Normal model, in matches with 200 arrows, she is expected to get between 154.34 Builtscyes and 165.66 bull’s-eyes approximately 68% of the | 200 shots. time, between 148.68 and 171.32 bull’s-eyes 68% approximately 95% of the time, and between 143.02 and 176.98 bull’s-eyes approximately 99.7% Say waa wu lw ioe Wn Te of the time. « 9.7% »274 4) Part IV Randomness and Probability Using Binom(200, 0.80): P(at most 140 hits) = P(Y < 140) = PUY =0)+ PW =1) +... + PLY = 140) =( 80020 +(**}o8'020)" +...+(%)as0)"*a70/* = 0.0005 According to the Binomial model, the probability that she makes at most 140 bull’s-eyes ‘out of 200 is approximately 0.0005. Using N(160, 5.66): According to the Normal model, the probability that she hits at most 140 bull’s-eyes out of 200 is approximately 0.0002. 140-160 5.66 ge -3.534 200 shots 0.0002 140 160 22-3504 POY S$ 200) = P(z <-3.534) = .0002 Using either model, it is apparent that it is very unlikely that the archer would hit only 140 bull’s-eyes out of 200. 23, Frogs, part II. a) b) The frog examinations can be considered Bernoulli trials. ‘There are only two possible outcomes, having the trait and not having the trait. If the frequency of the trait has not changed, and the biologist collects a representative sample of frogs, then the probability of a frog having the trait is constant, at p = 0.125. The trials are not independent since the population of frogs is finite, but 150 frogs is fewer than 10% of all frogs. Since the biologist is collecting 150 frogs, use Binom(150,0.125). Let X= the number of frogs with the trait, from n = 150 frogs. E(X) = np = 150(0.125) = 18.75 frogs. SD(X) = npg = f750(0-125)(0.875) = 4.05 frogs. Since np = 18.75 and ng = 131.25 are both greater than 10, Binom(200,0.125) may be approximated by the Normal model, N(18.75, 4.05).Chapter 17 Probabilly Models 275 ), Using Binom(150, 0.125): Plat least 22 frogs) = P(X > 22) = PX = 22) +... P(X =150) =(:2)0.225/%10875) +...+(2)o125)* 0.875) = 0.2433 According to the Binomial model, the probability that at least 22 frogs out of 150 have the trait is approximately 0.2433, Using N(18.75, 4.05): P(X 2 22) = P(z> 0.802) = 0.2111 o 2-18.75 ‘ee 02111 According to the Normal 4.05 model, the probability that at 2~0.802 Teast 22 frogs out of 150 have ae the trait is approximately = 0,802 0211 Using either model, the probability that the biologist discovers 22 of 150 frogs with the trait simply as a result of natural variability is quite high. This doesn’t prove that the trait has ‘become more common. 24. Apples. a) A binomial model and a normal model are both appropriate for modeling the number of cider apples that may come from the tree. Let X = the number of cider apples found in the n = 300 apples from the tree. ‘The quality of the apples may be considered Bernoulli trials, There are only two possible outcomes, cider apple or not a cider apple. The probability that an apple must be used for a cider apple is constant, given as p = 0.06. The trials are not independent, since the population of apples is finite, but the apples on the tree are undoubtedly less than 10% of all the apples that the farmer has ever produced, so model with Binom(300, 0.06). E(X) = np = 300(0.06) = 18 cider apples. SD(X) = .npq = 300(0.06)(0.94) = 4.11 cider apples. Since np = 18 and ng = 282 are both greater than 10, Binom(300,0.06) may be approximated by the Normal model, N(18, 4.11)276 Part IV Randomness and Probability b) Using Binom (300, 0.06): P(at most 12 cider apples) = P(X $12) = P(X =0)+...+ P(X =12) =(P footy ++ (20.0604 = 0.085 According to the Binomial model, the probability that no more than 12 cider apples come from the tree is approximately 0.085. Using N(18, 4.11): P(X $12) = P(z<-1.460) = 0.072 Cider apples According to the Normal ipa eof model, the probability that no PP more than 12 apples out of 300 are cider apples is 12 13 SSS~S*CpProximnately 0.072. z= 1.460 ©) It is extremely unlikely that the tree will bear more than 50 cider apples. Using the Normal model, N(18, 4.11), 50 cider apples is approximately 7.8 standard deviations above the mean. 25, Lefties again. Let X = the number of righties among a class of n = 188 students. Using Binom(188, 0.87): These may be considered Bernoulli trials. There are only two possible outcomes, right- handed and not right-handed. ‘The probability of being right-handed is assumed to be constant at about 87%. The trials are not independent, since the population is finite, but a sample of 188 students is certainly fewer than 10% of all people. Therefore, the number of righties in a class of 188 students may be modeled by Binom(188, 0.87). If there are 171 or more righties in the class, some righties have to use a left-handed desk. P(at least 171 righties) = P(X 2171) = P(X=171)+...+ P(X = 188) -(S)o 87)'"(0.13)" + +(S)os7*e 13)° Flan = 0.061 According to the binomial model, the probability that a right-handed student has to use a left-handed desk is approximately 0.061.78 Part IV Randomness and Probability 2-2 P(X > 266) ~ P(e> 1.316) = 0.0941 266 261.25 Saatiune ‘According to the Normal 223g AFP booked) model, the probability that at #1316 least 266 passengers show up at is approximately 0.0941. z= 1316 27, Annoying phone calls. Let X= the number of sales made after making 1 = 200 calls. Using Binom(200, 0.12): These may be considered Bernoulli trials. There are only two possible outcomes, making a sale and not making a sale. The telemarketer was told that the probability of making a sale is constant at about p = 0.12. The trials are not independent, since the population is finite, but 200 calls is fewer than 10% of all calls. Therefore, the number of sales made after making 200 calls may be modeled by Binom(200, 0.12), P(at most 10) = P(X $10) = P(X =0)+...+ P(X =10) -(2 oa" say +...+(™)o.12)"(088)" = 0.0006 According to the Binomial model, the probability that the telemarketer would make at most 10 sales is approximately 0.0006. Using N(24, 4.60): E(X) = np = 200(0.12) = 24 sales. SD) = sfnpq = 200(0.12)(0.88) = 4.60 sales. Since np = 24 and ng = 176 are both greater than 10, Binom(200,0.12) may be approximated by the Normal model, N(24, 4.60). P(X < 10) ~ P(g <—3.043) = 0.0012 ‘According to the Normal model, the probability that the telemarketer would make at ‘most 10 sales is approximately 0.0012. 22-3083 Since the probability that the telemarketer made 10 sales, given that the 12% of calls result in sales is so low, it is likely that he was misled about the true success rate.Chapter 17 Probability Moo 2 Using N(163.56, 4.61): BCX) = np = 188(0,87) = 163.56 righties SPO) = npg = /T880.87V03} ~ 4.61 rights Since np = 163.56 and ng = 24.44 are both Breater than 10, Binom(188,0.87) may be “pproximated by the Normal model, N(163.56, 461) P(X2 171) = P(z> 1.614) = 0.08 According to the Normal model, the probability that there are at least 171 tighties in the class of 188 is approximately 0.0533 the class hee 0.0533 26. No-shows. Let X= the number of passengers that show up for the flight of n = 275 passengers, Using Binom (275, 0.95): These may be considered Bernoulli trials. There are only two possible outcomes, showing, up and not showing up. The airlines believe the probability of showing up is constant at about 95%. The trials are not independent, since the Population is finite, but a sample of 275 passengers is certainly fewer than 10% of all Passengers. Therefore, the number of Passengers who show up for a flight of 275 may be modeled by Binom(275, 0.95). 1£266 or more passengers show up, someone has to get bumped off the flight. Plat least 266 passengers) = P(x > 266) P(X = 266) +... P(X = 275) approximately 0.116, Using N(261.25, 3.61): EXO) = np = 275(0.95) = 261.25 passengers, SDD = npg = (2750.95 \(0.05) Since np = 261.25 and ng = 13.75 are both greater than 10, Binom(275,0.95) may beChapter 17 Probability Models 279 28. The euro. Let X = the number of heads after spinning a Belgian euro n = 250 times, Using Binom(250, 0.5): These may be considered Bernoulli trials. ‘There are only two possible outcomes, heads and tails. The probability that a fair Belgian euro lands heads is p = 0.5. The trials are independent, since the outcome of a spin does not affect other spins, Therefore, Binom(250, 0.5) may be used to model the number of heads after spinning a Belgian euro 250 times. P(at least 140) = P(X 2140) = P(X = 140) +...+ P(X = 250) -()o. 50.5)" +... +(Rjoso 5 0.0332 According to the Binomial model, the probability that a fair Belgian euro comes up heads at least 140 times is 0.0332, Using N(125, 7.91): E(X) = np = 250(0.05 SD(X) = npg = 250(05)(05) = 7.91 heads. Since np = 125 and nq = 125 are both greater than 10, Binom(250,0.5) may be approximated by the Normal model, N(125, 7.91). = 125 heads. P(X 2 140) = P(z > 1.896) = 0.0290 Number of bead ator 0.0290 According to the Normal 0-15” model, the probability that a 751 fair Belgian euro lands heads 21.896 aaa 13 140 at least 140 out of 250 spins is 2= 1.896, approximately 0.0290. Since the probability that a fair Belgian euro lands heads at least 140 out of 250 spins is low, itis unlikely that the euro spins fairly. However, the probability is not extremely low, and we aren’t sure of the source of the data, so it might be a good idea to spin it some more.280 Part IV Randomness and Probability 29. Seatbelts II. °) ‘These stops may be considered Bernoulli trials. There are only two possible outcomes, belted or not belted. Police estimate that the probability that a driver is buckled is 80%. (The probability of not being buckled is therefore 20%.) Provided the drivers stopped are representative of all drivers, we can consider the probability constant. The trials are not independent, since the population of drivers is finite, but the police will not stop more than 10% of all drivers. Let X= the number of cars stopped before finding a driver whose seat belt is not buckled. Use Geom(0.2) to model the situation. a t.5 e700 P@irst unbelted driver is in the sixth car) = P(X = 6)= (0.8)°(0.2) ~ 0.066 E(x) P(The first ten drivers are wearing seatbelts) = (0. 8)" = 107 Let Y= the number of drivers wearing their seatbelts in 30 cars. Use Binom(30, 0.8). E(Y) = np = 30(0.8) = 24 drivers. SD(Y) = npg = ¥30(0.8)(0.2) = 2.19 drivers Let W = the number of drivers not wearing their seatbelts in 120 cars. Using Binom(120, 0.2): P(at least 20) = P(W2 20) ‘According to the Binomial =PW= - model, the probability that at eee Jeast 20 out of 120 drivers are =(:2)o2y"08)"+.. +(Zyoa™@s? — notwearing thei seatbelts is , 20, approximately 0.848. = 0.848 Using N(24, 4.38): E(W)= np= 1200.2) SDW) = [npg = {120(0.2)(0.8) = 4.38 drivers. Since np = 24 and ng =96 are both greater than 10, Binom(120,0.2) may be approximated by the Normal model, N(24, 4.38). PCW 2120) = P(g >-0.913) = 0.8194 Daivers with o.io4 According to the Normal no seatbelts model, the probability that at Teast 20 out of 120 drivers stopped are not wearing their seatbelts is approximately 0.8194.Chapter 17 Probability Models 281 30. Rickets. The selection of these children may be considered Bernoulli trials. There are only two possible outcomes, vitamin D deficient or not vitamin D deficient. Recent research indicates that 20% of British children are vitamin D deficient. (The probability of not being vitamin D deficient is therefore 80%.) Provided the students at this school are Tepresentative of all British children, we can consider the probability constant. The trials are not independent, since the population of British children is finite, but the children at this school represent fewer than 10% of all British children. Let X= the number of students tested before finding a student who is vitamin D deficient, Use Geom(0.2) to model the situation. (First vitamin D deficient child is the eighth one tested) = P(X = 8) = (0.8)' (0.2) ~ 0.042 P(The first ten children tested are okay) = (0.8)"" = 107 daael —~= 5 kide. 70375 Kids Let Y= the number of children who are vitamin D deficient out of 50 children. Use Binom(50, 0.2) E(Y)= np =50(0.2)=10 children. SD) = Jnpq = 50(0.2)(0.8) ~ 2.83 children. Using Binom(320, 0.2 P(no more than 50 children have the deficiency) = P(X <50) =P(X=0)+...4 P(X =50) -(2)e 2°08) +...+(Z 02s” =0.027 According to the Binomial model, the probability that no more than 50 of the 320 children have the vitamin D deficiency is approximately 0.027. Using N(64, 7.16): E(X)= E(¥)= np = 320(0.2) = 64 children, SD(Y) = npq = ¥320(0.2)(0.8) ~7.16 children. Since np = 64 and ng = 256 are both greater than 10, Binom(320,0.2) may be approximated by the Normal model, N(64, 7.16).282 Part IV Randomness and Probability P(Y $50) = P(g <-1.955) = 0.0253 Children wih According to the Normal ygun? model, the probability that no (of320) more than 50 out of 320 children have the vitamin D deficiency is approximately 0.0253. 276 00088 31. ESP. Choosing symbols may be considered Bernoulli trials. There are only two possible outcomes, correct or incorrect. Assuming that ESP does not exist, the probability of a correct identification from a randomized deck is constant, at p = 0.20. The trials are independent, as long as the deck is shuffled after each attempt. Since 100 trials will be performed, use Binom(100, 0.2). Let X = the number of symbols identified correctly out of 100 cards. E(X) = np = 100(0.2) = 20 correct identifications. 10(0.2)(0.8) = 4 correct identifications. SD(X)= Ynpq = 1 Answers may vary. In order be convincing, the “mind reader” would have to identify at least 32 out of 100 cards correctly, since 32 is three standard deviations above the mean. Identifying fewer cards than 32 could happen too often, simply due to chance. 32, True-False. Guessing at answers may be considered Bernoulli trials. ‘There are only two possible outcomes, correct or incorrect. If the student was guessing, the probability of a correct response is constant, at p= 0.50. The trials are independent, since the answer to one question should not have any bearing on the answer to the next. Since 50 questions are on the test use Binom(500, 0.5). Let X = the number of questions answered correctly out of 50 questions. EX SD(X) = \npq = (50(0.5)(0.5) = 3.54 correct answers. np =50(0.5) = 25 correct answers, “Answers may vary. In order be convincing, the student would have to answer at least 36 out of 50 questions correctly, since 36 is approximately three standard deviations above the mean, Answering fewer than 36 questions correctly could happen too often, simply due to chance. 33. Hot hand. ‘Astreak like this is not unusual. ‘The probability that he makes 4 in a row with a 55% free throw percentage is (0.55\(0.55)(0.55)(0.55)~ 009. We can expect this to happen nearly one in ten times for every set of 4 shots that he makes. One out of ten times is not that unusual.Chapter 17 Probability Models 283 34, New bow. A streak like this is not unusual. The probability that she makes 6 consecutive bulls-eyes with an 80% bulls-eye percentage is (0.8)(0.8)(0.8)(0.8)(0.8)(0.8) = 0.26. If she were to shoot several flights of 6 arrows, she is expected to get 6 bulls-eyes about 26% of the time. An event that happens due to chance about one out of four times is not that unusual. 35. Hotter hand. The shots may be considered Bernoulli trials. There are only two possible outcomes, make or miss. The probability of success is constant at 55%, and the shots are independent of one another. Therefore, we can model this situation with Binom(32, 0.55). Let X = the number of free throws made out of 40. E(X)= np = 40(0.55) = 22 free throws made. SD(X)= npg = [40(0.55)(0.45) ~ 3.15 free throws. Answers may vary. The player's performance seems to have increased. 32 made free throws is (32-22) /3.15 ~3.17 standard deviations above the mean, an extraordinary feat, unless his free throw percentage has increased. This does NOT mean that the sneakers are responsible for the increase in free throw percentage. Some other variable may account for the increase. The player would need to set up a controlled experiment in order to determine what effect, if any, the sneakers had on his free throw percentage. 36. New bow, again. The shots may be considered Bernoulli trials. There are only two possible outcomes, hit or miss the bulls-eye. ‘The probability of success is constant at 80%, and the shots are independent of one another. Therefore, we can model this situation with Binom(50, 0.8). Let X = the number of bulls-eyes hit out of 50. E(X)= np = 50(0.8) = 40 bulls-eyes hit. SD(X) = .Jnpq = {50(0.8)(0.2) ~ 2.83 bulls-eyes. Answers may vary. The archer’s performance doesn’t seem to have increased. 45 bulls-eyes is (45 ~ 40) /2.83 = 1.77 standard deviations above the mean, This isn’t unusual for an archer of her skill level.284 Part IV Randomness and Probability Review of Part IV 1. Quality Control. Construct a Venn diagram of the disjoint outcomes, a) P(defect) = P(cosm,) + P(func.) ~ P(cosm. N func.) = 0.29 + 0.07 - 0.02 = 0.34 Or, from the Venn: 0.27 + 0.02 + 0.05 = 0.34 b) P(cosm. [| no func.) = P(cosm.) - P(cosm. 1 func.) ©) P(fune. | cosm.) Pee Deen) From the Venn, consider only the region inside the Cosmetic circle. The probability that the car has a functional defect is 0.02 out of a total of 0.29 (the entire Cosmetic circle). 4) ‘The two kinds of defects are not disjoint events, since 2% of cars have both kinds. ©) Approximately 6.9% of cars with cosmetic defects also have functional defects. Overall, the probability that a car has a cosmetic defect is 7%. The probabilities are estimates, so these are probably close enough to say that they two types of defects are independent, 2. Workers. Male | Female J Total Organize the counts in a two-way table. Management | 7 6 Supervision | 8 12 Production | 45 | 72 Total 60 150 Type 2 2% =06 150 a) i) P(female) = ii) ) P(female U production) = P(female) + P(production)-P (female () production) 90 117 72 =750"150” 150°” i) Consider only the production row of the table. There are 72 women out of 117 production workers. 72/117 = 0.615. Or, use the formula: (female 1) production) 72450 aie P(production) a 7/59 P(female | production) =Review of Part IV 285 iv) Consider only the female column. There are 72 production workers out of a total of 90 women. 72/90 = 0.8. Or, use the formula: P(production 1) female) _ 73459 _ ine ‘P(female) % 59 b) These data suggest that holding a production position may be associated with gender. P( production | female) = 60% of the plant employees are women, but 61.5% of the production workers are women, However, this is a small difference, and may be due to sampling error. 3. Airfares. a) Let C= the price of a ticket to China Let F = the price of a ticket to France. Total price of airfare = 3C + 5F b) w= EGC +5F)=3E(C)+5E(F) = 3(1000)+ 5(500) = $5500 = SDC +5P) = 3" (War(C))+ 5° Var(F)) = 37150") + 5°(100") ~ $672.68 9 H=E(C-F)= E(C)~ E(F)= 1000-500 = $500 0 =SD(C- F) = \Var(O)+ Var) = ¥150" +100? = $180.28 d) No assumptions are necessary when calculating means. When calculating standard deviations, we must assume that ticket prices are independent of each other for different countries but all tickets to the same country are at the same price. 4, Bipolar. Let X= the number of people with bipolar disorder in a city of 1 = 10,000 residents. ‘These may be considered Bernoulli trials. There are only two possible outcomes, having bipolar disorder or not having bipolar disorder. Psychiatrists estimate that the probability that a person has bipolar is about 1 in 100, so p = 0.01. We will assume that the cases of bipolar disorder are distributed randomly throughout the populations, The trials are not independent, since the population is finite, but 10,000 people represent fewer than 10% of all people. Therefore; the number of people with bipolar disorder in a city of 10,000 may be modeled by Binom(10000, 0.01). Since np = 100 and rq = 9900 are both greater than 10, Binom(10000,0.01) may be approximated by the Normal model, N(100, 9.95). E(X)= np = 10,000(0.01) = 100 residents. SD(X)= Jnpq = /10,0000.01)0.99) ~ 9.95 residents.286 Part IV Randomness and Probability We expect 100 city residents to have bipolar disorder. According to the Normal model, 200 cases would be over 10 standard deviations above this mean. The probability of this occurring is essentially zero. Technology can compute the probability according to the Binomial model. Again, the probability that 200 cases of bipolar disorder exist in the city is essentially zero. We use the Normal model in this case, since it gives us a more intuitive idea of just how unlikely this event is. 5. Agame. a) Let X=net amount won xX | $0 | #2 [Pe | 010 | 0.40 H= E(X) = 0(0.10) + 2(0.40) ~ 2(0.50) =-$0.20 0? = Var(X) = (0~ (-0.20))?(0.10) + (2 = (-0.20))* (0.40) + (-2— (-0.20))*(0.50) = 3.56 D(X) = \Var(X) = 3.56 = $1.89 b) X+X-= the total winnings for two plays, (X + X) = E(K) + EOD = (0.20) + (0.20) = -$0.40 Var(X) + VartX) = ¥3.56+3.56 = $2.67 6. Emergency switch. He o=SDX+X) Construct a Venn diagram of the disjoint outcomes a) From the Venn diagram, 3% of the workers were unable to operate the switch with either hand i P( left right) 0.51 b) Pet sigh) = USE oe About 62% of the workers who could operate the switch with 3% .622, their right hands could also operate it with left hands. Overall, the probability that a worker could operate the switch with his right hand was 66%. Workers who could operate the switch with their right hands were less likely to be able to operate the switch with their left hand, so success is not independent of hand. ) Success with right and left hands are not disjoint events. 51% of the workers had success with both hands.Review of PartIV 287 7. Twins. The selection of these women can be considered Bernoulli trials. There are two possible outcomes, twins or no twins. As long as the women selected are representative of the population of all pregnant women, then p = 1/90. (If the women selected are representative of the population of women taking Clomid, then p = 1/10.) The trials are not independent since the population of all women is finite, but 10 women are fewer than 10% of the population of women. Let X = the number of twin births from = 10 pregnant women. Let Y = the number of twin births from n = 10 pregnant women taking Clomid, a) Use Binom(10, 1/90) Pfat least one has twins) = 1— P(none have twins) =1-P(X=0) ~-(Sha(8) 1 P(none have twins) =1-P(Y =0) wy 1ayr9y° (0) Gs) =0.651 ©) Use Binom(5, 1/90) and Binom(S, 1/90). P(at least one has twins) = 1— P(no twins without Clomid)P(no twins with Clomid) (ea) Ga) a) a) o}90) (90) Loo) (20 = 0.442 b) Use Binons(10, 1/10) Pat least one has twins) 8. Deductible. = E(cost) = 500(0.005) = $2.50 0 = Var(cost) = (2.50 500)°(0.005)+ (2.50 —0)*(0.995) = 1243.75, o = SD(cost) = War(cost) = V 1243.75 = $35.27 Expected (extra) cost of the cheaper policy with the deductible is $2.50, much less than the $12 surcharge for the policy with no deductible, so on average she will save money by going with the deductible. The standard deviation, at $35.27, is quite high compared to the $12 surcharge, indicating a high amount of variability. The value of the car shouldn't influence the decision.288 Part IV Randomness and Probability 9. More twins. In Exercise 7, it was determined that these were Bernoulli trials. Use Binom(5, 0.10) Let X = the number of twin births from n = 5 pregnant women taking Clomid. a) ») P(none have twins) = P(X =0) P(exactly one has twins) = P(X = 1) = (Jonesy = (ore 9 = 0.590 = 0328 ¢ P(at least three will have twins) = P(X =3)+ P(X = 4)+ P(X = 5) = (Jose oy + (Jon 0.9) +(Jc 150.9) = 0.00856 10. At fault. If we assume that these drivers are representative of all drivers insured by the company, then these insurance policies can be considered Bernoulli trials. There are only two possible outcomes, accident or no accident. The probability of having an accident is constant, p = 0.005. The trials are not independent, since the populations of all drivers is finite, but 1355 drivers represent fewer than 10% of all drivers. Use Birom(1355, 0.005). a) Let X = the number of drivers who have an at-fault accident out of n = 1355. E(X) = np = 1,385(0.005) =6.75 drivers. SD(X) = Jnpq = J1,355(0.005)(0.995) ~ 2.60 drivers. b) Since np = 6.775 < 10, the Normal model cannot be used to model the number of drivers who are expected to have accidents. The Success/ Failure condition is not satisfied. 11. Twins, part IIL. In Exercise 7, it was determined that these were Bernoulli trials. Use Binom(152, 0.10). Let X= the number of twin births from n = 152 pregnant women taking Clomid. a) E(X)=np = 152(0.10)= 15.2 births. SD(X)= Jnpq = J152(0-10)(0.90) = 3.70 births. b) Since np = 15.2 and nq = 136.8 are both greater than 10, the Success /Failure condition is satisfied and Binom(152, 0.10) may be approximated by N(15.2, 3.70).Review of PartIV 289 ©) Using Binom(152, 0.10): P(no more than 10) = P(X < 10) =P(X=0)+...+ PX =10) =(3} 02090090) +...+(%}01"@90" = 0.097 According to the Binomial model, the probability that no more than 10 women would have twins is approximately 0.097. Using N(15.2, 3.70): P(X $10) = Pz < ~1.405) = 0.080 Twin binhs According to the Normal ‘women model, the probability that no more than 10 women would have twins is approximately 0.080. 10 15.2 z= 1.405 12. Child’s play. a) Let X = the number indicated on the spinner X | 5 | 10 | 20 P(X) | 05 | 0.25 | 0.25 b) 1= E(X) =5(0.5) + 10(0.25) + 20(0.25) = 10 6° = Var(X) = (5-10)'(0.5) + (10 - 10)*(0.25) + (20 10)*(0.25) = 37.5, o = SD(X) = War(X) = V375 = 6.12 ©) Let ¥ = the number indicated on the die y Jolal2]3]4 a}ajajasa P ee eee | (ls |elele|s # rior( Jee) oGfe wo B62) IHE-BIOHO-RICMMG) 0 = SDV) = WWar(¥) = v2.22 =1.49 e) H=E(X+Y)=E(X)+ EW) = 1041.67 = 11.67 spaces 0 = SD(K+Y)= \VarlX)+ Var) = ¥375+2.22 = 6.30 spaces290 Part IV Randomness and Probability 13, Language. Assuming that the freshman composition class consists of 25 randomly selected people, these may be considered Bernoulli trials. There are only two possible outcomes, having a specified language center or not having the specified language center. The probabilities of the specified language centers are constant at 80%, 10%, or 10%, for right, left, and two- sided language center, respectively. ‘The trials are not independent, since the population of people is finite, but we will select fewer than 10% of all people. a) Let L= the number of people with left-brain language control from n = 25 people. Use Binom(25, 0.80). P(no more than 15) = P(L $15) = P(L=0)+...+ P(L=15) fe 80)°(0.20)" +. +(Zesn%e 20)" = 0.0173 According to the Binomial model, the probability that no more than 15 students in a class. of 25 will have left-brain language centers is approximately 0.0173. b) Let T= the number of people with two-sided language control from n = 5 people. Use Binom(5, 0.10). P(none have two-sided language control) = P(T =0) 7 (eo 10)'(0.90)° =0.590 ©) Use Binomial models: E(t) = np, = 1200(0.80) = 960 people Eight) = np, = 1200(0.10) = 120 people E(two- sided) = np, = 1200(0.10) = 120 people d) Let R = the number of people with right-brain language control. E(R) = npg = 1200(0.10)= 120 people SD(R)= \/nppdq = ¥1200(0.10)(0.90) ~ 10.39 people.Review of Part IV 291 €) Since np, =120 and ng, = 1080 are both greater than 10, the Normal model, N(120, 10.39), may be Nigh brain used to approximate Birton(1200, 0.10). According Soo to the Normal model, about 68% of randomly Aa selected groups of 1200 people could be expected to have between 109,61 and 130.39 people with = right-brain language control. About 95% of Bee aa aneaee randomly selected groups of 1200 people could be expected to have between 99.22 and 140.78 people with right-brain language control. About 99.7% of, oe. randomly selected groups of 1200 people could be expected to have between 88.83 and 151.17 people with right-brain language control. 14, Play again. N= E(X—Y)= E(X)- EY) = 10- 1.67 = 8.33 spaces D(X -¥) = \Var(X)+Var(¥) = V375+2.22 = 6.30 spaces 15. Beanstalks. a) The greater standard deviation for men’s heights indicates that men’s heights are more variable than women’s heights. b) Admission to a Beanstalk Club is based upon extraordinary height for both men and women, but men are slightly more likely to qualify. The qualifying height for women is about 2.4 standard deviations above the mean height of women, while the qualifying height for men is about 1.75 standard deviations above the mean height for men. ©) Let M= the height of a randomly selected man from N(69.1, 2.8). Let W=the height of a randomly selected woman from N(64.0, 2.5). M-W- the difference in height of a randomly selected man and woman. d) E(M-W)= E(M)~ E(W)=69.1-64.0=5.1 inches e) SD(M -W)= \Var(M)+ Var(W) = 2.8" + 2.5" = 3.75 inches f) Since each distribution is described by a Normal model, the distribution of the difference in height between a randomly selected man and woman is N(5-1, 3.75). According to the Normal p= 0-51 model, the probability that a “ ¥287 4257 randomly selected man is 1.359 taller than a randomly selected woman is approximately 0.913.292 Part IV Randomness and Probability 2) If people chose spouses independent of height, we would expect 91.3% of married couples to consist of a taller husband and shorter wife. The 92% that was seen in the survey is close to 91.3%, and the difference may be due to natural sampling variability. Unless this survey is very large, there is not sufficient evidence of association between height and choice of partner. 16. Stocks. a) (market will rise for 3 consecutive years)=(0.73) = 0.389 b) Use Binom(5, 0.73). (market will rise in 3 out of 5 years) = (lozave. 27) = 0.284 ©) P(fall in at least 1 of next 5 years) = 1 ~ P(no fall in 5 years) = 1—(0.73)° = 0.793 4) Let X = the number of years in which the market rises. Use Binom(10, 0.73). P(tises in the majority of years in a decade)= P(X 2.6) = PUX = 6) +... P(X = 10) -( so 73) 0.27) +... (") (0.73)"0.27" =0.896 17. Multiple choice. Guessing at questions can be considered Bernoulli trials. There are only two possible outcomes, correct or incorrect. If you are guessing, the probability of success is p = 0.25, and the questions are independent. Use Binom(50, 0.25) to model the number of correct guesses on the test a) Let X= the number of correct guesses. P{at least 30 of 50 correct) = P(X 230) = P(X =30)+...+ P(X =50) ‘) (0.25)"°(0.75) +... (3 (0.25) (0.75) = 000000016 ‘You are very unlikely to pass by guessing on every question. b) Use Binom(50, 0.70). P(at least 30 of 50 correct) = P(X 230) = P(X =30)+...+ P(X =50) =(2°)(0.70) (0.30) +... +[3}(0.70)" 0.30)" 20 51 = 0.952 According to the Binomial model, your chances of passing are about 95.2%. «) Use Geom(0.70), P(first correct on third question) = (0.30)*(0.70) = 0.063Review of Part IV 293 18. Stock strategy. a) This does not confirm the advice. Stocks have risen 75% of the time after a two-year fall, but there have only been eight occurrences of the two-year fall. The sample size is very small, and therefore highly variable, b) Stocks have actually risen in 73% of years. This is not much different from the strategy of the advisors, which yielded a rise in 75% of years (from a very small sample of years.) 19. Insurance. The company is expected to pay $100,000 only 2.6% of the time, while always gaining $520 from every policy sold. When they pay, they actually only pay $99,480, E(profit) = $520(0.974) - $99,480(0.026) = - $2,080. ‘The expected profit is actually a loss of $2,080 per policy. The company had better raise its premiums if it hopes to stay in business. 20. Teen smoking. Randomly selecting high schoo! students can be considered Bernoulli trials. ‘There are only two possible outcomes, smoker or nonsmoker. The probability that a student is a smoker is p= 0.30, The trials are not independent, since the population is finite, but we are not sampling more than 10% of all high school students. a) P(none of the first 4 are smokers) = (0.7)* = 0.2401 b) Use Geom (03). P(first smoker is the sixth person) = (0.7)°(0.3) ~ 0.050 ©) Use Binom(10, 0.3). Let X = the number of smokers among n= 10 students, (no more than 2 smokers of 10) = P(X <2) = P(X =0)+ P(X =1)+ P(X =2) -(sJe 30)°(0.70)" + ("| (0.30)*(0.70)" + (le 30)°(0.70)° ~0.383 21. Passing stats. Pass (7020/08) «0.4667 Organize the information in a tree 08 diagram. a) Scedastc P(Passing Statistics) 7an20 o2 = P(Scedastic ( Pass) Fa (7ortzoyo.2) «0.1167 + P(Kurtosis 1) Pass) = 0.4667 +0.25 =0.717 Pass (60/120)(0.6)=0.25, orto os Kurtis os Fail (60/20/04) ~ 0.1667294 Part IV Randomness and Probability b) P(Kurtosis | Fail) _ P(Kurtosis () Fail) a P(Fail) 0.1667 ~ 0.41167 +0.1667 22, Teen smoking II. In Exercise 22, it was determined that the selection of students could be considered to be Bernoulli trials. = 0.588 a) Use Binom(120, 0.30) to model the number of smokers out n = 120 students. E(number of smokers) = np = 120(0.30) = 36 smokers. b) SD(number of smokers) = /npq = y'120(0.30)(0.70) = 5.02 smokers. ©) Since np = 36 and ng = 84 are both greater than 10, the Success/Failure condition is satisfied and Birom(120, 0.30) may be approximated by N(36, 5.02). 4) According to the Normal model, approximately 68% of samples of size mt = 120 are expected to have between 30.98 and 41.02 smokers, approximately 95% of the samples are expected to have between 25.9 and 46.04 smokers, and approximately 99.7% of the samples are expected 2094 25.95,3098 36 41.02 4604 51.06 to have between 20.94 and 51.06 smokers. Number of smokers ‘out oF 120 23. Random variables. a) b) = E(X +50) = E(X)+50=50+50= 100 o = SD(X +50)= SD(X)=8 o (LOY) = 10£(¥) = 10(100) = 1000 D(LOY )= LOSD(¥ )= 60 B= E(X+05Y)= E(X)+0.5E(Y) =50+0.5(100)= 100 M=E(X-Y)= E(X)- E(Y) = 50-100=-50 = SD(X-Y)= \VarX+ Var) = SD(X-+0.5Y) = War(x) +05 Vari) ee = PHOS) = 8.54 ec Me E(X, + Xz) = E(X)+ E(X)= 50+ 50 = 100 0 = SD(X, + X,) = yWar(X) + Var(X) = V8? +8? = 11.31Review of Part IV 295 24. Merger. Small companies may run in to trouble in the insurance business. Even if the expected Profit from each policy is large, the profit is highly variable. There is a small chance that a company would have to make several huge payouts, resulting in an overall loss, not a profit. By combining two small companies together, the company takes in profit from more policies, making the larger company more resistant to the possibility of a large payout. This is because the total profit is increasing by the expected profit from each additional policy, but the standard deviation is increasing by the square root of the sum of the variances. The larger a company gets, the more the expected profit outpaces the variability associated with that profit. 25. Youth survey. a) Many boys play computer games and use email, so the probabilities can total more than 100%. There is no evidence that there is a mistake in the report. b) Playing computer games and using email are not disjoint. If they were, the probabilities would total 100% or less. 9) Emailing friends and gender are not independent. 76% of girls emailed friends in the last week, but only 65% of boys emailed. If emailing were independent of gender, the probabilities would be the same. d) Let X = the number of students chosen until the first student is found who does not use the Internet. Use Geom(0.07). P(X = 5) = (0.93)*(0.07) = 0.0524 . 26. Meals. Let X= the amount the student spends daily. a) m=E(X+X)= E(X)+ E(X)=13.50+ 13.50 = $27.00 o = SD(X + X)= Var(X) + Var(X) =V7? +7? =$9.90 b) Inorder to calculate the standard deviation, we must assume that spending on different days is independent. This is probably not valid, since the student might tend to spend less ona day after he has spent a lot. He might not even have money left to spend! ) MH E(X4+X4+X+X4X4X4X)=7E(X) =7G13.50) = $94.50 SD(X+X+X+X+X+X+X) Var(X) + Var(X) + Var(X) + Var(X) + Var(X)+ Var(X)+ Var(X) = 77) = $18.52 4) Assuming once again that spending on different days is independent, it is unlikely that the student will spend less than $50. This level of spending is about 2.4 standard deviations below the weekly mean. Don’t try to approximate the probability! We don’t know the shape of this distribution.296 Part IV Randomness and Probability 27. Travel to Kyrgyzstan. a) Ifyou spend an average of 4237 soms per day, you can stay about ae = 21 days. b) Assuming that your daily spending is independent, the standard deviation is the square root of the sum of the variances for 21 days. o = {21(360) = 1649.73 soms ¢) The standard deviation in your total expenditures is about 1650 soms, so if you don’t think you will exceed your expectation by more than 2 standard deviations, bring an extra 3300 soms. This gives you a cushion of about 157 soms for each of the 21 days. 28, Picking melons. a) = E(First—Second) = E(First)— E(Second) = 2-18 = 4 Ibs. b) o=SD(First - Second) = /Var( jrst) + Var(Gecond) = V2.5? + 2? ~ 3.20 Ibs, According to the Normal model, the probability that a melon from the first store weighs more than a melon from the second store is approximately 0.894. 29, Home sweet home. Since the homes are randomly selected, these can be considered Bernoulli trials. There are only two possible outcomes, owning the home or not owning the home. The probability of any randomly selected resident home being owned by the current resident is 0.66. The trials are not independent, since the population is finite, but as long as the city has more than 8200 homes, we are not sampling more than 10% of the population. The Binomial model, Binom(820, 0.66), can be used to model the number of homeowners among the 820 homes surveyed. Let H = the number of homeowners found in 1 = 820 homes. E(H) = np = 820(0.66) = 541.2 homes. SDH) = Jnpq = {820(0.66)(0.34) = 13.56 homes. ‘The 523 homeowners found in the candidate's survey represent a number of homeowners that is only about 1.34 standard deviations below the expected number of homeowners. It is not particularly unusual to be 1.34 standard deviations below the mean. There is little support for the candidate’ claim of a low level of home ownership.Review of Part IV 297 30. Buying melons. ‘The mean price of a watermelon at the first store is 22(0.32) = $7.04. At the second store the mean price is 18(0.25) = $4.50. The difference in the price of the watermelons is expected to be $7.04 ~ $4.50 2.54, The standard deviation in price at the first store is 2.5(0.32) = $0.80, At the second store, the standard deviation in price is 2(0.25) = $0.50. ‘The standard deviation of the difference is 0.80" +0.50° ~ $0.94. 31. Who's the boss? a) P(first three owned by women) = (0.26)' ~ 0.018 b) (none of the first four are owned by women) ¢) (sixth firm called is owned by women | none of the first five were) = 0.26 Since the firms are chosen randomly, the fact that the first five firms were owned by men has no bearing on the ownership of the sixth firm. 32. Jerseys. a) P(all four kids get the same color) = (3) = 0.0156 (There are four different ways for this to happen, one for each color.) * b) Paall four kids get white) = (3) = 0.0089 ©) P(all four kids get white) = (\3) =0,0026 33. When to stop? a) Since there are only two outcomes, 6 or not 6, the probability of getting a6 is 1/6, and the trials are independent, these are Bernoulli trials. Use Geom(1/6). Eveseay ”% b) If 6's are not allowed, the mean of each die roll is 1#2*3+4+5 : 3. You would expect to get 15 if you rolled 5 times. 9) PGrolls without a6) ~(2) 0.402298 Part IV Randomness and Probability 34, Plan B. 142434445 a) If6's are not allowed, the mean of each die roll is b) Let X= your current score. You expect to lose it all z of the time, so your expected loss per 1 roll is =X. roll is ¢ c) Expected gain equals expected loss when x =3, So, X=18. 4) Roll until you get 18 points, then stop. 35. Technology on campus. Construct a Venn diagram of the disjoint outcomes. a) Compu P(neither tech.) = 1— P(either tech.) =1-[P(calculator) + P(computer) — P(both)| -[0.51+0.31-0.16] =0.24 Or, from the Venn: 0.34 (the region outside both circles) This is MUCH easier. 0.45 0.34 34% of students use neither type of technology. b) P(cale. and no comp.) = Pale.) ~ P(cale. and comp.) = 0.51-0.16 = 0.35 Or, from the Venn: 0.35 (region inside the Calculator circle, outside the Computer circle) 35% of students use calculators, but not computers. P(comp. 1) calc.) P(cale.) About 31.4% of calculator users have computer assignments. ) P(computer | calculator) = d) The percentage of computer users overall is 31%, while 31.4% of calculator users were computer users. These are very close. There is no indication of an association between computer use and calculator use. 36. Dogs. Since the outcomes are disjoint, probabilities may be added and subtracted as needed. a) P(no dogs) = (0.77)(0.77) = 0.5929 b) P(some dogs) = 1 P(no dogs) = 1 (0.77)(0.77) = ©) P(both dogs) = (0.23)(0.23) = 0.0529 4) P(more than one dog in each) = (0.05)(0.05) = 0.0025 A071Review of Part IV 299 37. Socks. Since we are sampling without replacement, use conditional probabilities throughout. a) rawu)-(4)(2)- 24 nin) 32° 7 6 42 57 » reoge-(S\ in) iz 9 8 60 5 Piat k black) = 1— P(no black) = 1-] — |] — |=—-=— 6) Motlestonebng)=1-mrovac=1-(2)(#)= «3 d) P(green)=0 (There aren’t any green socks in the drawer.) e) (match) = P(2 blue) + P(2 grey) + P(2 black) = (2)8) (34) (3)4) 38. Coins. Coin flips are Bernoulli trials. There are only two possible outcomes, the probability of each outcome is constant, and the trials are independent. a) Use Binom(36, 0.5). Let H = the number of heads in 1 = 36 flips. = E(H)= np =36(0.5)= 18 heads. = SD(H) = \npq = (36(0.5)(0.5) = 3 heads. b) Two standard deviations above the mean corresponds to 6 “extra” heads observed. ©) The standard deviation of the number of heads when 100 coins are flipped is = npg = {100(0.5\0.5) = 5 heads. Getting 6 “extra” heads is not unusual, d) Following the “two standard deviations” measurement, 10 or more “extra” heads would be unusual. e) What appears surprising in the short run becomes expected in a larger number of flips. ‘The “Law of Averages” is refuted, because the coin does not compensate in the long run. Accoin that is flipped many times is actually less likely to show exactly half heads than a coin flipped only a few times. The Law of Large Numbers is confirmed, because the percentage of heads observed gets closer to the percentage expected due to probability. 39. The Drake equation. a) N-f, represents the number of stars in the Milky Way Galaxy expected to have planets. b) N-f,-n,-f, represents the number of planets in the Milky Way Galaxy expected to have intelligent life. ©) _fi-f is the probability that a planet has a suitable environment and has intelligent life300 Part IV Randomness and Probability 4) f, = P(life| suitable environment). This is the probability that life develops, if a planet has a suitable environment. {f= P(intelligence | life). This is the probability that the life develops intelligence, ifa planet already has life. J, = P(communication | intelligence). This is the probability that radio communication develops, if a planet already has intelligent life. 40. Recalls. Organize the information in a tree diagram Recall (0740.02) =0.014 one a) American P(recall) = P(American recall) aN + P(Japanese recall) © \\ Recal (0710.98) =0.686 o7 + P(German recall) oor Recall (0.2}(0.01) = 0.002 = 0.014 + 0.002+0.001 ah =0.017 . iz 099 b) e Recall (0.24099) = 0.188 (American | recall) = American 1 recall) aot Pecal (0.14001) =0.001 P(recall) iE 0.014 German i 0.014 +0.002+0.001 0.99 Recall (0:1(0.98) = 0.089 = 0.824 41, Pregnant? Organize the information in a tree Poste Tet 07 a) one diagram. 088, P(pregnant | positive test) Pregnant P(pregnant () positive test) o7 ood P(positive test) Nogative Test (0.70.02) =0.014 _ 0.686 © 0.686-+0.006 Postive Test (0:30.02) =0.006 03 oo, =0.991 Not Pregnant od Nogatve Test (0:9)(0.98)=0.294Review of Part IV 301 42. Door prize. a) The probability that the first person in line wins is 1 out of 100, or 0.01, b) Ifyou are third in line, the two people ahead of you must not win in order for you to win. The probability is (0.99)(0.99)(0.01) = 0.009801, ©) There must be 100 losers in a row. The probability is (0.99)! = 0.366. 4) The first person in line has the greatest chance of winning at p = 0.01. The probability of winning decreases from there, since winning is dependent upon everyone else in front of you in line losing, ©) Position is irrelevant now. Everyone has the same chance of winning, p = 0.01. One way to visualize this is to imagine that one ball is handed out to each person. Only one person out of the 100 people has the red ball. It might be you! If you insist that the probabilities are still conditional, since you are sampling without replacement, look at it this way: Consider P(sixth person wins) = (2 i S\ ZZ (23) a we302 Part V From the Data at Hand to the World at Large Chapter 18 - Sampling Distribution Models 1. Coin tosses. a) The histogram of these proportions is expected to be symmetric, but not because of the Central Limit Theorem. The sample of 16 coin flips is not large. The distribution of these proportions is expected to be symmetric because the probability that the coin lands heads is the same as the probability that the coin lands tails. b) The histogram is expected to have its center at 0.5, the probability that the coin lands heads. ©) The standard deviation of data displayed in this histogram should be approximately equal to the standard deviation of the sampling distribution model, {2 7 oe =0:35. 4) The expected number of heads, np = 16(0.5) = 8, which is less than 10. The Success/Failure condition is not met. The Normal model is not appropriate in this case. 2. M&M's. a) The histogram of the proportions of green candies in the bags would probably be skewed slightly to the right, for the simple reason that the proportion of green M&M's could never fall below 0 on the left, but has the potential to be higher on the right. b) The Normal model cannot be used to approximate the histogram, since the expected number of green M&M's is np = 50(0.10) = 5, which is less than 10. The Success/ Failure condition is not met. ©) The histogram should be centered around the expected proportion of green M&M's, at about 0.10. 4) The proportion should have standard deviation f= {ene =0.042, A 5 3. More coins. a) H,= p= 05and o(p)=,"4 Pane Vn of Hs “About 68% of the sample proportions are expected 68% to be between 0.4 and 0.6, about 95% are expected to be between 0.3 and 0.7, and about 99.7% are Goreeparroacunsiuaeiaio7=i0a expected to be between 0.2 and 0.8. 95% SeChapter 18 Sampling Distribution Models 303 b) First of all, coin flips are independent of one another. There is no need to check the 10% Condition. Second, np = ng = 12.5, so both are greater than 10, The Success/Failure condition is met, so the sampling distribution model is N(0.5, 0.1). eee ne ©) Hy=p=O5and o(9)= f# (oer 0.0625 Proportion of of Sth fps About 68% of the sample proportions are expected to be between 0.4375 and 0.5625, about 95% are 7 expected to be between 0.375 and 0.625, and about, L 99.7% are expected to be between 0.3125 and 09125 0375 04313 O35 05625 ,0605 06875 0.6875. < 95% \ Coin flips are independent of one another, and n= nq =32, 0 both are grater than 10 (~ om) The Success /Failure condition is met, so the sampling distribution model is N(0.5, 0.0625). d) As the number of tosses increases, the sampling distribution model will still be Normal and centered at 0.5, but the standard deviation will decrease. The sampling distribution model will be less spread out. 4, Bigger bag. a) Randomization condition: The 200 M&M's in the bag can be considered representative of all M&M's, and they are thoroughly mixed. 10% condition: 200 is certainly less than 10% of all M&M's, Proportion of green M&M's in 200 0057 0058,0079 01 0421 014? D1I6s Success/Failure condition: np = 20 and ng = 180 are both greater than 10. b) The sampling distribution model is Normal, with: [pq n About 68% of the sample proportions are expected to be between 0.079 and 0.121, about 95% are expected to be between 0.058 and 0.142, and about 99.7% are expected to be between 0.037 and 0.163. ©) Ifthe bags contained more candies, the sampling distribution model would still be Normal and centered at 0.1, but the standard deviation would decrease. The sampling distribution model would be less spread out.