Question 1

1. A confidence interval is computed from a sample selected from a population. Based on the concept of confidence interval, please answer questions 1-2.
2. There are two simple random samples selected separately from two different populations. The population with low variability will produce estimates with large margins of error assuming two samples have same sample size.
3. a. True
4. b. False

20 points

Question 2

1. There are two simple random samples selected from the same population. The sample with small sample sizes will tend to have large margins of error.
2. a. True
3. b. False

20 points

Question 3

1. The mean of a simple random sample selected from a population is used to estimate the population mean. Based on statistical inference, please answer questions 3 - 5.
2. The mean from a sample of 9 is 10 with sample standard deviation 3.
3. This sample is used to test the hypothesis:
4. Given that the critical value of the t-statistic with 8 degree freedom and alpha-level of 0.05 is 2.306, the null hypothesis is
5. a. Can not be determined
6. b. Rejected
7. c. Retained

20 points

Question 4

1. The mean from a sample of 9 is 10 with sample standard deviation 3. Given that the critical value of t-statistics with 8 degree freedom and alpha-level of 0.05 is 2.306, the 95% conference interval calculated from this sample includes the value 8:
2. a. False
3. b. Can not be determined
4. c. True

20 points

Question 5

1. Please answer which statistical test is most appropriate for questions 5-8.
2. 10 new born boys and 10 new born girls are randomly selected from a hospital. The test which is used to compare mean of birth weight between boys and girls is
3. a. ANOVA
4. b. Paired-sample
5. c. One-sample
6. d. Two-sample

20 points

Question 6

1. 10 new born babies are randomly selected from a hospital. The test which is used to compare the mean of birth weight to a known national mean is
2. a. One-sample
3. b. Two-sample
4. c. ANOVA
5. d. Paired-sample

20 points

Question 7

1. 30 students who took the MCAT once only are selected. 10 took the MCAT at the end of second college year, 10 at the end of third college year and 10 after college. The test which is used to compare the mean of MCAT scores among these three students groups is
2. a. Two-sample
3. b. ANOVA
4. c. One-sample
5. d. Paired-sample

20 points

Question 8

1. 10 male students and 10 female students are randomly selected from those students who took the SAT test twice. The first SAT was taken before taking Kaplan preparation course, and the second SAT after taking Kaplan preparation course. The test which is used to evaluate the Kaplan preparation course is
2. a. ANOVA
3. b. Two-sample
4. c. Paired-sample
5. d. One-sample

20 points

Question 9

1. A correlation coefficient is used to present a relationship between two variables. It can not be negative number.
2. a. False
3. b. True

20 points

Question 10

1. A multiple regression equation of new born babies' systolic blood pressure (SBP) accounted by systolic blood pressures from mother and father is
2. Baby_SBP = - 5 + 0.5*Mother_SBP + 0.4*Father_SBP
3. Based on this equation, please answer questions 13-15.
4. If mother A's SBP is 10 mmHg higher than mother B's SBP, and if both fathers have the same SBP, the predicted Baby A SBP is higher than Baby B by
5. a. 4
6. b. 5
7. c. Can not be determined
8. d. -5

20 points

Question 11

1. If a mother's SBP is 120, the baby's SBP is
2. a. 83.91
3. b. 120
4. c. Can not be determined
5. d. 99.6

20 points

Question 12

1. If father A's SBP is higher than father B's SBP, Baby A's SBP is higher than Baby B's.
2. a. True
3. b. False

20 points

Question 13

1. The proportion of successes in a sample is 0.5. The sample size is 100. The 95% confidence interval is
2. a. 0.402 to 0.598
3. b. 0.500 to 0.598
4. c. 0.402 to 0.500
5. d. 0.302 to 0.698

20 points

Question 14

1. Disease data for two independent groups are associated with the summary statistics reported in the table below. Given summary statistics in the table, please answer questions 17-18.
2. Group
3. Disease
4. Non-disease
5. Total
6. 1
7. 8
8. 92
9. 100
10. 2
11. 4
12. 96
13. 100
14. Total
15. 12
16. 188
17. 200
20. The Fisher exact test is
21. a. unnecessary
22. b. necessary
23. c. Can not be determined

20 points

Question 15

1. What is the overall sample proportion of disease (combining two groups together)?
2. a. Can not be determined
3. b. 6%
4. c. 8%
5. d. 3%
6. e. 4%

20 points

Question 16

1. What is the relative risk ratio (group 1 over group 2)?
2. a. 3
3. b. 2
4. c. 1.5
5. d. 1

20 points

Question 17

1. What is the risk difference (group 1 - group 2)?
2. a. 0.03
3. b. 0.08
4. c. 0.06
5. d. 0.04

20 points

Question 18

1. An investigator is trying to state a null hypothesis of relative risk ratio. The null hypothesis is
2. a. 0.5
3. b. 1
4. c. 1.22
5. d. 1.49

20 points

Question 19

1. A disease data for two independent groups are summarized into 2 by 2 tables stratified by gender. Based on Mantel-Haenszel procedure, the Mantel-Haenszel summary relative risk ratio is 1.6 and the crude relative risk ratio 1.5, the gender variable is
2. a. Can not be determined
3. b. Positive confounders
4. c. Negative confounders

20 points

Question 20

1. A disease data for two independent groups are summarized into 2 by 2 tables stratified by gender. The Chi-Square test for interaction between group variable and gender variable shows that a statistical significant interaction does NOT exist. Which following statistic should be reported?
2. a. The Mantel-Haenszel summary odds ratio
3. b. Can not be determined
4. c. The crude odds ratio from combined data
5. d. The strata-specific odds ratio