Chapter Eight 8.83 n 100 , x $273 , , and s $60; s x s / n 60 / 100 6 therefore a. Point estimate of x $273 Margin of error 1.96 s x 1.96 (6) $11.76 b. The 95 % confidence interval for μ is: 8.84 n 100 , x $2640 , and s $578; x zs x 273 1.96(6) 273 11.76 $261 .24 to $284.76 s x s / n 578 / 100 $57 .80 then a. Point estimate of x $2640 Margin of error 1.96 s x 1.96(57.80) $113 .29 b. The 97% confidence interval for μ is: x zs x 2640 2.17 (57.80) 2640 125 .43 $2514 .57 to $2765.43 8.85 n 36, x 24.015 inches, and .06 inches; so x / n .06 / 36 .01 The 99% confidence interval for μ is: x z x 24.015 2.58(.01) 24.015 .026 23.989 to 24.041 inches. Since the upper limit of the confidence interval is 24.041, which is greater than 24.025, the machine needs an adjustment. 8.86 n 50, x 3.99 inches, and .04 inches; so x / n .04 / 50 .00565685 The 98% confidence interval for μ is: x z x 3.99 2.33(.00565685 ) 3.99 .013 3.977 to 4.003 inches. Since the lower limit of the confidence interval is less than 3.98 inches, the machine needs an adjustment. 203 204 8.87 Chapter Eight n 32, x 1779, x 2 142,545 x x / n 1779 / 32 55.59 minutes s 2 x 2 x / n 142 ,545 (1779 ) 2 / 32 37 .52148578 minutes n 1 32 1 s x s / n 37 .52148578 / 32 6.63292426 minutes a. Point estimate of x 55.59 minutes Margin of error 1.96 s x 1.96(6.63292426 ) 13.00 minutes b. The 98% confidence interval for is: x zs x 55.59 2.33(6.632924 ) 55.59 15.45 40.14 to 71.04 minutes 8.88 n 36, x 547, x 2 9265 x x / n 547 / 36 15.19 hours s 2 x 2 x / n 9265 (547 ) 2 / 36 5.21984917 n 1 36 1 s x s / n 5.21984917 / 36 .86997486 hour a. Point estimate of x 15.19 hours Margin of error 1.96 s x 1.96(.86997486 ) 1.71 hours b. The 99% confidence interval for μ is: x zs x 15.19 2.58(.86997486 ) 15.19 2.24 12.95 to 17.43 hours 8.89 n 25, x $685 , and s $74; therefore s x s / n 74 / 25 $14 .80 df n 1 25 1 24, / 2 .5 (.99 / 2) .005 and t = 2.797 The 99% confidence interval for μ is: x ts x 685 2.797 (14.80) 685 41.40 $643 .60 to $726.40. 8.90 n 18, x 24 minutes, and s 4.5 minutes; so s x s / n 4.5 / 18 1.06066017 minutes df n 1 18 1 17, / 2 .5 (.95 / 2) .025 and t = 2.110 The 95% confidence interval for is: x ts x 24 2.110 (1.06066017 ) 24 2.24 21.76 to 26.24 minutes Mann – Introductory Statistics, Fifth Edition, Solutions Manual 8.91 n 20, x 9.75 hours, and s 2.2 hours; so 205 s x s / n 2.2 / 20 .49193496 hour df n 1 20 1 19, / 2 .5 (.90 / 2) .05 and t = 1.729 The 90% confidence interval for is: x ts x 9.75 1.729 (.49193496 ) 9.75 .85 8.90 to 10.60 hours 8.92 n 20, x 4.5 hours, and s .75 hours; so s x s / n .75 / 20 .16770510 hour df n 1 20 1 19, / 2 .5 (.98 / 2) .01 and t = 2.539 The 98% confidence interval for is: x ts x 4.5 2.539 (.16770510 ) 4.5 .43 4.07 to 4.93 hours. 8.93 n 12, x 26 .80, and s x 2 62.395 . This means x x / n 26.80 / 12 2.23 hours, 2 x 2 x / n 62 .395 (26 .80 ) 2 / 12 .48068764 hours, and n 1 12 1 s x s / n .48068764 / 12 .13876257 hours df n 1 12 1 11, / 2 .5 (.95 / 2) .025 and t = 2.201 The 95% confidence interval for is: x ts x 2.23 2.201(.13876257 ) 2.23 .31 1.92 to 2.54 hours. 8.94 n 10, x 1514 , and s x 2 229,646 ; Then x x / n 1514 / 10 151 .40 calories 2 x 2 x / n 229 ,646 (1514 ) 2 / 10 6.88315173 calories, and n 1 10 1 s x s / n 6.88315173 / 10 2.17664369 calories df n 1 10 1 9, / 2 .5 (.99 / 2) .005 and t = 3.250 The 99% confidence interval for is: x ts x 151 .40 3.250 (2.17664369 ) 151 .40 7.07 144 .33 to 158.47 calories 8.95 n 50, pˆ .12 , and qˆ 1 .12 .88; therefore s pˆ pˆ qˆ / n (.12)(.88) / 50 .04595650 a. Point estimate for p pˆ .12 or 12% Margin of error 1.96s pˆ 1.96(.04595650) .090 or 9.0% 206 Chapter Eight b. The 99% confidence interval for p is: pˆ zs pˆ .12 2.58(.04595650) .12 .119 .001 to .239 or .1% to 23.9% 8.96 n 500 , pˆ .44 , and qˆ 1 .44 .56 ; therefore s pˆ pˆ qˆ / n (.44)(.56) / 500 .02219910 a. Point estimate of p pˆ .44 or 44% Margin of error 1.96s pˆ 1.96(.02219910) .044 or 4.4% b. The 90% confidence interval for p is: pˆ zs pˆ .44 1.65(.02219910) .44 .037 .403 to .477 or 40.3% to 47.7% 8.97 n 20, pˆ 8 / 20 .40, and qˆ 1 .40 .60 ; hence, s pˆ pˆ qˆ / n (.40)(.60) / 20 .10954451 The 99% confidence interval for p is: pˆ zs pˆ .40 2.58(.10954451) .40 .283 .117 to .683 or 11.7% to 68.3% 8.98 n 16, pˆ 5 / 16 .313 , , and qˆ 1 .313 .687 ; hence s pˆ pˆ qˆ / n (.313)(.687) / 16 .11592859 The 97% confidence interval for p is: pˆ zs pˆ .313 2.17(.11592859) .313 .252 .061 to .565 or 6.1% to 56.5% 8.99 n 30; 95% confidence interval: $8.46 to $9.86 95% confidence interval for a large sample: x 1.96 s x a. x $9.86 $8.46 $9.16 2 b. x 1.96 , s x 9.86 , hence s x 9.86 x 9.86 9.16 .35714286 1.96 1.96 Confidence level = 99%, i.e. z = 2.58 so the 99% confidence interval is x 2.58 s x $8.24 to $10.08 8.100 a. Let x = time devoted to commercials per hour on Channel 66. A sample of 30 twenty – minute time intervals yielded a mean of 4.68 minutes of commercials with a standard deviation of 1.30 minutes. This is equivalent to: x 3(4.68) 14 .04 minutes and s 3(1.30) 3.90 minutes for a 60–minute time interval. s x s / n 3.90 / 30 .71203932 minute. Mann – Introductory Statistics, Fifth Edition, Solutions Manual 207 The 90% confidence interval for is: x zs x 14.04 1.65(.71203932 ) 14.04 1.17 12.87 to 15.21 minutes per hour b. n 30, ˆ 7 / 30 p The point estimate of is (7/30) 60 minutes per hour = 14 minutes per hour a. Both estimates yield approximately 14 minutes per hour. 7 23 pˆ qˆ / n / 30 .07722022 and pˆ 7 / 30 .233 30 30 d. s pˆ So the 90% confidence interval for p̂ is pˆ 1.65s pˆ .233 1.65(.07722022) .106 to .360. Hence the 90% confidence interval for p × 60minutes per hour = 6.36 to 21.60 minutes per hour. e. Maximum error of Waldo’s estimate: z s p 60 minutes per hour = 1.65 (.07722022 ) (60 ) minutes per hour = 7.64 minutes per hour Maximum error of class estimate: z s x 1.65 (.71203932 ) = 1.17 minutes per hour f. We need a new s pˆ such that s x s pˆ 60 minutes per hour; hence, .71203932 (7 / 30 ) (23 / 30 ) 60 , so n n (7 / 30 ) (23 / 30 ) (.71203932 / 60 ) 2 1270 .2 1271 . It does not seem to be feasible to achieve the same accuracy as that of the class using Waldo’s method. 8.101 Let: p1= proportion of 12–18 year old females who expect a female president within 10 years p2= proportion of 12–18 year old females who expect a female president within 15 years p3= proportion of 12–18 year old females who expect a female president within 20 years p4 = proportion of 12–18 year old females who do not expect a female president not within their lifetime n 1250 , pˆ 1 .40, pˆ 2 .25, pˆ 3 .21, pˆ 4 .14 ˆ and nqˆ exceed 5 for all these proportions, so the sample is considered large. np s pˆ1 pˆ 1qˆ1 / n .40(.60) / 1250 .01385641 The 95% confidence interval for p1 is: pˆ 1 zs pˆ1 .40 1.96(..01385641) .40 .027 .373 to .427 or 37.3% to 42.7% s pˆ 2 pˆ 2 qˆ 2 / n .25(.75) / 1250 .01224745 208 Chapter Eight The 95% confidence interval for p2 is: pˆ 2 zs pˆ 2 .25 1.96(.01224745) .25 .024 .226 to .274 or 22.6% to 27.4% s pˆ 3 pˆ 3 qˆ 3 / n .21(.79) / 1250 .01152042 The 95% confidence interval for p3 is: pˆ 3 zs pˆ 3 .21 1.96(.01152042) .21 .023 .187 to .233 or 18.7% to 23.3% s pˆ 4 pˆ 4 qˆ 4 / n .14(.86) / 1250 .00981428 The 95% confidence interval for p4 is: pˆ 4 zs pˆ 4 .14 1.96(.00981428) .14 .019 .121 to .159 or 12.1% to 15.9% A confidence interval is a range of numbers (in this particular case proportions or percentages) which give an estimate for the true value (i.e. proportion of 12–18 year old females who feel this way). The 95% means that we are 95% confident that this interval actually contains the true value. A single percentage that we assign as an estimate would almost always differ from the true value, hence a range with the associated confidence level is more informative. We assume that the 1200 people are a random sample of 12–18 year old females. 8.102 n 100 , and pˆ 10 / 100 .10 a. s pˆ .10 (.90 ) .030 100 The 95% confidence interval for p is: pˆ 1.96s pˆ .10 1.96(.030) .10 .059 .041 to .159 or 4.1% to 15.9% b. No, 18% does not lie within the confidence interval in part a. This suggests that the vaccine is effective to some degree. c. There are several things that may have distorted the outcome of this experiment: we only point out two. It is not mentioned, although it is reasonable to assume that all the dogs that were vaccinated did not have Lyme disease to begin with. Assuming this, the vaccinated dogs were exposed only 1 year to possible tick bites, whereas the other dogs in the area probably were exposed much longer. Also, the owners who agreed to participate in this experiment may be more concerned about their dogs’ health and restrict the area that the dogs can access, thereby decreasing the exposure to the ticks as well. 8.103 s 4.1 miles, E 1.0 and z 1.96 . may be estimated by s. Hence, n z 2 s 2 / E 2 (1.96) 2 (4.1) 2 /(1.0) 2 64.58 65 Thus, an additional 65–20=45 observations must be taken. Mann – Introductory Statistics, Fifth Edition, Solutions Manual 8.104 209 The major problem with this procedure is that the sample is not drawn from the whole target population. This introduces nonsampling error. i. Households with no cars would be excluded from this sample. Furthermore, households with more than 2 cars could be counted more than once in the sample. Both of these problems would result in an upwardly biased estimate of p. ii. Because of the characteristics of this particular gas station, the clientele may not be typical of all gas station clientele (with respect to multiple car ownership). Other problems may be present, such as the next 200 gasoline customers may not be typical of all customers, or there may be dependency between gasoline customers from the same household. If the attendant took a random sample of 200 households, and determined how many of these households owned more than 2 cars, the sampling error would still be present, but could be reduced by taking a larger sample. 8.105 A. Here, 170 , E 100 , and 1 .99 , and z 2.58 Thus, the required sample size is n z 2 2 E2 (2.58) 2 (170 ) 2 (100 ) 2 19 .24 or 20 days Note that since n < 30, we must assume that the number of cars passing each day is approximately normally distributed. Or we may take a large (n 30 ) sample. b. Since n z 2 2 / E 2 , then, z E n / 100 20 / 272 1.64 which corresponds to a confidence level of approximately 90%. c. Since, n z 2 2 E 2 , then, E z n 2.58(130 ) 75 .00 20 Thus, they can be 99% confident that their point estimate is within 75 cars of the true average. 8.106 No, the student’s analysis does not make sense. The relevant parameter, p, is the proportion of all U.S. senators in favor of the bill. The value, .55, is not a sample proportion: instead it is the population proportion, p. Since, p = .55 is known there is no need to estimate it. Self – Review Test for Chapter Eight 1. a. Estimation means assigning values to a population parameter based on the value of a sample statistic. b. An estimator is the sample statistic used to estimate a population parameter. 210 Chapter Eight c. The value of a sample statistic is called the point estimate of the corresponding population parameter. 2. 7. b 3. a 4. a 5. d 6. b n 36 , x $159,000, and s $27,000; hence s x s / n 27 ,000 / 36 $4500 a. Point estimate of x $159 ,000 Margin of error 1.96 s x 1.96(4500 ) $8820 b. The 99% confidence interval for is: x zs x 159 ,000 2.58(4500 ) 159 ,000 11,610 $147 ,390 to $170,610 8. n 25 , x $410,425, and s $74,820 df n 1 25 1 24, / 2 .5 (.95 / 2) .025 and t = 2.064 s x s / n 74,820 / 25 $14,964 The 95% confidence interval for is: x ts x 410 ,425 2.064 (14,964 ) 410 ,425 30,885 .70 $379 ,539 .30 to $441,310.70 9. n 612 , pˆ .41, and qˆ 1 .41 .59, so s pˆ pˆ qˆ / n (.41)(.59) / 612 .01988118 a. Point estimate of p pˆ .41 Margin of error 1.96s pˆ 1.96(.01988118) .039 pˆ zs pˆ .41 1.96(.01988118) .41 .039 .371 to .449 b. The 95% confidence interval for p is: 10. E .65 houses, 2.2 houses, and z 1.96 , then 11. E .05, z 1.65, 12. E .05, p q .50 , then n z 2 pq E2 n z 2 2 E2 (1.65) 2 (.50 )(. 50 ) (.05) 2 pˆ .70, qˆ 1 pˆ 1 .70 .30, and z 1.65; so n z 2 pˆ qˆ E2 (1.96 ) 2 (2.2) 2 (.65) 2 44 .01 ≈ 45. 272 .25 273 . (1.65) 2 (.70 )(. 30 ) (.05) 2 228 .69 229 . Mann – Introductory Statistics, Fifth Edition, Solutions Manual 211 13. The width of the confidence interval can be reduced by: 1. Lowering the confidence level 2. Increasing the sample size The second alternative is better because by lowering the confidence level we will simply lower the probability that our confidence interval includes . 14. To estimate the mean number of hours that all students at your college work per week: 1. Take a random sample of 12 students from your college who hold jobs 2. Record the number of hours each of these students worked last week 3. Calculate x , s, and s x from these data 4. After choosing the confidence level, find the value for the t distribution with 11df and for an area of / 2 in the right tail. 5. Obtain the confidence interval for by using the formula x ts x You are assuming that the hours worked by all students at your college have a normal distribution. 15. To estimate the proportion of people who are happy with their current jobs: 1. Take a random sample of 35 workers 2. Determine whether or not each worker is happy with his or her job 3. Calculate pˆ , qˆ, and s p 4. Choose the confidence level and find the required value of z from the normal distribution table 5. Obtain the confidence interval for p by using the formula pˆ zs p