advanced biostat.pdf
- 1. 1 DAMBI DOLLO UNIVERSITY COLLEGE OF MEDICINE and HEALTH SCIENCES ADVANCED BIOSTATISTICS FINAL EXAMINATION Time allotted: 2:30 Hrs. Date: March /2023 Name:______________________________________________ID No:___________ Department: ___________________________________ Instructions: - Make sure that you are writing your name, ID number and department. Check whether the exam paper contains 13 pages including the cover page and II parts Read each instruction carefully and attempt all questions clearly Switch off your cell phone and use only your property i.e., Calculator, etc. Write clearly, unreadable answers will not be marked.
- 2. 2 Part I. Match items under column A with appropriate items under column B. Part I. Choose the best answer for the following questions 1. The researcher wants to study whether there is a significant difference in mean weight gain among of two groups of children after feeding two different diets for six months. To construct a valid 95% confidence interval, for the difference of population mean weight gain, all of the following assumptions are required; except? A. Distribution of the population whether normally distributed or not B. Independence of the population C. Sample size D. Equality variance 2. Consider question #1, if 95% confidence interval for the difference in the mean weight gain of the two population is (-2.15 kg, -1.04 kg), what will be your possible conclusion? (Assuming the direction of difference is the mean of the first population minus the mean of the second population) A. The mean weight gain for the first population is less than the mean weight gain for the second population. B. The mean weight gain for the second population is less than mean weight gain for the first population. C. There is no difference in the mean weight gain between the two populations or groups. D. It impossible to make any conclusion about the mean difference between the two populations or groups from the given confidence interval. 3. Consider question #2, what will be the precession of the estimate? A. 1.04 B. -2.15 C. 1.595 D. -1.11
- 3. 3 A pediatrician working at Dambi Dollo hospital wants to study whether the birth weight of children from rural mothers is different from children born from urban mothers. Previously it is known that the average birth weight of babies from urban mothers was 2.5 kg. A random sample infants was selected from rural infant population and, their weight was measured. From data obtained from rural babies and the average birth weight of babies from urban mothers, the following result was obtained. Based on the above-given information answer the following questions #4-6. 4. What will be the possible null hypothesis for the above research question? A. The mean birth weight of children from a rural area is less than the mean birth weight of children from an urban area. B. The mean birth weight of children from a rural area is greater than the mean birth weight of children from urban areas. C. There is no difference between the mean birth weight of children from rural and urban areas. D. It is not possible to change the given research question into statistical hypothesis. 5. Based on the given out to put what will be your conclusion? A. The alternative hypothesis is rejected B. The alternative hypothesis accepted C. It needs an extra sample size to make a decision on whether to reject or accept the null hypothesis. D. No hypothesis can be tested from the given output 6. What is degree of freedom for the test? A. 98 B. 49 C. 50 D. 25
- 4. 4 7. How can we tell which means are different in an ANOVA test? A. p-value associated with the f-test B. Alpha-level C. f-ratio value D. Post hoc tests A researcher wants to study whether there is a difference in mother’s average length of stay after giving birth in the four hospitals. With this objective, the collected data from the four hospitals were analyzed and the following output was obtained. Based on the above-given data, (Assuming that the data meet all the assumptions) give the answer to the following questions. 8. What will be the possible alternative hypothesis for this research question? A. There is no difference in the mean stay of mothers after delivery in the hospitals. B. The mean stay of mothers in the hospital is different for all hospitals. C. At least the mean stay of mothers is different for at least one of the hospitals. D. It is not possible to state the hypothesis for the current study. 9. What do we conclude about the assumption of homogeneity of variance from the given output? A. The variance of Najo hospital is greater than the variance in other hospitals. B. There is a difference in the variance of the length of stay in the hospitals. C. The length of stay in the hospital is heterogeneous.
- 5. 5 D. The output for test homogeneity of variance is not given. 10. How many individuals have participated in this study from all hospitals A. 109 B. 112 C. 113 D. 107 11. Based on the above output what can you say about the post-hock test? A. Since the null hypothesis was rejected conducting post hock test is not relevant. B. Since the null hypothesis was rejected, it needs to conduct post hock test. C. Since the null hypothesis was accepted conducting post-hock test is not relevant. D. Since the null hypothesis was accepted, it needs to conduct post hock test. 12. All of the followings are assumptions of linear regression, except? A. The dependent variable should be continuous type. B. The relationship between the response and the explanatory variables is linear in parameters. C. The independent variable must be continuous D. The residual should be normally distributed 13. Which of the following method(s) is not used for predicting continuous dependent variables? A. Linear Regression C) Two-way ANOVA B. Logistic Regression D) Survival Analysis 14. Selection of variables by using the combinations forward and backward methods is known as: A. Forward selection C) Stepwise selection B. Backward Elimination D) A & B 15. A correlation between age and the health of a person was found to be 0.9. On the basis of this, you would tell the doctors that: A. Age is good predictor of health C) A & B B. Age is a poor predictor of health D) None of these
- 6. 6 16. Which of the following measurement (metrics) can be used for evaluating regression models? A. Adjusted R Squared C) R Squared B. F Statistics D) All 17. The following output was obtained to test the existence of a linear relationship between Albumin level and age person. Based on the output, what do you conclude about the relationship between the two variables? A. There is no linear association between the two variables. B. There is a positive weak linear association between the Albumin level and the age of the person. C. When the age of the person increases by one year the Albumin level decreases by - 0.123. D. It is not possible to conclude the possible association between Albumin level and the age of the person from the plot. 18. Suppose that the following two models have identical R2 values. Which one of the following statements is true? A. The two models will also have identical values of adjusted R-square B. Model 2 must have a higher value of adjusted R-square C. Model 2 must have a lower value of adjusted R-square
- 7. 7 D. It is not possible to determine which model will have the higher R-square without knowing the sample size. Work Sheet Note: Give the answer on the shaded region Name of student:______________________ ID No. __________________ Part I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Part II. Workout 1. Suppose that we are interested in identifying factors associated with HIV status of young adults. During the course of the study, the data were collected on the relevant variables identified from the literature. In the data, HIV status is a dichotomous random variable that takes the value of 1 if the person is positive and 0 if it is not. The study used a sample of 10000 young adults to estimate the probability, p, being positive. For this study Age of the respondents (in a year), drinking habits, physically active and marital statuses were included in the study as the explanatory variable. Based on the available data on the study variables logistic regression was fitted and the following output was obtained from SPSS.
- 8. 8
- 9. 9 Based on the above outputs answer the following questions: A. Based on the above SPSS output, write the equation for the estimated log-odds of HIV positive for a young adult with (Age of x1, physically active of x2 and marital status of x3). ( 2-points)
- 10. 10 B. State the hypothesis to test the association between Age and HIV status. Test this hypothesis at alpha value 0.05. Based on your test what do you conclude? ( 2-points)
- 11. 11 C. State the hypothesis for Hosmer and Lemeshow model fit test. What do you conclude if the null hypothesis is accepted? ( 2-points) D. What percentage of observation can be correctly predicted by the model? ( 2-points) E. Interpret the estimated coefficients of age and marital status using the fitted values and appropriate unit of measurement. ( 3-points)
- 12. 12 F. Interpret the Odds of being HIV positive for marital status and being physically active. ( 3-points) G. What is the 95% confidence interval for the OR for marital status that adjusts for other predictor variables? Does the interval contain the null value of 1? What does it tell us? ( 2-points) H. Write a short summary report for the above studies based on the given outputs. ( 3-points)
- 13. 13