252y0313a 10/02/03 Don’t waste paper by copying more of this document than you need. IV. Solution to first problem on TAKE HOME SECTION 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. 1. An airline president is tracking late arrivals and believes that the proportion is at most p 0 . Suppose that a sample of 200 flights is selected and 82 were late. Do the following: a) To find p 0 , take the third digit of your Social Security Number divide by 100 and add it to .30. For example, my Social Security Number is 265398248 and the third digit is 5, so the value of p 0 that you would use is 30 5% or .35. (no point credit for this section.) b) Using the value of p 0 that you found in a) prepare to conduct a test to determine whether there is evidence that the proportion is at most p 0 by stating your null and alternative hypotheses. (1) c) Find a critical value for the sample proportion for the hypotheses in b), using a significance level of 1% specify what your ‘reject’ region is and use it to test the null hypothesis. (2) d) (Extra credit) Assume that the actual population proportion is p 0 .03 . (Since my p 0 was .35, I would assume that p1 was .38) , find the power of the test in c). If I had used a lower significance level, explain whether the power would be higher, lower, or the same. (3.5) e) Compute a test ratio for the hypotheses in c) and test the hypotheses using a significance level of 1% (2) f) Use your test ratio in e) to get a p-value for the hypothesis in c) and explain whether and why you would reject the null hypothesis if the significance level was 3%. (2) g) Test the hypotheses in c) using an appropriate confidence interval and a significance level of 1% (2) h) If you were doing a 2-sided 99% confidence interval for the proportion of flights that were late and wanted the proportion to be known within .01 , how large a sample would you use if you expected the proportion to be about 20%? What if you thought the proportion was about 4%? (2) i) (Extra credit) do a power curve for the test in c), using a few carefully chosen values of p1 that are above your p 0 . (4.5) Solution: From the formula table we have: Interval for Confidence Hypotheses Interval Proportion p p z 2 s p H 0 : p p0 pq n q 1 p sp H1 : p p0 Test Ratio z p p0 p Critical Value pcv p0 z 2 p p0 q0 n q0 1 p0 p Note: Rather than compute the standard error of p more than twice in each solution, I have inserted a table p1 q1 at the end of this document. To use the solutions find your value of p 0 . A 2-page n solution for each version of the problem will follow. of p 1 252y0313a 10/02/03 a) p0 .30 The problem says that n 200 , x 82 and p sp pq n x 82 .41 . q 1 p .59. so, for later n 200 .41.59 .0012095 .0347779 200 p0 q0 .30 .70 .00105 .0324037 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .30, H1 : p .30 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .30 2.327.0324037 .30 .0754 .3754. The ‘reject, region must be above .3754. Make a diagram showing a Normal curve with a mean at p0 .30 and a shaded reject region above .3754. Since p .41 is in this ‘reject’ region, reject H 0 . p1 q1 .33.67 .0011055 .0332491 n 200 not reject H 0 : p .30 if p pcv .3754 . So P p .3754 p .33 d) p1 p0 .03 .30 .03 .33 . From this p We do .3754 .33 P z Pz 1.37 .5 .4147 .9147 . Power 1 .0853. .0332491 Since pcv p0 z p .30 2.327.0324037 .30 .0754 .3754, if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0324037 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a mean of zero and shade the area above 2.327. Since z p p0 p .41 .30 3.39 is in the ‘reject region, .0324037 we reject H 0 . f) Since this is a right-tail test, p value Pz 3.39 .5 .4997 .0003 . Since this p-value is below 3%, reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .30 is not on this interval, reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 2 252y0313a 10/02/03 i) We have for the following values of 1 : .30, .33, .3754, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .41 and .45. Our table will read as below, with supporting computations following. 1 Power 1 .30 .99 .01 .33 .9147 .0853 .3754 .5 .5 .41 .1736 .8264 .45 .0197 .9803 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .30 and .45 and graph the ‘Power’ points. P p .3754 p .41 P z .3754 .41 Pz 0.94 .5 .3264 .1736 . .0347779 We can write P p .3754 p .45 P z .3754 .45 Pz 2.06 .5 .4803 .0197 . .0351781 3 252y0313a 10/02/03 a) p0 .31 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .31.69 .00106950 .0327032 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .31, H1 : p .31 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .31 2.327.0327032 .31 .0761 .3861. The ‘reject, region must be above .3861. Make a diagram showing a Normal curve with a mean at p0 .31 and a shaded reject region above .3861. Since p .41 is in this ‘reject’ region, reject H 0 . d) p1 p0 .03 .31 .03 .34 . From this p .34 .66 .001122 .0334963 200 p1q1 n We do not reject H 0 : p .31 if p p cv .3861 . So P p .3861 p .34 .3861 .34 P z Pz 1.38 .5 .4162 .9162 . Power 1 .0838. .0334963 Since pcv p0 z p .31 2.327.0327032 .31 .0761 .3861, if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0327032 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a mean of zero and shade the area above 2.327. Since z p p0 p .41 .31 3.06 is in the ‘reject region, .0327032 we reject H 0 . f) Since this is a right-tail test, p value Pz 3.06 .5 .4989 .0011 . Since this p-value is below 3%, reject H 0 . g) Since the alternate hypotheses is H1 : p .31 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .31 is not on this interval, reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 4 252y0313a 10/02/03 i) We have for the following values of 1 : .31, .34, .3861, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .42 and .46. Our table will read as below, with supporting computations following. 1 Power 1 .31 .99 .01 .34 .9162 .0838 .3861 .5 .5 .42 .1660 .8340 .46 .0179 .9821 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .31 and .50 and graph the ‘Power’ points. P p .3861 p .42 P z .3861 .42 Pz 0.97 .5 .3340 .1660 . .0348999 We can write P p .3861 p .46 P z .3861 .46 Pz 2.10 .5 .4821 .0179 . .0352420 5 252y0313a 10/02/03 a) p0 .32 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .32 .68 .001088 .0329848 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .32, H1 : p .32 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .32 2.327.0329848 .30 .0768 .3968. The ‘reject, region must be above .3968. Make a diagram showing a Normal curve with a mean at p0 .32 and a shaded reject region above .3968. Since p .41 is in this ‘reject’ region, reject H 0 . d) p1 p0 .03 .32 .03 .35 . From this p p1q1 .35 .65 .0011375 .0337268 n 200 We do not reject H 0 : p .32 if p p cv .3968 . So P p .3968 p .35 .3968 .35 P z Pz 1.38 .5 .4162 .9162 . Power 1 .0838. .0337268 Since pcv p0 z p .32 2.327.0329848 .30 .0768 .3968. if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0329848 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a p p0 .41 .32 2.73 is in the ‘reject region, mean of zero and shade the area above 2.327. Since z p .0329848 we reject H 0 . f) Since this is a right-tail test, p value Pz 2.73 .5 .4968 .0032 . Since this p-value is below 3%, reject H 0 . g) Since the alternate hypotheses is H 1 : p .32 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .32 is not on this interval, reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 6 252y0313a 10/02/03 i) We have for the following values of 1 : .32, .35, .3968, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .43 and .47. Our table will read as below, with supporting computations following. 1 Power 1 .32 .99 .01 .35 .9162 .0838 .3968 .5 .5 .43 .1771 .8289 .47 .0192 .9808 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .32 and .50 and graph the ‘Power’ points. P p .3968 p .43 P z .3968 .43 Pz 0.95 .5 .3289 .1771 . .0350071 We can write P p .3968 p .47 P z .3968 .47 Pz 2.07 .5 .4808 .0192 . .0352916 7 252y0313a 10/02/03 a) p0 .33 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .33.67 .0011055 .0332491 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .33, H1 : p .33 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .33 2.327.0332491 .33 .0774 .4074. The ‘reject, region must be above .4074. Make a diagram showing a Normal curve with a mean at p0 .33 and a shaded reject region above .4074. Since p .41 is in this ‘reject’ region, reject H 0 . d) p1 p0 .03 .33 .03 .36 . From this p .36 .64 .001152 .0339411 200 p1q1 n We do not reject H 0 : p .33 if p pcv .4074 . So P p .4074 p .36 .4074 .36 P z Pz 1.39 .5 .4177 .9177 . Power 1 .0823. .033941 Since pcv p0 z p .33 2.327.0332491 .33 .0774 .4074. if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0324491 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a mean of zero and shade the area above 2.327. Since z p p0 p .41 .33 2.41 is in the ‘reject region, .0332491 we reject H 0 . f) Since this is a right-tail test, p value Pz 2.41 .5 .4920 .0080 . Since this p-value is below 3%, reject H 0 . g) Since the alternate hypotheses is H 1 : p .33 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .33 is on this interval, do not reject H 0 . This is an unusual result, which I knew was possible, but had never seen before. If we use p .0324491 in place of s p .0347779, we get a rejection. h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 8 252y0313a 10/02/03 i) We have for the following values of 1 : .33, .36, .4074, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .44 and .48. Our table will read as below, with supporting computations following. 1 Power 1 .33 .99 .01 .36 .9177 .0823 .4074 .5 .5 .44 .1762 .8238 .48 .0197 .9803 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .33 and .50 and graph the ‘Power’ points. P p .4074 p .44 P z .4074 .44 Pz 0.93 .5 .3238 .1762 . .0350999 We can write P p .4074 p .48 P z .4074 .48 Pz 2.06 .5 .4803 .0197 . .0353270 9 252y0313a 10/02/03 a) p0 .34 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .34 .66 .001122 .0334963 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .34, H1 : p .34 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .34 2.327.0334963 .34 .0779 .4179. The ‘reject, region must be above .4179. Make a diagram showing a Normal curve with a mean at p0 .34 and a shaded reject region above .4179. Since p .41 is not in this ‘reject’ region, do not reject H 0 . d) p1 p0 .03 .34 .03 .37 . From this p .37 .63 .0011655 .0341394 200 p1q1 n We do not reject H 0 : p .34 if p pcv .4179 . So P p .4179 p .37 .4179 .37 P z Pz 1.40 .5 .4192 .9192 . Power 1 .0808. .0341394 Since pcv p0 z p .34 2.327.0334963 .34 .0779 .4179. if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0334963 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a mean of zero and shade the area above 2.327. Since z p p0 p .41 .34 2.09 is not in the ‘reject .0334963 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 2.09 .5 .4817 .0183 . Since this p-value is below 3%, reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .34 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 10 252y0313a 10/02/03 i) We have for the following values of 1 : .34, .37, .4179, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .45 and .49. Our table will read as below, with supporting computations following. 1 Power 1 .34 .99 .01 .37 .9177 .0823 .4179 .5 .5 .45 .1762 .8238 .49 .0207 .9793 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .34 and .50 and graph the ‘Power’ points. P p .4179 p .45 P z .4174 .45 Pz 0.93 .5 .3238 .1762 . .0351781 We can write P p .4179 p .49 P z .4179 .49 Pz 2.04 .5 .4793 .0207 . .0353483 11 252y0313a 10/02/03 a) p0 .35 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .35.65 .0011375 .0337268 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .35 2.327.0337268 .35 .0785 .4285. The ‘reject, region must be above .4285. b) H 0 : p .35, H 1 : p .35 and, for later, p Make a diagram showing a Normal curve with a mean at p0 .35 and a shaded reject region above .4285. Since p .41 is not in this ‘reject’ region, do not reject H 0 . p1 q1 .38.62 .0011780 .0343220 n 200 ..4285 . So P p .4285 p .38 d) p1 p 0 .03 .35 .03 .38. From this p not reject H 0 : p .35 if p p cv We do .4285 .38 P z Pz 1.41 .5 .4207 .9207 . Power 1 .0793. .034322 Since pcv p0 z p .35 2.327.0337268 .35 .0785 .4285, if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0337268 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a p p0 .41 .35 1.78 is not in the ‘reject mean of zero and shade the area above 2.327. Since z p .0337268 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 1.78 .5 .4625 .0375 . Since this p-value is above 3%, do not reject H 0 . g) Since the alternate hypotheses is H 1 : p .35 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p 0 .35 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 12 252y0313a 10/02/03 i) We have for the following values of 1 : .35, .38, .4285, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .46 and .50. Our table will read as below, with supporting computations following. 1 Power 1 .35 .99 .01 .38 .9207 .0793 .4285 .5 .5 .46 .1867 .8133 .50 .0217 .9783 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .35 and .55 and graph the ‘Power’ points. P p .4285 p .46 P z .4285 .46 Pz 0.89 .5 .3133 .1867 . .0352420 We can write P p .4285 p .50 P z .4285 .50 Pz 2.02 .5 .4783 .0217 . .0353553 13 252y0313a 10/02/03 a) p0 .36 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .36.64 .0011520 .0339411 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .36 2.327.0339411 .36 .0790 .4390. The ‘reject, region must be above .4390. b) H 0 : p .36, H 1 : p .36 and, for later, p Make a diagram showing a Normal curve with a mean at p0 .30 and a shaded reject region above .4390. Since p .41 is not in this ‘reject’ region, do not reject H 0 . p1 q1 .39.61 .0011895 .0347779 n 200 .4390 . So P p .4390 p .39 d) p1 p 0 .03 .36 .03 .39 . From this p not reject H 0 : p .36 if p p cv We do .4390 .39 P z Pz 1.41 .5 .4207 .9207 . Power 1 .0793. .0347779 Since pcv p0 z p .36 2.327.0339411 .36 .0790 .4390. if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0324037 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a p p0 .41 .36 1.47 is not in the ‘reject mean of zero and shade the area above 2.327. Since z p .0339411 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 1.47 .5 .4292 .0708 . Since this p-value is above 3%, do not reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p 0 .36 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 14 252y0313a 10/02/03 i) We have for the following values of 1 : .36, .39, .4390, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .47 and .51. Our table will read as below, with supporting computations following. 1 Power 1 .36 .99 .01 .39 .9207 .0793 .4390 .5 .5 .47 .1894 .8106 .51 .0222 .9778 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .36 and .55 and graph the ‘Power’ points. P p .4390 p .47 P z .4390 .47 Pz 0.88 .5 .3106 .1894 . .0352916 We can write P p .4390 p .51 P z .4390 .51 Pz 2.01 .5 .4778 .0222 . .0353483 15 252y0313a 10/02/03 a) p0 .37 later s p The problem says that n 200 , x 82 and p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .37 .63 .0011655 .0341394 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .37 2.327.0341394 .37 .0794 .4494. The ‘reject, region must be above .4494. b) H 0 : p .37, H 1 : p .37 and, for later, p Make a diagram showing a Normal curve with a mean at p 0 .37 and a shaded reject region above .4494. Since p .41 is not in this ‘reject’ region, do not reject H 0 . p1 q1 .40 .60 .0012000 .0346410 n 200 .4494 . So P p .4494 p .40 d) p1 p 0 .03 .37 .03 .40. From this p not reject H 0 : p .37 if p p cv We do .4494 .40 P z Pz 1.43 .5 .4236 .9236 . Power 1 .0764. .0346410 Since pcv p0 z p .37 2.327.0341394 .37 .0794 .4494, if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0341394 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a p p0 .41 .37 1.17 is not in the ‘reject mean of zero and shade the area above 2.327. Since z p .0341394 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 1.17 .5 .3790 .1210 . Since this p-value is above 3%, do not reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p 0 .37 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 16 252y0313a 10/02/03 i) We have for the following values of 1 : .37, .40, .4494, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .48 and .52. Our table will read as below, with supporting computations following. 1 Power 1 .37 .99 .01 .40 .9236 .0764 .4494 .5 .5 .48 .1922 .8078 .52 .0228 .9772 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .37 and .55 and graph the ‘Power’ points. P p .4494 p .48 P z .4494 .48 Pz 0.87 .5 .3078 .1922 . .0353270 We can write P p .4494 p .52 P z .4494 .52 Pz 2.00 .5 .4772 .0228 . .0353270 17 252y0313a 10/02/03 a) p0 .38 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .38.72 .0011780 .0343220 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .38 2.327.0343220 .38 .0799 .4599. The ‘reject, region must be above .4599. b) H 0 : p .38, H 1 : p .38 and, for later, p Make a diagram showing a Normal curve with a mean at p 0 .38 and a shaded reject region above .4599 Since p .41 is not in this ‘reject’ region, do not reject H 0 . Since pcv p0 z p .38 2.327.0343220 .38 .0799 .4599, if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. p1 q1 .41.69 .0012095 .0347779 n 200 .4599 . So P p .4599 p .41 d) p1 p 0 .03 .38 .03 .41. From this p not reject H 0 : p .38 if p p cv We do .4599 .41 P z Pz 1.43 .5 .4236 .9236 . Power 1 .0764. .0347779 p p0 e) The test ratio formula is z . p .41 . We already know that p .0343220 and that p z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a p p0 .41 .38 0.87 is not in the ‘reject mean of zero and shade the area above 2.327. Since z p .0343220 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 0.87 .5 .3078 .1922 . Since this p-value is above 3%, do not reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p 0 .38 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 18 252y0313a 10/02/03 i) We have for the following values of 1 : .38, .41, .4598, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .49 and .53. Our table will read as below, with supporting computations following. 1 Power 1 .38 .99 .01 .41 .9236 .0764 .4598 .5 .5 .49 .1977 .8023 .53 .0233 .9767 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .38 and .55 and graph the ‘Power’ points. P p .4598 p .49 P z .4598 .49 Pz 0.85 .5 .3023 .1977 . .0353483 We can write P p .4598 p .53 P z .4598 .53 Pz 1.99 .5 .4767 .0233 . .0352916 19 252y0313a 10/02/03 a) p0 .39 The problem says that n 200 , x 82 and p later s p pq n x 82 .41 . q 1 p .59. so, for n 200 .41.59 .0012095 .0347779 200 p0 q0 .39 .61 .0011895 .0347779 n 200 c) Since the alternate hypothesis is H1 : p p0 , the critical value must be above p0 , and we must use a b) H 0 : p .39, H1 : p .39 and, for later, p right-tail test. .01 and z z.01 2.327 . Use pcv p0 z p .39 2.327.0347779 .39 .0809 .4709. The ‘reject, region must be above .4709. Make a diagram showing a Normal curve with a mean at p0 .39 and a shaded reject region above .4709. Since p .41 is not in this ‘reject’ region, do not reject H 0 . d) p1 p0 .03 .39 .03 .42 . From this p .42 .58 .001218 .0348999 200 p1q1 n We do not reject H 0 : p .39 if p pcv .4709 . So P p .4709 p .42 .4709 .42 P z Pz 1.46 .5 .4279 .9279 . Power 1 .0721. .0348999 Since pcv p0 z p .39 2.327.0347779 .39 .0809 .4709. if I had used a lower significance level, I would use a larger value of z , which means a larger value of pcv , which would make my test ratio larger, which would make larger and the power smaller. e) The test ratio formula is z p p0 p . p .41 . We already know that p .0347779 and that z z.01 2.327 . Our ‘reject’ region is above 2.327. We make a diagram showing a Normal curve with a mean of zero and shade the area above 2.327. Since z p p0 p .41 .39 0.575 is not in the ‘reject .0347779 region, we do not reject H 0 . f) Since this is a right-tail test, p value Pz 0.575 .5 .2174 .2826 . Since this p-value is above 3%, do not reject H 0 . g) Since the alternate hypotheses is H1 : p .30 , the confidence interval is p p z s p .41 2.327 .0347779 or p .3291. Since p0 .39 is on this interval, do not reject H 0 . h) From the outline: The usually suggested formula is n forgets that we covered. p .20, e .01, .01, z 2 z.005 2.576 and n p .04, e .01, .01, z 2 z.005 2.576 and pqz 2 e2 , ………… This is the formula everyone .20 .80 2.576 2 .012 .04 .96 2.576 2 n .012 10617 .2 10618 . 2548 .1 2549 . 20 252y0313a 10/02/03 i) We have for the following values of 1 : .39, .42, .4709. Let’s try two points above the critical value, say .51 and .55. i) We have for the following values of 1 : .39, .42, .4709, since for 1 0 , power = , and for 1 xcv , power = 0 . Let’s try two points above the critical value, say .50 and .54. Our table will read as below, with supporting computations following. 1 Power 1 .39 .99 .01 .42 .9279 .0721 .4709 .5 .5 .50 .2061 .7939 .54 .0250 .9750 Graph off your y-axis with intervals of 0.1 for power between 0 and 1.0 and set up your x-axis with divisions of .2 for 1 between .39 and .55 and graph the ‘Power’ points. P p .4709 p .50 P z .4709 .50 Pz 0.82 .5 .2939 .2061 . .0353553 We can write P p .4709 p .54 P z .4709 .54 Pz 1.96 .5 .4750 .0250 . .0352420 21 252y0313a 10/02/03 Table of p n 200 p1 q1 n p1 q1 .30 .31 .32 .33 .34 .35 .36 .37 .38 .39 .40 .41 .42 .43 .44 .45 .46 .47 .48 .49 .50 .51 .52 .53 .54 .55 .70 .69 .68 .67 .66 .65 .64 .63 .62 .41 .60 .59 .58 .57 .56 .55 .54 .53 .52 .51 .50 .49 .48 .47 .46 .45 p1 q1 n 0.00105000 0.00106950 0.00108800 0.00110550 0.00112200 0.00113750 0.00115200 0.00116550 0.00117800 0.00118950 0.00120000 0.00120950 0.00121800 0.00122550 0.00123200 0.00123750 0.00124200 0.00124550 0.00124800 0.00124950 0.00125000 0.00124950 0.00124800 0.00124550 0.00124200 0.00123750 p p1 q1 n 0.0324037 0.0327032 0.0329848 0.0332491 0.0334963 0.0337268 0.0339411 0.0341394 0.0343220 0.0347779 0.0346410 0.0347779 0.0348999 0.0350071 0.0350999 0.0351781 0.0352420 0.0352916 0.0353270 0.0353483 0.0353553 0.0353483 0.0353270 0.0352916 0.0352420 0.0351781 22