MAT222 Test 2 Fall 2009 1. Out of a random sample of 50 accounts in a certain department store, 40 out of 50 customers had paid their outstanding bills on time for a given month. (a) Construct a 95% confidence interval of the true proportion of customers who had paid their outstanding bills on time. (10 points) (b) Test the null hypothesis that p= .6 versus the alternative hypothesis that p >.6 at .05 level of significance. (10 points) (a). x 40 pˆ .8 n 50 SE (b). p̂(1 - p̂ ) n .16 .0032 .0565 50 95% confidence interval for p: p̂ z * SE = (.8) 1.96(.0565) =.8 .1108 [.7892, .9108] H0: p = .6l vs Ha: p > .6 z p̂ - .6 p0 (1 p0 ) n .8 - .6 .2 2.89 p-value = P(Z > 2.89) < .0019 < .6(1 .6) .0693 n 05 Reject H0 at level = .05. 2. A test question is considered good if it differentiates between prepared and unprepared students. The first question on a test was answered correctly by 62 of 80 prepared students and by 26 of 50 unprepared students. (a) Perform a significance test for the null hypothesis H 0 : p1 p2 versus H a : p1 p2 where p1 is the population proportion of correct answer for prepared students and p 2 is the population proportion of correct answer for unprepared students. What do you conclude? (10 points) (b) Find a 95% confidence interval for p1 - p 2 . (10 points) 62 26 pˆ 1 .775. pˆ 2 .52 (a) 80 50 62 26 88 pˆ pooled .677 80 50 130 SE Dp p̂(1 - p̂)( z 1 1 1 1 ) .677 .323( ) .0843 n 1 n2 80 50 p̂1 - p̂ 2 .775 .52 3.02 SE Dp .0843 p-value = P(Z > 3.02) = ..0018 < 05 Reject H0 at level = .05; The proportion of correct answer for prepared students is higher than the proportion of correct answer for unprepared students. (b) SE Dˆ p̂1 (1 - p̂1 ) p̂ 2 (1 - p̂ 2 ) .775(1 .775) .52(1 .52) .00218 00499 .0847 n1 n2 80 50 95% confidence interval for p1 - p2: D z * SE D̂ = (.775 - .52) 1.96(.0847) =.255 .166 [.089, .4221] 3. The following table pertains to a study of the relationship between the speed of promotion and the standard of clothing of bank employees: Poorly dressed Slow 10 Speed of Promotion Average 18 Fast 14 Total 42 Well dressed 16 22 20 58 Total 26 40 34 100 (a) Give the joint distribution of “speed of promotion” and “dressed” for this table. (5 points) Speed of Promotion Slow Average Fast Poorly dressed .10 .18 .14 Well dressed .16 .22 .20 (b) Computer the marginal distribution for the speed of promotion. (5 points) P(slow) = 26/100 = .26, P(Average) = 40/100 = .40, P(fast)=34/100 = .34 (c) Compute the conditional distribution of the “speed of promotion” for “poorly dressed” and “well dressed”. (5 points) P(slow|poorly dressed) = 10/42 = .238 P(Average|poorly dressed) = 18/42 = .428 P(fast|poorly dressed)=14/42 = .333. P(slow|well dressed) = 16/58 = .276 P(Average|well dressed) = 22/58 = .379 P(fast|well dressed)=20/58 = .345. We wish to use a chi-square test at .05 level of significance to test whether the speed of promotion and the standard of clothing are independence. (d) State the null and alternative hypotheses for this problem. (5 points) H0 : Speed of promotion and Clothing are independent/or unrelated. Ha : Speed of promotion and Clothing are dependent/or related. . (e) Test the hypotheses in (d) at 5% level of significance and make a conclusion. (10 points) Expected counts: 10.92, 16.8, 14.28,; 15.08, 23.20, 19.72 2 (observed exp ected ) 2 .0775 .0857 .0055 .0561 .0620 .0039 .2907 exp ected degree of freedom = (2 – 1)(3 – 1) = 2 .25 < p-value Fail to reject H0 Speed of promotion and Clothing are independent/or unrelated. 4. (a) Find a 95% confidence interval for the slope in the following setting: n = 20, yˆ - 2.8 5.3 x , and SEb1 = 2.2 (5 points) Use d.f. = 18, t* = 2.093 -5.3 (2.101)(2.2) = -5.3 4.62 = [-9.92, -.68] (b) Test the null hypothesis that the slope is 0 versus the two-sided alternative. (5 points) H0: =0 versus Ha: 0, at =.05, t = -5.3/2.2 = -2.41, p-value = 2P{t19 < -2.41} d.f. = 18, 2.214 < 2.41 < 2.552 .01 < upper tail probability < .02 Hence .02 < p-value < .04 Reject H0 at level = .05 5. Exercise 2.144 on Page 163 of our textbook gives the modulus of elasticity (MOE) and the modulus of rupture (MOR) for 32 plywood specimens. We regress MOR on the MOE for the 32 specimens. Here is part of the MINITAB output for the regression: Results for: EX10_021.MTP Regression Analysis: mor versus moe The regression equation is mor = 2653 + 0.00474 moe Analysis of Variance Source Regression Residual Error Total DF 1 30 31 SS 75573461 45987465 121560926 MS F 75573461 49.30 1532915.5 P 0.000 (a) Complete the analysis of variance table by filling in the “Residual Error” row. (5 points) (b) What are the values of the regression standard errors s and the squared correlation r2? (5 points) S = square root of MSE =1239.1; r2 = SSM/SST = 75573461/121560926 = .62 (c) The standard deviation of the moe is 325293. Find the standard error for the leastsquares slope b1. (5 points) 1 Sx ( xi x ) 2 325293 ( x i x ) 2 325293 31 1781703.1 n 1 SEb1 s (x i x) 2 1239.1 .000695 1781703.1 (d) Give a 95% confidence interval for the slope 1 of the regression line. (5 points) d.f =30, t* = 2.042 95% confidence interval for the slope 1: b1 t * SEb1 .00474 2.042(.000695) .00474 .00142 [.00332,.00616]