4/14/98 252y9831 ECO252 QBA2 THIRD HOUR EXAM April 16, 1998 Name Hour of Class Registered (Circle) MWF 10 11 TR 12:30 2:00 Hour of Class Attended (If Different) ______________ I. (16 points) Do all the following. 1. Hand in your computer printouts.(6) 2. Do not do the following unless you handed in all three outputs: year 1 2 3 4 5 6 7 8 visitors 4068 3934 3931 4144 4780 5760 6457 5126 Fper$ 4.2566 4.2250 5.4396 6.5793 7.6203 8.7355 8.9799 6.9256 The material at the left is the number of visitors (in thousands) and the francs per dollar exchange rate for a sample of years. Two Minitab procedures were run on the data. There is no need for you to do any calculations from the original data. a)My first test read as follows: MTB ttest mu=5000 'visitors' SUBC alt=-1. TEST OF MU=5000 VS MU L.T. 5000 N MEAN STDEV SE MEAN visitors 8 4775 944 334 T -0.67 P VALUE 0.26 In the language used in class: (i) What are H 0 and H1 (1) (Do not do any unnecessary calculations) (ii) If .05 do we accept H 0 ? Why? (1.5) (iii) Sketch a diagram of the t-distribution to show the meaning of the p-value (2) H 0 : 5000 Solution: (i) H1 : 5000 (ii) Accept H 0 if pval so accept H 0 or accept H 0 if t lies above t , so accept H 0 . (iii) Note that a p-value is a probability, not just a point on the x axis. 4/14/98 252y9831 b)My next test read as follows: MTB regress 'visitors' on 1 'Fper$' 'resid' 'pred' The regression equation is visitors = 1768 + 456 Fper$ Predictor Constant Fper$ s = 462.6 Coef Stdev 1768.5 645.7 455.86 94.72 R-sq = 79.4% t-ratio p 2.74 0.034 4.81 0.003 R-sq (adj) = 76.0% Analysis of Variance SOURCE Regression Error Total DF 1 6 7 SS 4956347 1283854 6240202 MS 4956347 213976 F p 23.16 0.003 i) If the exchange rate is 5 Francs per dollar, how many visitors can Europe expect? (1.5) ii) Is the regression coefficient of Fper$ significant at the 1% level? Why? (2) iii) What two items on the printout could be used to compute R-sq? (2) Solution: (i) visitors = 1768 + 456(5)=4048. H0 : 1 0 6 1,6 (ii) Accept H 0 if t lies between t .005 or if p-value is above .01 or if F lies below F.01 . H : 0 1 1 Since these are all false we reject H 0 and say that 1 is significant. RSS 4956347 (iii) R 2 .794 TSS 6240202 2 4/14/98 252y9831 III. Do at least 2 of the following 3 Problems (at least 12 each) (or do sections adding to at least 24 points Anything extra you do helps, and grades wrap around) . Show your work! State H 0 and H1 where applicable. You must do 1a! 1. A researcher interviews 3 independent samples of 5 individuals of different ages and asks them to name the seven dwarfs. The amount of time in seconds for the three samples is below. For your convenience the following facts are given: x 3 350, x 2 3 x 1 208, x 2 1 ?, x 2 443, x 2 2 47029, 26300 . a) Find the sample variance of x1 (2) b)Do a test of the equality of the means at a significance level of 5% . (Assume that the underlying distributions are normal) (10) c) Do a confidence interval for the difference between the means of the second and third populations on the assumption that this interval is one of the three possible contrasts. (3) 2 x12 nx12 9194 5 414 . a) s x21 135.3 n 1 4 x1 x2 x3 x12 people people people under 12 12-25 over 25 60 120 75 3600 45 65 45 2025 38 60 75 1444 30 48 55 900 35 150 100 1225 Sum x 1001 208 443 350 b) nj 5 5 5 x. j 41.6 88.6 70.0 x SS 9194 47029 26300 82523 1730.56 7849.96 4900 x 2 .j 2 9194 x n 15 x n 66.7333 x 14480.47 x 2 2 .j x n x 82523 15 66.7333 15722.9333 SSB n x n x 5 x 15 x 514480.47 15 66.7333 SST 2 2 j 2 2 .j 2 2 .j H 0 : 1 2 3 H 1: Not all means equal. Source Between Within Total 2 1 SS 5602.282 10120.65 15722.933 2 ,12 Since F F.05 , accept H 0 . DF 2 12 14 MS 2801.1415 843.3875 F 3.321 2 5602.283 2 ,12 F.05 3.89 NS 3 4/14/98 252y9831 c) 2 3 x 2 x 3 1 1 1 1 88.6 70.0 2 389 . 843.3875 5 5 n1 n2 m 1 F m1,nm s 8.6 2624.6219 8.6 5123 . 4 4/14/98 252y9831 2. a) Data from 1) is repeated below. Assuming that the underlying distributions are not normal, do a test of the equality of medians. (6) x1 x2 x3 r1 r2 r3 people people people under 12 12-25 over 25 60 120 75 9 14 11 45 65 45 4 10 5 38 60 75 3 8 12 30 48 55 1 6 7 35 150 100 2 15 13 19 53.5 47.5 b) Complete the following ANOVA table for three speeds, four temperatures and three measurements per cell. Name the three tests and their conclusions (5). SOURCE speed temperature interaction within total SS DF 227.06 113.44 142.06 MS F 683.22 c) The author of the text mentions a randomized block experiment with 5 treatments and six blocks for which SSTR = 287, SST = 1446, and SSE = 180. Is there a difference between treatment means? (4) Solution: a) H 0 : 1 2 3 H 1: Not all medians equal. check: 19 535 . 47.5 120 H 1516 2 12 19 2 535. 2 47.5 2 5479.5 316 .05 48 6.795 5 5 5 1516 5 Since the Kruskal-Wallis table says that 5.78 is the 4.9% value and 7.98 is the 1% value, the p-value is below 5% and we reject H 0 . b) Total degrees of Freedom are (3)(4)(3)-1 = 36 - 1 = 35. Degrees of freedom for speed are 3-1=2. Degrees of Freedom for Temperature are 4 - 1 = 3. Degrees of Freedom for Interaction are (2)(3) = 6. SOURCE SS DF MS F F.05 2 ,24 F.05 3.40 Speed means not equal speed 227.06 2 113.53 13.57 3,24 F.05 3.01 Temperature means not equal temperature 113.44 3 37.813 4.52 6,24 F.05 2.51 Interaction is present interaction 142.06 6 23.68 2.83 within 200.66 24 8.3608 total 683.22 35 c) F.05 SOURCE SS DF MS F 4 ,20 F.05 2.71 S Treatments 287 4 71.75 7.97 5,20 F.05 2.60 S Blocks 979 5 195.8 21.76 Error 180 20 9.0 Total 1446 29 The F test shows a difference between both block and treatment means. 5 4/14/98 252y9831 3. The following data represents gross farm income and value of land and buildings (in billions of dollars) on farms in a random sample of 11 states. y income 2.6 1.8 3.9 15.4 3.7 0.4 0.5 4.9 4.1 2.6 8.1 x value 9.3 11.1 15.0 63.3 16.2 1.4 1.2 20.9 12.3 12.0 51.0 For your convenience the following values are given: x 213.7, x 8040.53, y 48.0, y 389.66, xy 1735.72, n 11 2 2 213.7 19.4273 11 48.0 y 4.3636 11 x a. Compute the regression equation Y b b x (6) Spare Parts: b. Compute R 2 . (4) c. Compute s e . (3) 3888.92 0 x 1 2 nx 2 8040.53 1119.4273 2 xy nxy 1735.72 1119.4273 4.3676 d. Compute sb1 and do a significance test on b1 803.21 y .(4) e. Do a prediction interval for Y when x 5 . (5) 2 ny 2 389.66 11 4.3636 2 180.21 TSS Solution: a. b1 xy nxy x nx 2 2 803.21 0.2065 3888.92 b0 y b1 x 4.3636 0.206519.4272 0.3518 Y b0 b1 x becomes Y 0.35180 0.2065x . b. R 2 b1 xy nxy 0.2065803.21 RSS 0.9206 or 180.21 TSS y ny xy nxy 803.21 .9206 x nx y ny 3888.92180.21 y ny b xy nxy 180.21 0.2065803.21 ESS 1594 s . where 2 2 2 R 2 2 2 2 c. 2 2 2 2 2 e ESS TSS RSS or 1 n2 se2 y n2 9 2 ny 2 n2 1 R or s y 2 2 e 2 x ny 2 b12 2 nx 2 n2 6 4/14/98 252y9831 se se2 1594 . 1263 . d. sb21 se2 x nx 2 2 1594 . 0.00041 3888.92 H 0 : 1 0 H 1: 1 0 t sb1 0.00041 0.0205 b1 10 b1 0 0.2065 . Since this is not between 101975 . sb1 sb1 .0205 n2 9 t .025 t .025 2.262 , reject H 0 and conclude that 1 is significant. e. If Y0 0.35180 0.2065x0 and x 0 5 , then Y0 0.35180 0.20655 1385 . 1 s 2y se2 0 n x0 x 2 x 2 nx 2 1 5 19.4273 2 1 1594 . 1 1824 . 11 3888.92 s y0 1824 . 1351 . . . 2.2621351 . 138 . 3.06 . So y 0 y 0 t s y0 1385 7