This is the second exam from Spring 1988. Unfortunately the take-home exam that accompanied it is not available since the disk that contains it is unreadable by this version of Word. 3/23/98 252y9822 ECO252 QBA2 Name SECOND HOUR EXAM Hour of Class Registered (Circle) February 18, 1998 MWF 10 11 TR 12:30 2:00 Hour of Class Attended (If Different) ______________ I. (14 points) Do all the following. x ~ N 5,3 1. 16 5 3 5 z P 3 x 16 P P 2.67 z 3.67 3 3 P 2.67 z 0 P 0 z 367 . .4962.4999 .9961 2. 3. 4. 5. 3 5 05 z . z 0.67 P 0 x 3 P P 167 3 3 P 167 . z 0 P 0.67 z 0 .4525.2486 .2039 0 5 2 5 z . P 2 x 0 P P 2.33 z 167 3 3 P 2.33 z 0 P 167 . z 0 .4901.4525 .0376 0 5 . P x 0 P z P z 167 3 P z 0 P 167 . z 0 .5000.4525 .0475 2 5 . F 2 (The Cumulative probability) P x 2 P z P z 100 3 P z 0 P 100 . z 0 .5000.3413 .1587 6. A symmetrical interval about the mean with 82% probability. We want two points x .09 and x.91 , so that P x.91 x x.09 .8200 . From the diagram, if we replace x by z, P 0 z z.09 .4100 . The closest we can come is . , and x z.09 5 134 P 0 z 134 . .4099 . So z.09 134 . 3 5 4.02 , or 0.98 to 9.02. check: 7. 9.02 5 0.98 5 z . z 134 . 2.4099 .8198 P 0.98 x 9.02 P P 134 3 3 x .18 We want a point x .18 , so that P x x .18 .18 . From the diagram, if we replace x by z, P 0 z z.18 .32 . The closest we can come is P0 z 0.91 .3186 or P0 z 0.92 .3212 . Use something between the two. So z.18 0.915 , and x z.18 5 0.915 3 5 2.745 , or 7.745. 7.745 5 check: P x 7.745 P z P z 0.92 P z 0 P 0 z 0.92 .5000.3212 .1788 3 2 3/20/98 252y9822 II. (6 points-2 point penalty for not trying part a.) Show your work! I wish to decide whether to become a high school teacher or a university professor. I will do so on the basis of earnings alone. I am interested in both a high salary and the certainty of getting it. The data that I get in my sample is shown below. Assume that the data represent independent samples taken from populations with the normal distribution. x1 x2 d High School Teachers University Professors Difference 38.7 39.2 0.50 42.5 68.5 -25.00 62.6 43.7 18.90 38.3 47.2 -8.90 49.2 51.3 -2.10 73.5 29.9 43.60 a. Compute s1 , the standard deviation for high school teachers. (3) Note that x 2 46.63, s2 12.99 , d 4.50, sd 2386 . . You will not necessarily need all of these in your computations. b. Test to see if means for salaries of both types of teacher are equal. You may assume that they come from populations with identical variances. Does this indicate that one of the occupations is more lucrative?(3) c. (Extra Credit) Test to see if variances are equal. What can you say about the relative safety of the two occupations?(2) a. Solution: x1 s12 x 1 n x 2 1 Item 304.8 50.80 6 nx12 n 1 16512.48 6 50.80 . s1 14.343 . 205728 . 5 2 1 2 3 4 5 6 Total b. Solution: From page 10 of the Syllabus Supplement: Interval for Confidence Hypotheses Interval Difference H 0: 0 d t 2 sd between Two H 1: 0 1 1 Means ( sd s p 1 2 n1 n2 unknown, variances DF n1 n2 2 assumed equal) x12 1497.69 1806.25 3918.76 1466.89 2420.64 5402.25 16512.48 x1 38.7 42.5 62.6 38.3 49.2 73.5 304.8 Test Ratio t s p2 d 0 sd Critical Value d cv 0 t 2 sd n1 1 s12 n2 1 s22 n1 n2 1 3 3/20/98 252y9822 H 0 : 0 H 0: 1 2 H 1 : 0 Same as H 1 : 1 2 1 2 if 0 0 s p2 n1 1 s12 n2 1 s22 sd s p n1 n2 1 = x1 511333 . , s12 197.835, d x1 x 2 4.50 5 205.728 512.78 10 2 x 2 46.63, s2 12.99 .05 or as you specified. DF n1 n2 2 6 6 2 10 205.728 163.3284 10 184.5542 t .025 2.228 2 1 1 1 1 . 7.8423 s 184.5542 615181 6 6 n1 n2 Test Ratio: t d 0 4.50 0 0.5737 This is between 2.228 . sd 7.8433 or Critical Value: d cv 0 t 2 sd 0 2.228 7.8433 17.75 d 4.50 is between these values. or Confidence Interval: d t 2 sd 4.50 2.228 7.8433 4.50 17.75 or -13.25 to 22.25. The interval includes 0. In all cases accept H 0 . There is no difference in salaries. c. Solution: From page 11 of the Syllabus Supplement: Interval for Confidence Hypotheses Interval Ratio of 22 s22 DF1 , DF2 H 0 : 12 22 2 F.5 .5 Variances 2 2 1 s1 H 1 : 12 22 1 DF1 , DF2 F1 2 DF1 , DF2 DF1 n1 1 F 2 DF2 n2 1 2 .5 .5 2 or 1 2 s2 205.728 5,5 F 12 1260 . s2 12.78 2 There is no difference in risk. Test Ratio F DF1 , DF2 Critical Value s12 s22 and F DF2 , DF1 s22 s12 s2 1 5,5 5,5 . Since both are below F.025 7.15 , accept H 0 . F 22 . s1 1260 4 3/23/98 252y9822 III. Do at least 2 of the following 4 Problems (at least 10 each) (or do sections adding to at least 20 points Anything extra you do helps, and grades wrap around) . Show your work! State H 0 and H1 where applicable. 1. In a random sample of 600 television sets made between Tuesday and Thursday from a production line 80 were defective. In a random sample of 200 sets made on Monday 40 were defective. a. Using a 5% test, test the hypothesis that the defect rate is larger on Monday than during the middle of the week. (5) b. If you conclude that the defect rate is higher on Monday, test the Hypothesis that it is more than 3% higher (3). c. If you found out that of 200 sets made on Friday 34 were defective, test for a difference between Monday, Tuesday-Thursday, and Friday. (6) On Monday On Tuesday-Thursday On Friday x1 40, n1 200, p1 40 .2000 200 x 2 80, n2 600, p2 80 .1333 600 x 3 34, n2 200, p1 34 .1700 200 a. Solution: From page 11 of the Syllabus Supplement: Interval for Confidence Interval p p z 2 sp Difference between proportions q 1 p p p1 p2 p1q1 p2 q 2 n1 n2 s p Hypotheses Test Ratio H 0 : p p0 H 1 : p p0 p 0 p 01 p 02 or p 0 0 z p p 0 p If p 0 p p01q 01 p02 q 02 n1 n2 Or use s p Critical Value pcv p0 z 2 p If p0 0 p p0 q 0 1 n1 1 n2 n p n2 p2 p0 1 1 n1 n 2 H 0 : p 0 H 0 : p1 p2 p p1 p2 .0667. Same as . . H 1 : p 0 H 1 : p1 p2 .05 or as you specified. z 1645 p0 .20.80 .1333.8677 p1q1 p2 q 2 n1 n2 sp 200 600 .00099 .03151 n1 p1 n2 p2 200.20 600.1333 40 80 .15 n1 n2 200 600 200 600 p p0 q 0 Test Ratio: z 1 n1 1 n2 p p 0 p .15.85 1 200 1 600 .000850 .02915 .0667 0 2.287 This is above 1.645. .02915 or Critical Value: pcv p0 z p 0 1645 . .02915 .0480 p .0667. is above this value. or Confidence Interval: p p z sp or p .0733 1645 . .02915 .0187 . The interval does not include 0. In all cases reject H 0 . 5 3/23/98 252y9822 b. Solution: H 0 : p .03 H 1 : p .03 Test Ratio: z p p 0 .0667 .03 1165 . This is below 1.645. s p .03151 or Critical Value: pcv p0 z p .03 1645 . .03151 .0818 p .0733. is below this value. or Confidence Interval: p p z sp or p .0733 1645 . .02915 .0187 . The interval does not contradict H 0 : p .03 In all cases accept H 0 . c. Solution: H 0 : Homogeneous H 1 : Not homogeneous DF r 1 c 1 1 2 2 2 2 .05 5.9915 O pr 40 80 34 154 E .154 160 520 166 846 .846 200 600 200 1000 1000 . O E 40 80 34 160 520 166 1000 30.8 92.4 30.8 169.2 507.6 169.2 1000.0 O2 E 51.9481 69.2641 37.5325 151.3002 532.7029 162.8605 1005.6083 pr 30 .8 92 .4 30 .8 .154 169 .2 507 .6 169 .2 .846 200 .0 600 .0 200 .0 1.000 O E O2 n 1005.6083 1000 5.6083 E E Since this is less than 5.9915. accept H 0 . 2 6 3/23/98 252y9822 2. A group of students with insomnia are asked to test a new sleeping pill. The number of hours of sleep each gets is summarized by the following data: x1 refers to the sleeping pill, x 2 refers to the placebo (sugar pill) and d x1 x2 x1 6.88, s1 0.52 , n1 11 , x 2 6.40, s2 0.75 , n2 11 , d 0.48, sd 0.33 a. Assume that the variances of the parent populations are not equal and that the samples are independent, but that the normal distribution applies. Test the hypothesis that the mean hours for the sleeping pill is greater than the placebo. (i) Compute the number of degrees of freedom for the test (5). (ii) Do the test at the 95% level using either a critical value or a test ratio. (5) (iii) Do a two -sided confidence interval for the difference between the means. (2) b. Assume that this represents paired data and redo the test (4) a. Solution: From page 10 of the Syllabus Supplement: Interval for Confidence Hypotheses Test Ratio Critical Value Interval 2 s12 s22 n1 n2 sd DF s12 s22 n 1 n2 s12 2 n1 1 2 s 22 t 1 2 d cv 0 t 2 sd d 0 sd Same as H 0: 1 2 2 H 1: 1 2 n1 1 H 0 : 0 H 1: 0 H 0 : 0 H 1: 0 d t sd Difference between Two Means( unknown, variances assumed unequal) 2 if 0 0 n2 n2 1 H 0: 1 2 Same as H 1: 1 2 s12 0.52 n1 11 2 s22 0.75 n2 11 2 s12 n1 0.02458 s22 n2 0.05114 0.07572 sd s12 s22 n1 n 2 0.07572 0.2752 (i) DF s12 s22 n1 n 2 2 s12 2 n1 n1 1 s22 2 n2 n2 1 (ii)Test Ratio: t 0.07572 2 .02458 2 .05114 2 11 1 17.81 17 . . So use 17 degrees of freedom. t .05 1740 11 1 d 0 0.48 0 1744 . This is above 1740 . . sd 0.2752 or Critical Value: d cv 0 t sd 0 1740 . 0.2752 0.479 d 0.48 is above this value. With both methods reject H 0 . 17 (iii) Confidence Interval: t .025 2.110. d t 2 sd 0.48 2.110 0.2752 0.48 0.58 or -0.10 to 1.06. 7 3/23/98 252y9822 b. Solution: H 0 : 0 H 0: 1 2 Same as H 1: 0 H 1: 1 2 1 2 Test Ratio: t d 0 0.48 0 4.824 sd 0.0995 sd ss n 0.33 11 DF n 1 10 01089 . 0.0995 11 t .10 . 05 1812 This is above 1812 . . or Critical Value: d cv 0 t sd 0 1812 . 0.0995 01803 . d 0.48 is above this value. or Confidence Interval: d t 2 sd 0.48 1812 . 0.0995 0.2997 . With all methods reject H 0 . 8 3/23/98 252y9822 3. The failure times in thousands of hours are given for a random sample of 8 components. 1.700 1.000 1.450 1.500 1.560 1.600 1.650 1.725 a. Test the hypothesis that these numbers came from a normal distribution. Use a 5% significance level. (5) b. Test the hypothesis that the above data came from a normal distribution with a mean of 1.700 and a standard deviation of 2.000 (5) c. A television set distributorship believes that television ownership in the local area is distributed according to a Poisson distribution with a mean of 4. A sample of 100 is taken. Is this true? Use a 5% significance level. (5) Number of TV sets: 0 1 2 3 4 5 or more. Number of Households: 7 27 28 18 10 10 a. Solution: H 0 : N ?, ? H 1 : Not Normal Because the mean and standard deviation are unknown, this is a Lilliefors problem. xx From the data we find that x 1523 and s 0.2314 . t . . s x 1000 . 1450 . 1500 . 1560 . 1600 . 1650 . 1700 . 1725 . t Ft O O n Fo D 2.26 0.32 010 . 016 . 0.33 0.54 0.76 0.87 .0119 .3745 .4602 .5636 .6293 .7054 .7764 .8078 1 1 1 1 1 1 1 1 Max D .1924 O n 8 0125 . 0125 . 0125 . 0125 . 0125 . 0125 . 0125 . 0125 . 0125 . 0.250 0.375 0.500 0.625 0.750 0.825 1000 . .1131 .1255 .0852 .0636 .0043 .0446 .0986 .1924 Since the Critical Value for .05 is .288, accept H 0 . b. Solution: H 0 : N 17 . , 2 H 1 : Not N 17 . , 2 Because the mean and standard deviation are known, this is a Kolmogorov-Smirnov problem. x z . x 1000 . 1450 . 1500 . 1560 . 1600 . 1650 . 1700 . 1725 . z 0.35 0125 . 010 . 0.07 0.05 0.025 0 0.013 F z .3632 .4503 .4602 .4721 .4801 .4900 .5000 .5497 O O n Fo D 1 1 0125 . 0125 . .2382 0125 . 0.250 .2003 1 1 1 0125 . 0125 . 0125 . 0.375 0.500 0.625 .0852 .0279 .1449 1 0125 . 0.750 .2600 1 1 0125 . 0125 . 0.825 1000 . .3750 .4502 Mas D .4502 O n 8 Since the Critical Value for .05 is .457, accept H 0 . 9 3/23/98 252y9822 c. Solution: H 0 : Poisson 35 . H 1 : Not Poisson 35 . This can be done as a chi-square or KolmogorovSmirnov problem. f e and Fe come from the Poisson table. x 0 O 7 O n .07 Fo .07 fe .018316 E 1832 . Fe .0183 D .0517 1 27 .27 .34 .073263 7.326 .0916 .2484 2 3 28 18 .28 .18 .62 .80 .146525 14.626 .195367 19.537 .2381 .4335 .3919 .3665 4 10 .10 .90 .195367 19.537 .6288 .2712 5 10 .10 100 . 100 100 . .37116 37.116 100000 . 97.974 0 For the Kolmogorov-Smirnov Method the 5% critical value is O2 O E 34 9.158 E 126.228 28 14.626 18 19.537 53.603 16.596 10 19.537 5118 . 10 38116 . 100 97.974 2.624 204.169 100 104.169 136 . . This is less than the 0136 . 100 2 4 maximum value of D , which is .3919, so reject H 0 . For the Chi-Squared Method. .05 9.4877 , this is less than our computed 2 , which is 104.169, so reject H 0 . Note that, for the Chi-Squared method. the first two cells must be merged, since the first has an expected value well below 5. 10 3/23/98 252y9822 4. a. The lifetimes of independent samples of two brands of power equipment are given below. The distributions are not believed to be normal, so compare the medians at the 10% significance level.(5) Brand 1: 2.4 3.8 6.0 6.9 3.6 Brand 2: 2.0 3.9 5.0 5.7 6.5 b A sample of eight married couples are compared as to their spending on clothing. The distributions are not believed to be normal, so that medians are compared. Use the 10% significance level to answer whether husbands and wives spend equally in this category.(5) Couple Number: 1 2 3 4 5 6 7 8 Husband's Spending: 310 250 240 260 240 310 481 240 Wife's Spending: 390 385 190 400 410 490 482 330 a. Solution: Wilcoxon-Mann-Whitney Method H 0 : 1 2 x1 r1 x 2 r2 2.0 1 2.4 3.6 3.8 2 3 4 Check: 29 26 55 1011 H 1: 1 2 2 For a 10% two-tailed test, the critical values are 18 and 37. W 27 is between them, so accept H 0 . 3.9 5 5.0 6 5.7 7 6.0 8 6.5 9 6.9 10 __ 27 28 b. Solution: Wilcoxon Signed rank test for paired data. H 0 : 1 2 x1 x 2 difference rank 310 390 80 3 250 385 135 5 240 190 -50 -2 260 400 140 6 240 410 170 7 310 490 180 8 481 482 1 1 240 330 90 4 T 2, T 34 H 1: 1 2 . Check : T T 2 34 36 Since TL TL .05 6 , and the smaller T is below it, reject H 0 . 8 9 2 2 Note: I defined the difference as x 2 x1 . If I had defined it the other way around, it would not have affected the results. 11