252y0112 2/28/01 ECO252 QBA2 FIRST HOUR EXAM February 20, 2001 Name ____key___________ Hour of class registered _____ Class attended if different ____ Show your work! Make Diagrams! Note that since relatively few people took this exam, commentary on 252y0111 is more complete. I. (14 points) Do all the following. x ~ N 3,7 93 9`3 z 1. P9 x 9 P P 1.71 z 0.86 7 7 P1.71 z 0 P0 z 0.86 .4564 .3051 .7615 93 0`3 z 2. P0 x 9 P P 0.43 z 0.86 7 7 P0.43 z 0 P0 z 0.86 .1664 .3051 .4715 28 3 9`3 z 3. P9 x 28 P P0.86 z 3.57 7 7 P0 z 3.57 P0 z 0.86 .4998 .3051 .1947 53 4. Px 5 P z Pz 0.29 7 Pz 0 P0 z 0.29 .5 .1141 .6141 03 53 z 5. P5 x 0 P P 1.14 z 0.43 7 7 P1.41 z 0 P0.43 z 0 .3729 .1664 .2065 6. A symmetrical interval about the mean with 85% probability. We want two points x .075 and x.925 , so that Px.075 x x.925 .8500 . Make a diagram. If you do a diagram for z , it will show two points, z .075 and z .925 . z .925 (which has 92.5% above it!) is below zero and z .075 is above zero. Since zero is halfway between these two points, the diagram will show 85% split between the two sides of zero, so that 42.5% is between z .925 and zero, and 42.5% is between zero and z .075 . The probability below z .925 and the probability above z .075 are both 7.5%. From the diagram, if we replace x by z, P0 z z.075 .4250 . The closest we can come is P0 z 1.44 .4251 . So z.075 1.44 , and x z.075 3 1.447 3 10 .08 or -7.08 to 13.08. 13 .08 3 7.08 3 z To check this note that P7.08 x 13.08 P P 1.44 z 1.44 7 7 2P0 z 1.44 2.4251 .8502 85%. x.135 We want a point x .135 , so that Px x .135 .135 . Make a diagram. The diagram for z will show one point , z .135 which has 13.5% above it (and 86.5% below it!) and is above zero because zero has only 50% below it. Since zero has 50% above it, the diagram will show 36.5% between zero and z .135 . From the diagram, if we replace x by z, P0 z z.135 .3650 . The closest we can come is P0 z 1.10 .3643 . So z.135 1.10 , and x z.135 3 1.107 3 7.70 , or 10.70. 7. 10 .70 3 To check this note that Px 10 .70 P z Pz 1.10 Pz 0 P0 z 1.10 7 .5 .3643 .1357 13.5%. 252y0112 2/21/01 II. (6 points-2 point penalty for not trying part a.) The Watched Potts Bank of Pottstown wishes to monitor the debt-to-equity ratio of the firms to which it has made commercial loans. A random sample of 8 of the firms is taken. Assume that the bank is sampling from a population with a normal distribution. firm 1 2 3 4 5 6 7 8 Debt/equity ratio 1.34 1.04 1.46 1.22 1.20 1.79 1.38 1.42 a. Compute the sample standard deviation, s , of the debt-to-equity ratios. Show your work! (3) b. Compute a 95% confidence interval for the mean debt-to-equity ratio, .(3) Solution: a) x x 10.85 1.35625 n x Firm 1 2 3 4 5 6 7 8 8 15 .0621 81.35625 2 n 1 7 0.0495411 or s 0.2226 . s2 2 Total nx 2 10.85 b) x (Ratio) 1.34 1.04 1.46 1.22 1.20 1.79 1.38 1.42 x2 1.7956 1.0816 2.1316 1.4884 1.4400 3.2041 1.9044 2.0164 15.0621 From the problem statement .05 . From Table 3 of the syllabus supplement, if the x t sx and t n21 t .7025 2.365 . population variance is unknown 2 s 0.0495411 0.2226 sx 0.07869 . So 1.35625 2.365 0.07869 1.356 0.186 or 1.170 8 n 8 to 1.542. 2 252y0112 2/16/01 III. Do at least 3 of the following 4 Problems (at least 10 each) (or do sections adding to at least 30 points Anything extra you do helps, and grades wrap around) . Show your work! State H 0 and H 1 where appropriate. You have not done a hypothesis test unless you have stated your hypotheses, run the numbers and stated your conclusion. Use a 95% confidence level unless another level is specified. Error: On this page, 1.43 should be changed to 1.33. Remove section f!!! 1. The Watched Potts Bank of Pottstown is a subsidiary of the Umongous Bancorp. The Bancorp is threatening to remove the current officers of the bank if the firms to which they are making loans have an average debt-to-equity ratio that exceeds 1.33. Are the officers in trouble? The data are on the previous page. For your convenience the data are repeated below. firm 1 2 3 4 5 6 7 8 Debt/equity ratio 1.34 1.04 1.46 1.22 1.20 1.79 1.38 1.42 Test to see if the mean is above 1.33 using the sample mean and standard deviation you found in part II. a. Choose the correct null hypothesis from the list below and write the alternative hypothesis. (2) H 0 : x 1.33 H 0 : 1.33 H 0 : 1.33 H 0 : x 1.33 H 0 : 1.33 H 0 : x 1.33 H 0 : x 1.33 H 0 : p 1.33 H 0 : 1.33 H 0 : p 1.33 H 0 : 1.33 H 0 : x 1.33 b. Find a critical value appropriate for this problem, using a confidence level of 95%.(3) c. Use your critical value to test the hypothesis. State clearly whether you reject the null hypothesis. (2) d. Repeat the test using (i) a test ratio (2) and (ii) a confidence interval. (2) e. Test the hypothesis that the average debt-to-equity ratio is exactly 1.33 at the 95% confidence level. (2) Solution: From Table 3 of the Syllabus Supplement: Interval for Confidence Hypotheses Interval Mean ( x z 2 x H0 : 0 Known) H : 1 Mean ( Unknown) x t 2 s x Test Ratio z 0 H0 : 0 t Critical Value x 0 xcv 0 z 2 x x 0 sx xcv 0 t 2 s x x H1 : 0 DF n 1 a) We are afraid that 1.33 . Because this does not contain an equality, it cannot be a null hypothesis. So we must test its opposite H 0 : 1.33 and our alternate hypothesis is H1 : 1.33 7 8 1.895 (If you chose a 2-sided test, t .025 2.306 ) b) n 8, 0 1.32, DF n 1 7, .05, tn1 t .05 From the previous page: x 1.35625 and s x s n 0.0495411 0.2226 0.07869 . 8 8 3 252y0112 2/21/01 Critical Value: Since this is a one-sided test, xcv 0 t sx 1.33 1.895 0.07869 1.33 0.1491 or 1.4791. It might help to remember that x 0 is always in the 'accept' region. c) The critical value in b) means that we reject H 0 if the sample mean is above 1.4791. Since x 1.35625 is not above this critical value, do not reject H 0 . d) (i) Test Ratio: t x 0 1.35625 1.33 0.3336 . This is in the ‘accept’ region sx 0.07869 7 below tn 1 t .05 1.895 , so do not reject H 0 . It might help to remember that zero is always in the 'accept' region for t. (ii) Confidence Interval: Since this is a one-sided test, and the alternate hypothesis is H1 : 1.33 , x t sx 1.35625 1.950 0.07869 1.35625 0.14911 or 1.207 . This does not contradict H 0 : 1.33 , because any mean in the range 1.207 to 1.33 satisfies both statements, so do not reject H 0 . 7 e) H 0 : 1.33 , H 1 : 1.33 . n 8, 0 1.33, DF n 1 7, .05, tn1 t.025 2.365 . 2 x 1.34625 and sx 0.07741 . Use one of the three methods below. (i) Critical Value: Since this is a two-sided test, xcv 0 t sx 1.33 2.365 0.07741 1.33 0.1831 2 or 1.1469 and 1.5131. Since x 1.34625 is between these two values, do not reject H 0 . x 0 1.34625 1.32 0.3391 . This is in the ‘accept’ region sx 0.07741 7 7 between tn1 t.025 2.365 and tn1 t.025 2.365 , so do not reject H 0 . (ii) Test Ratio: t 2 2 (iii) Confidence Interval: Since this is a two-sided test, use x t s x 2 1.346 2.365 0.29477 1.346 0.1831 . We can thus say that 1.163 1.529. Since this interval includes 0 1.33, we cannot reject H 0 . 4 252y0112 2/21/01 2. Ken Black tells us that, nationwide, 17% of Americans drink milk as their primary breakfast beverage. A milk producer in Wisconsin believes that the proportion of Wisconsin residents that drink milk for breakfast is above 17%. She takes a survey of 500 Wisconsin residents and finds that 105 drink milk as their primary beverage. Use a 90% confidence level in all parts of this problem. a. Do a 90% confidence interval for the proportion of residents that use milk as their primary beverage. (3) b. Assume that the milk producer wants to estimate the proportion of Wisconsin residents that use milk as their primary beverage within .005 , how large a sample does she need? Assume that the correct proportion is 17% in this section. (3) c. Her original purpose in collecting the data presented above was to test the hypothesis that more than 17% of the people in Wisconsin drink milk as their primary breakfast beverage. Test this hypothesis now. (i) State your null and alternative hypotheses. (2) (ii) Do a test of your null hypothesis using a test ratio and the 99% confidence level (3) (iii) Using your test ratio find a p-value for your null hypothesis. (2) (iv) Do a test of your null hypothesis using a critical value for the observed proportion (2) (v) Do a test of your null hypothesis using an appropriate confidence interval (2) d. (extra credit - this was covered in your text, but not in class) How would you modify your work and conclusion in either (ii) or (iii) in c) if you found that your sample of 500 residents had been taken in a state with only 3000 residents? (3) Solution: From Table 3: Interval for Confidence Interval Proportion p p z 2 s p pq n q 1 p sp Hypotheses Test Ratio H 0 : p p0 H 1 : p p0 z p p0 p Critical Value p cv p 0 z 2 p p p0 q 0 n .210 .790 .01831 , z z 1.645 and x 105 .210 , s p .05 2 500 n 500 p p z s p .21 1.645.01831 .210 .030 or .180 to .240. a) .10 , n 500 , p 2 b) From the outline n n pqz 2 e2 pqz 2 e2 .17 .83 1.645 2 .005 2 . p .17, e .005 and since .10 , use z z.05 1.645 . Thus 2 15272 .805 and we must use a sample of at least 15273. H 0 : p .17 c) (i) H 1 : p .17 p .17 .83 500 (ii) Test Ratio: z Note that p 0 .17 , n 500 , p x 105 .210 , so that n 500 .0002822 .016799 . q 0 1 p 0 . p p0 p .210 .17 2.381 . Since .10 , and this is a one-sided test, .016799 use z .10 1.282 . Since 2.381 is more than z .10 , reject H 0 . Make a diagram. Show a normal curve with a mean at 0 and a 'reject' region starting at z .10 1.282 and show where 2.381 falls on this diagram. (iii) pvalue P p .210 Pz 2.38 .5 .4613 .0087 , so reject H 0 . 5 252y0112 2/21/01 (iv) Critical Value: pcv p0 z p .17 1.282.016799 .1915 . Since p .210 is above this value, reject H 0 . Make a diagram. Show a normal curve with a mean at 0.17 and a 'reject' region starting at .1915 and show where .210 falls on this diagram. (v) Confidence Interval: Since the alternate hypothesis is H1 : p .17 , the one-sided confidence interval will be p p z s p .210 1.282.01831 .1865 . Since p .1865 contradicts H0 : p .17 , reject H 0 . This is an extremely rare case. Make a diagram. Show a normal curve with a mean at .210 and represent the confidence interval with a region above .1865. Represent the null hypothesis with a region below .17. show that they do not overlap. d) Because the population size is now N 3000 , which is less than 20 times the sample size, n 550 , we need a finite population correction factor. With this factor, the formula for the variance of the sample proportion becomes p0 q0 N n .17 .83 3000 500 .1411 2500 .000235245 .01534 . n N 1 550 3000 1 550 2999 Note that the text on page 324 omits the -1. p p 0 .210 .17 2.607 . Since .10 and this is a one-sided test, (ii) Test Ratio: z p .01534 p use z .10 1.282 . Since 2.607 is more than z .10 , reject H 0 . Or (iii) pvalue P p .210 Pz 2.61 .5 .4955 .0045 , so reject H 0 . The conclusions only became stronger. 6 252y0112 2/21/01 3. Ken Black says that a soft drink manufacturer takes a sample of 50 12-ounce cans of soda to check that the mean is at least 12 ounces and finds that the sample mean is 11.985 ounces. From long experience, the producer assumes that the population standard deviation is 0.10 ounces. Use a 99% confidence level in a) d) in this problem. a. State the null hypothesis and find the p-value for it. (3) b. Find a critical value for the sample mean. (2) c. What would the power of the test in b) be if the mean was, in fact, 11.98 ounces? (3) d. Doing appropriate calculations, find and sketch the power curve for the test in b). (5) e. Do a 73% confidence interval for the mean based on the sample mean of 11.985 (3) Solution: a) First, state the problem and find a critical value or values. H 0 : 12 0.10 0.10, n 60, .01 so x 0.014142 . z z.01 2.327 . n 50 H 1 : 12 x 0 11 .985 12 1.06 and the p-value is Px 11.985 Pz 1.06 .5 .3554 .1446 . .014142 Note: Since last time you had to use the t table to find the p-value, most of you assumed that you had to do that this time too. Think! Since you were given a population variance, you should have used z . Moreover, just because you have to double a p-value in a 2-sided test, doesn't mean you should double it in a 1-sided test. z x b) Since this is a one-sided test, the formula for a two-sided critical value x cv 0 z x becomes 2 xcv 0 z x , so that xcv 12 2.327 0.014142 11.9671 . So we will not reject H 0 if the sample mean x is greater than or equal to 11.9671. Make a diagram. Put 12 in the middle and show a 'reject' region below 11.9671. Shade this area to represent the significance level. c) Since a type II error is wrongly ‘accepting’ the null hypothesis, we compute the probability that the sample mean will be above or equal to the critical value for each value of 1 . Our computations are below. x 1 Note that, in general, for this one-sided hypothesis P z cv . Here 1 11 .98 , so x x 1 11 .9671 11 .98 P z cv Pz Pz 0.91 .5 .3186 .8186 . The power is 1 - .8186 x 0.014142 = .1814. d) Decide on what values of 1 to use to compute , the probability of a type II error. The usual set of values includes the mean from the null hypothesis, (12), the critical value (11.9671), a point about midway between these values (11.98) and two points, one further out beyond the critical value by a distance about equal to the distance between the null hypothesis mean and the critical value (11.93), and another roughly halfway between this point and the critical value (11.95). Compute for each value of 1 . Since a type II error is wrongly ‘accepting’ the null hypothesis, we compute the probability that the sample mean will be above or equal to the critical value for each value of 1 . Our computations are below. Note that, in general, for this one-sided hypothesis xcv 1 . x We know that if 1 12 , we will find 1 .99 and the power will be .01. At the critical value, .5 and the power is also .5. Pz 7 252y0112 2/28/01 1 11 .95 Px 11 .9671 11 .95 P z 11 .9671 11 .95 Pz 1.21 .5 .3869 .1131 .014142 power 1 .8869 1 11 .93 Px 11 .9671 11 .93 P z 11 .9671 11 .93 Pz 2.62 .5 .4956 .0044 .014142 power 1 .9956 You now have the power for 5 points below or equal to 12. Make a diagram. e) From page 3 of this document, x z x , and we know that 2 x x 11.985 and 0.10 0.014142 . 1 .73 .27 . z z.135 . To find z.135 (You found this on page 1!), 2 n 50 make a diagram. Show the mean at zero, that the probability above z.135 is .135, so that the probability between zero and z.135 must be 50% - 13.5% = 36.5%. So P0 z z.135 .3650 . The best we can do on the Normal table is So and P0 z 1.10 .3643 . z.135 1.10 , x z 2 x 11.985 1.100.014142 11.985 0.016 or 11.969 to 11.1156. 8 252y0112 2/21/01 4. (Use a 99% confidence level in all parts of this problem.) My old friend Ken Black talks about a worker who has an average of 20 tubes at his work station. To satisfy the requirements of just-in-time inventory, the superintendent wants the standard deviation of the number of tubes at the station to be less than or equal to 3. The station is visited 8 times and, from the eight measurements taken, she calculates a sample mean number of tubes at the station of 20.875 and a sample standard deviation of 4.5806. a. For starters, do a 2-sided test of the statement that the population standard deviation is 3. (3) b. Do a confidence interval for the standard deviation. (3) c. Actually the test in a) should have been a 1-sided test, so do it again. (2) d. The superintendent visits the work station 61 times during the month. She now calculates the mean at 20.475 and the sample standard deviation at 3.4520. Repeat the test in a) with these new numbers. (2) e. Now repeat the test in c) with these new numbers. (2) f. A paper company will cut farmed timber only if the median height of the trees is over 40 feet. A sample of 40 trees is taken and 27 of them are over 40 feet. State your null and alternate hypotheses and tell whether we would reject the null hypothesis at the 99% confidence level. Finally, tell whether your conclusion means that they do or do not cut. (4) g. I am sure that the Friendlyville, Alabama police department is writing an average of more than 25 speeding tickets a day on its one-block-long main street. Sho' 'nuff , in a one-day visit I see 38 tickets written. Assuming that the Poisson distribution applies, state the null and alternative hypotheses and test the null hypothesis. (3) h. (Extra Credit) We agree that one day is not enough time to test the hypotheses in g) and so I stay for a full week. In that time I see 185 tickets written. Repeat the test. (3) H 0 : 2 9 H : 3 n 8, DF n 1 7, s 4.5806 , .01, and 2 .005 . Solution: a) 0 or H1 : 2 9 H1 : 3 From the outline, the test ratio formula is 2 n 1s 2 02 74.5806 2 16 .31925 . From the chi-square 9 7 20.2778 and 27 0.9893 . Make a diagram. Show a chi-squared distribution with a mean table, .2005 ..995 at about DF 7 , and a 99% 'accept region between the two values of chi-squared just given. Show one 2.5% 'reject' region below the 99.5% value and a second 'reject' region above the 0.5% value. Since 16.31925 falls in the 'accept' region, do not reject H 0 . b) According to the outline, a two-sided confidence interval would be 74.5806 2 2 74.5806 2 n 1s 2 22 2 n 1s 2 12 2 . If we fill or 7.2430 2 148 .4618 . If we take square 20 .2778 0.9893 roots, we get 2.6913 12.1895 . H 0 : 2 9 H 0 : 3 n 8, DF n 1 7, s 4.5806 and .01 . Make a diagram. Show a c) or H1 : 2 9 H1 : 3 chi-squared distribution with a mean at about DF 7 , and a 99% 'accept region below 27 18.4753 . in the blanks, we get Show the 1% 'reject' region above the 1% value of chi-squared. Since 2 n 1s 02 .01 2 = 16.31925 falls in the 'accept' region, do not reject H 0 . 9 252y0112 2/21/01 H 0 : 2 9 H : 3 d) 0 or H1 : 2 9 H1 : 3 2 n 1s 2 02 n 61, DF n 1 60 , s 3.4520 .01, and 2 .005 . 60 3.4520 2 79 .4420 . For large samples z 2 2 2DF 1 9 279 .4420 260 1 158 .88 119 12 .6049 10 .9081 1.6920 . Make a diagram. Show a normal curve with a mean at 0 and 'reject' regions below z.005 2.576 and above z.005 2.576 and show where 1.6920 falls on this diagram. Do not reject H 0 . H 0 : 2 9 H : 3 e) 0 or . Make a diagram. Show a normal curve with a mean at 0 and a 'reject' H1 : 2 9 H1 : 3 region above z.01 2.327 and show where 1.6920 falls on this diagram. Do not reject H 0 . f) From outline point B6a for the sign test for a median: Hypothesis about Hypotheses about a proportion a median If p is the proportion If p is the proportion above 0 H 0 : 0 H 1 : 0 H 0 : p .5 H 1 : p .5 below 0 H 0 : p .5 H 1 : p .5 H : 40 H : p .5 We are told that x 27 are above 40 feet and that n 40 . So 0 becomes 0 and H1 : 40 H 1 : p .5 there are several methods to do this in B6 and the problem solutions. If we use the formula used by the text 2x 1 n (quoted as z in the outline) without a continuity correction we find pval Px 27 n 227 40 Pz 2.21 .5 .4864 .0136 . Since this is below .01, we reject H 0 and cut P z 40 the timber. H : 25 g) 0 From the Poisson(25) table, pval Px 38 1 Px 37 1 .99076 .00924 . Since H1 : 25 this is below .01, we reject H 0 . H : 175 h) 0 This time we follow Problem B4d) and say (using the Normal approximation to the H1 : 175 185 175 Pz 0.74 .5 .2704 .2296 . Since this Poisson distribution) pval Px 185 P z 175 is not below .01, we do not reject H 0 . 10