Design and Analysis of Experiments, Eighth Edition The 100(1 ~ @) confidence interval for the ratio ofthe population variances io} Sip SE wanciat 5S GF 27 Problems 59 oF Si ot St ‘To illustrate the use of Equation 2.50, the 95 percent confidence interval for the ratio of vari- ances r/o} in Example 2.2 is, using Foosss1: = 3:59 and Fasysou1 = VFocassi9= VB.92 = 0255, 4s << 45 10gO™) = 3 = 108°? om <2 49 2 2.7 Problems 2, Computer output for a random sample of data is shown below. Some of the quantities are missing. Compute the values of the missing quantities. 22, Computer output for a random sample of data is ‘shown below. Some of the quantities are missing. Compute ‘the values of the missing quantities. 23. Suppose that we are testing Hy: = ji versus 1H, 4 # po. Calculate the P-value for the following observed ‘values ofthe test statistic: @ %=225 @)%=155 @Z=195 @%=-010 24, Suppose that we are testing Hy: f= jo versus 4H, w> po. Caleulate the P-value for the following observed ‘values ofthe test statistic: (@) = 245 (b)Z = -153 @Z=195 (@)%=-025 25. Consider the computer output shown below. ©%=210 (©Z~215 (@) Fil in the missing values in the output. What concln- sion would you draw? (©) Is this a one-sided or two-sided test? (© Use the output and the normal abe to find 299 percent Clon the mean. (@ What is the P-value if the altemative hypothesis is Hy: p> 307 26. Suppose that we are testing Hf, : yx; = pl, versus Hy 1 # py where the two sample sizes are n, = ny = 12, Both Sample variances are unknown but assimed equal. Find ‘bounds onthe P-value forthe following observed values of the test statisti, (2) = 230) =341 (©) 44= 1.95 (@) = 245 27. ‘Suppose that we ae testing Hy: = ps VETS Hy ‘my > pa where the two sample sizes ae nj = m; = 10. Both ‘ample variances are unkrown but assimed equal. Find ‘bounds onthe P-value forthe following observed valves of the test statisti. (@) = 231 (b) f= 3.60 (6) = 1.95 (A) = 2.19 28, Consider the following sample daa: 9.37, 1304, 1169,821, 1.18, 1041, 13.15, 1151, 1321, and 775. Isit reasonable to assume that this data i a sample from anormal distribution? Is there evidence to support claim that the mean of the population i 10? 29. A computer program has produced the following out- put for a hypothesis-testing problem: (2) Whats the missing value forthe standard error? (b) Is this a two-sided or a one-sided test? (© Ifa = 0.05, what sre your conclusions? (@) Find a 90% two-sided CI on the difference inDesign and Analysis of Experiments, Eighth Edition 60 chapter 2 m Simple Comparative Experiments 2.10. A computer program has produced the following out- put for a hypothesis-esting problem: (a) What is the missing value forthe standard enor? () Is this a two-sided or a one-sided test? (© Ifa = 0.05, what are your conclusions? (@) Find a 95% two-sided Cl onthe difference in means. 21. Suppose that we are testing Hy: 4 = Mo versus Hy: 4 > iy with a sample size of n = 15. Caleulate bounds (on the P-value for the following observed values of the test statistic: (@) = 235 (b) = 355 (©) 4 = 200 (@) = 155 242. Suppose that we are testing Hy: 4 = My versus Hy:4 wy with a sample size of n = 10. Calculate bounds on the P-value forthe following observed values ofthe test statistic (@) = 248 (0) @ = 188 (©) 395 (6) 4 = 2.69 125 213. Consider the computer output shown below. (a) Fill in the missing values in the output. Can the null hypothesis be rejected atthe 0.05 level? Why? (b) Is this a one-sided or a two-sided test? (6) If the hypotheses had been Hy : = 90 versus H,: 4 # 90 would you reject the null hypothesis atthe 00S level? (@) Use the output and the table to find a 99 percent two- sided CI on the mean. (©) What is the P-value if the altemative hypothesis is Hip> 9? 2.14. Consider the computer output shown below. (a) How many degrees of freedom are there on the test statistic? (b) Fill in the missing information, 2.15. Consider the computer output shown below. Two SampleF (a) Can the mull hypothesis be rejecied atthe 0.05 level? why? (@) Is this a one-sided ora two-sided test? (6) If the hypotheses had been Hj; ~ 2 = 2 versus HY: fy ~ 4% 2 would you reject the nll hypothesis at the 0.05 level? (@ If the hypotheses had been Hj; ~ up = 2 versus 14, py ~ uy < 2 would you reject the nll hypothesis at the 005 level? Can you answer this question with- ut doing any addtional calculations? Why? (@) Use the output and the ¢ table to find a 95 perceat upper confidence bound on the difference in means. (® What is the P-value if the hypotheses are Hy: a — y= 2 versus Hj ~ py #2? 2.16. ‘The breaking strength of a fiber is required to be at least 150 psi. Past experience has indicated thatthe standard deviation of breaking strength is = 3 psi A random sample Of four specimens is tested, and the results are y, = 145, 153,y5 = 150,and y, = 147. (a) State the hypotheses that you think shouldbe tested in this experiment. () Test these hypotheses using a = 0.05. What are your conclusions? (6) Find the P-value for the test in pat (0). (@) Construct a 95 percent confidence interval onthe mean breaking strength. 27. The viscosity of a liquid detergent is supposed to average 800 centistokes at 25°C. A random sample of 16 batches of detergent is collected, andthe average viscosity is 812, Suppose we know thatthe standard deviation of viscosity is o = 25 centistokes. (a) State the hypotheses that should be tested. (b) Test these hypotheses using a = 0.08. What are your conclusions? (©) Whats the P-value for the test? (@) Find « 95 perceat confidence interval on the mean. 248. The diameters of stel shafts produced by a certain ‘manufacturing process should have a mean diameter of 0.255 inches. The diameter is known to have a standard deviation of| ‘= 0.0001 inch. A random sample of 10 shafts has an aver- age diameter of 0.2545 inch.Design and Analysis of Experiments, Eighth Edition (@) Setup appropriate hypotheses on the mean (b) Test these hypotheses using a = 0.05. What are your conclusions? (© Find the P-value for this test. (@ Construct a 95 percent confidence interval onthe mean shaft diameter. 249. A normally distributed random variable has an ‘unknown mean and a known variance o° = 9. Find the sam- pie size required to consiuct a 95 percent confidence interval ‘onthe mean that has total length of 1.0. 220. The shelf life ofa carbonated beverage is of interest. ‘Ten bottles are randomly selected and tesed, andthe follow- ing results ae obtained: Days 108138 m4 168 m4 19 106134 us 139 (@) We would like to demonstrate that the mean shelf life exceeds 120 days, Set up appropriate hypotheses for investigating this claim. (b) Test these hypotheses using a ~ 0.01. What are your conclusions? (©) Find the P-value forthe test in part (). (@) Construct a 99 percent confidence interval on the mean. shelf life. 221. Consider the shelf life data in Problem [Link] shelf life be described or modeled adequately by a normal distibu- tion? What effect would the violation of this assumption have ‘n the test procedure you used in solving Problem 2.15? 222. The time to repair an electronic instrument isa normal- ly distibuted random variable measured in hours. The repair times for 16 such instruments chosen at random are as follows: 19280212 2s 319179 m2 362168250 149260485170 (a) You wish to know ifthe mean repair time exceeds 225, hours. Set up appropriate hypotheses for investigating this issue. (b) Test the hypotheses you formulated in part (a). What are your conclusions? Use a = 0.05. (©) Find the P-value forthe test (@) Construct a 95 percent confidence interval on mean ‘repair time. 27 Probtems 61 223. Reconsider the repair time data in Problem 2.22. Can ‘repair time, in your opinion, be adequately modeled by a nor- ‘mal distribution? 224. ‘Two machines are used for filing plastic bots with ‘2 net volume of 16.0 ounces. The filling processes can be ‘assumed to be normal, with standard deviations of 7, = 0.015 ‘and a = 0.018. The quality engineering department suspects that both machines fill tothe same net volume, whether or not this volume is 16.0 ounces. An experiment is performed by taking a random sample from the output of each machine. “Machine 1 Machine 2 1603-1601 -—=«1602—=—«16.03 1604 15961597 16.04 1605 1598 1596 16.02 1605 160216011601 160215991599 16.00 (@) State the hypotheses that should be tested in this experiment (b) Test these hypotheses using a = 0.05. What are your conclusions? (© Find the P-value for this test. (@) Find 8 95 percent confidence interval on the difference {in mean ill volume for the two machines. 225. Two types of plastic are suitable for use by an elec twonie calculator manufacturer. The breaking strength of this plastic is important, It is known that 7, = a = 1.0 psi. From random samples of n, = 10 and n; ~ 12 we obtain, ~ 1625 and y, = 155.0. The company will not adopt plastic 1 unless its breaking strength exceeds that of plastic 2 by atleast 10 psi, Based on the sample information, should they use plastic 1? In answering this question, set up and test appropriate hypotheses using a = 0.01, Construct a 99 percent confidence {interval onthe true mean difference in breaking strength. 226. The following are the buming times (in minutes) of chemical flares of two different formulations. The design engineers are interested in both the mean and variance ofthe bbuming times ‘Type ‘Type? 6 2 oo 36 8 or n o 37 9 83 ™ 66 8 9 2 82 7” 6s n () Test the hypothesis thatthe two variances are equal Usea = 005. (b) Using the results of (@), test the hypothesis that the ‘mean burning times are equal. Use a = 0.05. What is the P-value for this test?Design and Analysis of Experiments, Eighth Edition (Chaptor 2 ml Simple Comparative Experiments (6) Discuss the role of the normality assumption in this problem. Check the assumption of normality for both types of flares. 227. An article in Solid State Technology, “Orthogonal Design for Process Optimization and Its Application to Plasma Etching” by G. Z. Yin and D. W. Jillie (May 1987) describes an experiment to determine the effect of the CF flow rate on the uniformity of the etch on 2 silicon wafer used in integrated circuit manufacturing. All of the runs were made in random order. Data for two flow rates are as follows: GR, Flow __ Uniformity Observation jccm 1 2 3 4 #5 6 ps 27 46 26 «30 32 38 200 46 44 29 35 41 51 (a) Does the Cs, low rate affect average etch uniformi- ty? Use a = 005. () Whats the P-value forthe test in part (@)? (©) Does the CF, flow rate affect the wafer-to-wafer vari- ability in etch uniformity? Use a= 0.05. (@ Draw box plots to assist in the interpretation of the dat from this experimen. 228. Anew filtering device is installed in a chemical unit. Before its installation, a random sample yielded the follow. {ng information about the percentage of impurity: j, = 123, S = 101.17,and n= 8. After installation, a random sample Yyeded y, = 10.2, 53 = 94.73,5= 9. (a) Can you conclude thatthe two varlances are equal? Use = 00 (©) Has the filtering device reduced the percentage of impurity significantly? Use a = 0.05, 229, Photoresst is a light-sensitive material applied to semiconductor wafers so that the circuit pattern can be imaged on tothe wafer. After application the coated wafers are baked to remove the solvent in the photoresist mixture and to harden the resist. Here are measurements of photore- sist thickness (in KA) for eight wafers baked at two differ- ent temperatures, Assume tht all ofthe runs were made in random onde. eC 100°C 11176 5263) 7.089 6748 8097 7461 11739 7015 11291 8.133 10.759 7418 6461 3772 8315 8963 (a) Is there evidence to support the claim thatthe high- er baking temperature results in wafers with a lower ‘mean photoresist thickness? Use a = 0.05 (b) Whats the P-value forthe test conducted in part (a)? (© Find 295 percent confidence interval onthe difference in ‘means. Provide a practical interpretation of this interval. (@) Draw dot diagrams to assist in interpreting the results from this experiment (©) Check the assumption of normality of the photoresist thickness. (fH Find the power ofthis test for detecting an actual dif- ference in means of 25 kA. (@) What sample size would be necessary to detect an actual difference in means of 1.3 KA with a power of atleast 0.92 230, Front housings for cell phones are manufactured in ‘an injection molding process. The time the partis allowed to cool in the mold before removal is thought to influence the occurrence of a particularly troublesome cosmetic defect, flow lines in the finished housing. After manufac- turing, the housings are inspected visually and assigned a score between I and 10 based on their appearance, with 10 corresponding to a perfect part and 1 corresponding t0 completely defective part. An experiment was conducted using two cool-down times, 10 and 20 seconds, and 20 housings were evaluated at each level of cool-down time. ‘AIL 40 observations in this experiment were run ia random ‘order. The data are as follows. 3 i 5 (a) Is there evidence to support the claim thatthe longer cool-down time results in fewer appearance defects? Usea = 005. (b) What is the P-value forthe test conducted in part (a)? (©) Find a 95 percent confidence interval on te difference in means. Provide a practical interpretation of this interval, (@) Draw dot diagrams to assist in interpreting the results from this experiment (© Check the assumption of normality for the data from this experimentDesign and Analysis of Experiments, Eighth Edition 231. Twenty observations on etch uniformity on silicon ‘wafers are taken during a qualification experiment fora plas- sma etcher. The data areas follows: 534 665476598728 600785 S54 58261 so] 738 S48 439498 525 635461 = 600532. (@) Construct a 95 percent confidence interval estimate of. (b) Test the hypothesis that «* = 1,0, Use a = 005. What are your conclusions? (© Discuss the nomality assumption and its role in this problem. (@) Check normality by constructing a normal probability plot. What are your conclusions? 232. The diameter of a ball bearing was measured by 12 inspectors, each using two different kinds of calipers. The results were Inspector Caliper T Caliper 2 1 0265 0264 2 0263 0265 3 0266 0264 a; 0267 0266 5 0267 0267 6 0265 0268 1 0267 0264 8 0267 0265 9 0265 0265 10 0268 0267 " 0268 0268 2 0263 0269 27 Problems 63 3 799 842 4 744 684 5 648 643 6 799 876 1 632 632 8 160 762 9 603 659 10 152 761 (a) Is the assumption that the difference in score is nor- rally distributed reasonable? (b) Find a 95% confidence interval on the difference in ‘mean score. Is there any evidence that mean score depends on birth order? (© Test an appropriate set of hypotheses indicating that the mean score does not depend on birth order. 234, Am article in the Journal of Strain Analysis (vol. 18, ‘no. 2, 1983) compares several procedures for predicting the sear strength for ste! plate girders. Data for nine girders in the form of the ratio of predicted to observed load for two of these procedures, the Karlsruhe and Lehigh methods, ae as follows Girder Karlsruhe Method Lehigh Method si 1186 1.061 su Last 0.992 sal 1322 1.063 san 1339 1.062 ssi 1200 106s su 1402 1.178 22 1365 1037 523 1537 1.086 S24 1539 Losz (a) Is there a significant difference between the means of the population of measurements from which the two samples were selected? Use a = 0.05. (b) Find the P-value forthe test in part (a). (© Construct a 95 percent confidence interval onthe dif- ference in mean diameter measurements for the two types of calipers 233, _Anarticle inthe joumal Newrofogy (1998, Vol. 50, pp. 1246-1252) observed that monozygotic twins share numerous physical, psychological, and pathological traits. The investi- ators measured an intelligence score of 10 pairs of twins. "The data obtained are as follows: Pair Birth Order: 1 Birth Order: 2 1 6.08, 373 a 622 5.30 (a) Is there any evidence to support a claim that there is a Gifference in mean performance between the two methods? Use a = 005. (b) What isthe P-value forthe tet in par (8)? (©) Construct a 95 percent confidence interval forthe dif ference in mean predicted to observed load. (@) Investigate the normality assumption for both samples. (© Investigate the normality assumption forthe difference in ratios for the two methods. (H Discuss the role of the normality assumption in the paired test, 2235. The deflection temperature under load for two difer- cent formulations of ABS plastic pipe is being studied. Two samples of 12 observations each are prepared using each for- mulation and the deflection temperatures (in ‘F) are reported below: anDesign and Analysis of Experiments, Eighth Edition 64 chapter 2 m Simple Comparative Experiments Formalation 1 Formulation 2 26 ~«193=«1STTS«THSC«R 188 2072101978588. 205 «185 «19420620089 187189) 178201197203, (@) Construct normal probability plots for both samples. Do these plots support assumptions of normality and ‘equal variance for both samples? (b) Do the data support the claim that the mean deflection temperature under load for formulation 1 exceeds that of formulation 2? Use a = 0.05. (©) What isthe P-value forthe test in part (2)? 236. Refer tothe data in Problem 2.35. Do the data support ‘claim that the mean deflection temperature under load for formulation 1 exceeds tha of formulation 2 by atleast 3°F? 237. In semiconductor manufacturing wet chemical ‘etching is often used to remove silicon from the backs of wafers prior to metaizaton. The etch rate is an important characteristic of this process. Two different etching solutions fae being evaluated. Eight randomly selected wafers have been etched in each solution, and the observed etch rates (in mils min) are as follows. Solution Solution 2 99 106 102 106 94 103 100 102 10 93 107 104 03 98 105 103 (a) Do the data indicate thatthe claim that both solutions have the same mean etch rate is valid? Use a = 0.05, and assume equal variances, () Find a 95 percent confidence interval on te difference inmean etch rates. (6) Use normal probability plots to investigate the adequa- cy of the assumptions of normality and equal variances. 238. ‘Two popular pain medications are being compared on the basis of the speed of absorption by the body. Specifically, tablet 1 is claimed to be absorbed twice as fast as tablet 2. Assume that? and o are known. Develop atest statistic for By:2my = os 2s # hy 239. Continuation of Problem 2.38. An article in Nate (1972, pp. 225-226) reported on the levels of monoamine oxi- ‘dase in blood platelets for a sample of 43 schizophrenic patients resulting in J; = 2.69 and s, = 2.30 while for a sam- ple of 45 normal patients the results were Ys = 6.35 and s) = 403. The units are am/mg protein’. Use the results of the previous problem to test the claim thatthe mean monoamine ‘das level for normal patients is at lst twice the mean level for schizophrenic patients. Assume thatthe sample sizes are lege enough to use the sample standard deviations as the tue parameter values. 2A0. Suppose we are testing Hosa. Hm * a where «7 > 0} are known, Our sampling resources are con- strained such that n, + m= N. Show that an allocation ofthe observation tothe two samp that lead the most powerful testis in the ratio m/ns = oy, 241. Continuation of Problem 240. Suppose that we ‘want to construct a 95% two-sided confidence interval onthe ifference in two means wiere the two sample standard devi- ations are known tobe 0 = 4 and o; = 8. The total sample size is restricted to N = 30. What isthe length of the 95% CL ifthe sample sizes used by the experimenter are n= m= 15? How much shorter would the 95% CI have been if the exper- menter bad used an optimal sample size allocation? 242, Develop Equation 2.46 fora 100(1 ~ a) percent con- fidence imerval forthe vaiance of a normal distribution. 243, Develop Equation 2.0 fora 100(1 ~ a) percent con- fidence interval for the ratio r/o, where and a are the variances of two normal distributions 244, Develop an equation for finding a 100 (1 ~ a) percent confidence interval on the difference inthe means of two nor- ral distributions where 2} Z 0}. Apply your equation tothe Portland cement experiment data, and find a 95 percent coni- ence interval. 2AS, Construct a data set for which the paired t-test satis- tics very large, but for which the usual two-sample or pooled ‘est statistic is smal. In general, describe how you created the data. Does this give you any insight regarding how the paired rest works? 246, Consider the experiment described in Problem 2.26. If the mean burning times of the two flares differ by as mach as 2 minutes, find the power of the test. What sample size ‘would be required to detect an actual difference in mean burn- ing time of 1 minute witha power of atleast 0.90? 247, Reconsider the bottle filing experiment described in Problem 2.24. Rework this problem assuming that the two population variances are unknown but equal 248, Consider the data from Problem 2.24. Ifthe mean il Yolume of the tvo machines differ by s much as 0.25 ounces, ‘what is the power ofthe test used in Problem 2.19? What sam ple size would result in a power of at least 0.9 if the actual dif- ference in mean fill volume is 0.25 ounces?