Sei sulla pagina 1di 3

Math 54 Chapter 10 problem handout answers Problems 1 5 Two proportions, independent samples (10.3) 1.

1. This one was done in both classes 2. Pop 1 no seat belt Pop 2 Seat belt n1 = 2823 X1 = 31 n2 = 7765 x2 = 16

a. 90% confidence interval : (.00559, .01226) We are 90% confidence that the difference in the proportion killed not wearing a seat belt and those wearing a seat belt is between 0.00559 and 0.01226 (.559% to 1.226%). Since the interval is entirely positive, we are 90% certain that the proportion killed while not wearing seat belt is greater than the proportion kill while wearing a seat belt. b. Claim: p1>p2 Ho: p1p2 Ha: p1>p2 Test statistic: z = 6.11 p-value = 5.134 x 10-10 since p-val < = 0.10 reject Ho. There is sufficient evidence to conclude that the proportion killed while not wearing a seat belt is greater than the proportion killed while wearing a seat belt. (Possible type 1 error) It appears that seat belts reduce the proportion killed. 3. Pop 1 Female survivors Pop 2 Male survivors n1 = 2739 n2 = 1352 X1 = 1397 x2 = 436 a. Claim: p1p2 Ho: p1=p2 Ha: p1p2 Test statistic: z = 11.35 p-value = 7.88 x 10-30 since p-val < = 0.01 reject Ho. There is sufficient evidence to conclude that the proportion of female survivors that developed thyroid disease is different than the proportion of male survivors. (possible type 1 error) b. 99% confidence interval: (.1466,.2285) We are 995 confident that the difference in the proportion of female and male survivors who developed thyroid disease is between .1466 and .2285 (14.66% and 22.85%) Since the interval is entirely positive, we are 99% certain that the proportion of female survivors who developed thyroid disease is greater than the proportion of male survivors. c. Since the hypothesis test was two-tailed they do give the same result. 4. This one was done in both classes. 5. This one was done in both classes.

Problems 6 9 are comparing two means from independent populations section 10.2 6. Done in both classes 7. Pop 1 high interest Pop 2 low interest

x1 =594.8

x2 =785.2

s2=16.3 s1=12.2 n1=40 n2=40 a. Claim: 1<2 Ho: 12 Ha: 1<2 Test statistic: t = -59.15 p-val = 0 <=0.01 Reject Ho. (Possible type I error) There is sufficient sample evidence to support the claim that the mean FICO score for high interest loans is less than that for low interest loans. b. 98% confidence interval: (-198.1, -182.7) We are 98% confident that the difference in the mean FICO score between low interest loans and high interest loans is between -198.1 and -182.7. Since the interval is entirely negative the mean for low interest loans is less at 98% confidence. 8. Done in both classes Note: For problem #8 there is a typo in the data set for recent winners, the first number should be 19.5 not 49.5. All calculations below use 19.5.

9. First find the sample statistics for each group using the given data sets. Pop 1 recent winners Pop 2 (20s and 30s winners)

x1 =18.6

x2 =20.16

s2=1.48 s1=1.20 n1=10 n2=10 a. Claim: 1<2 Ho: 12 Ha: 1<2 Test statistic: t = -2.48 p-val = 0.0119 <=0.05 Reject Ho. (Possible type 1 error) There is sufficient sample evidence to conclude that the mean BMI for recent winners is less Than that of winners from the 20s and 30s. b. 90% confidence interval: (-3.1, 0.05) We are 90% confident that the difference in the mean BMI for recent winners and winners from the 20s and 30s is in the interval (-3.1, 0.05). Since the interval contains 0, we conclude that the mean BMI for recent winners and for winnners from the 20s and 30s are statistically equal. *note the hypothesis test and the interval reach different conclusions because they are not equivalent. Note in both cases the decision was close. c. Based on the hypothesis test it does appear that the mean BMI has decreased, but based on the interval we cannot conclude that it has.

Problems 10 13 are matched pair samples, section 10.1 For each of these problems the first step is to find the list of differences and calculate the sample statistics for the set of differences. I have given the sample statistics, but not the set of difference. 10. Sample statistics: xd = -0.9 sd=1.35 n=6

a. 95% confidence interval: (-2.3, 0.52). We are 95% confident that the mean difference between morning and noon temperatures is between -2.3 degrees and 0.52 degrees. Since the interval contains 0, we conclude that the mean temperatures are statistically equal. b. Claim: d 0 Ho: d =0 Ha: d 0 Test statistic: t = -1.63 Critical values: 2.571 P-value: 0.1634 Using either method the conclusion would be Fail to Reject Ho, so there is not sufficient evidence to conclude that there is a difference in temp. (Possible type 2 error) 11. Sample statistics: xd = 839.56 sd=1935.08 n=9 Ho: d =0 Ha: d 0 a. Claim: d 0 Test statistic: t = 1.30 Critical values: 2.306 P-value: 0.2293 Using either method, fail to reject Ho, so there is not sufficient evidence to conclude that there is a difference in the two costs. (Possible type 2 error) b. 95% interval: (-647, 9.23) We are 95% confident that the mean difference between the costs is between -$647 and $9.37. Since the interval contains 0 we conclude tha the costs are statistically equal. 12. This one was done in both classes data was written on the board. 13. Sample statistics: xd = -3.33 sd=3.01 n=6 Ho: d 0 Ha: d <0 a. Claim: d <0 Test statistic: t = -2.71 Critical value: -2.015 P-value: 0.0211 Using either method, Reject Ho, there is sufficient evidence to conclude hospital admissions were higher on Friday the 13th. It does not imply Friday the 13th is unlucky, there could be many other factors in this situation. b. 95% interval: (-6.5, -0.17) We are 95% confident that the mean difference in hospital admissions is between -6.5 admissions and -0.17 admissions. Since the interval is entirely negative we conclude that the admissions on Friday the 13th were higher.

Potrebbero piacerti anche